Tropical Forestry Handbook [2 ed.] 9783642546006, 9783642546013

Table of contents :
Front Matter....Pages i-lx
Front Matter....Pages 1-1
Front Matter....Pages 3-45
Front Matter....Pages 47-90
Front Matter....Pages 91-91
Front Matter....Pages 93-283
Front Matter....Pages 285-291
Front Matter....Pages 293-301
Front Matter....Pages 303-331
Front Matter....Pages 333-342
Front Matter....Pages 343-361
Front Matter....Pages 363-390
Front Matter....Pages 391-403
Front Matter....Pages 405-412
Front Matter....Pages 413-427
Back Matter....Pages 429-444
....Pages 445-450

Citation preview

Laslo Pancel Michael Köhl Editors

Tropical Forestry Handbook Second Edition

1 3Reference

Tropical Forestry Handbook

Laslo Pancel • Michael Köhl Editors

Tropical Forestry Handbook Second Edition

with 995 Figures and 370 Tables

Editors Laslo Pancel Deutsche Gesellschaft f€ur Internationale Zusammenarbeit (GIZ) GmbH La Libertad, El Salvador

Michael Köhl Center for Wood Sciences Institute of World Forestry University of Hamburg Hamburg, Germany

ISBN 978-3-642-54600-6 ISBN 978-3-642-54601-3 (eBook) ISBN 978-3-642-54602-0 (print and electronic bundle) DOI 10.1007/978-3-642-54601-3 Library of Congress Control Number: 2015957807 # Springer-Verlag Berlin Heidelberg 1993, 2016 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer-Verlag GmbH Berlin Heidelberg

We dedicate this Handbook to our Heavenly Father and entrust ourselves to His guidance and protection as we reverently manage His marvelous donation to humanity so that future generations may experience the harmony, peace, and spirit of tropical forests.

Preface

Long ago, European forests were wild and undisturbed. Our ancestors did not experience those forests as locations of harmony and well-being but as spooky and scary scenes of creepy stories with kidnapped princesses, ruthless witches, and cruel noblemen. Then, humans entered the forests, utilized their timber, and cleared forests for other land use. They shaped forests according to their interests and later invented treatments to increase their sustainable productivity, which resulted in uniform, single-species stands. At the beginning of the last century, a slow shift of paradigm took place in Europe. Forest sustainability was understood as the interplay of different ecological, economic, and social functions and a continuous cover forestry with mixed species stands, layered structures, and selective cutting became the desirable forest practice. It looks as Europe has been saved by the bell. Today, tropical forests are one of the top ecological trouble spots in the world. They seem to share the fate of early European forests, but due to modern technology, global markets, and pursuit of profit at a much faster progress, they are destroyed at an alarming pace. Tropical forests find themselves in a field of tension between different user interests. Ecosystems and their biodiversity have to be protected; tropical forests have to be maintained as important drivers of the global climate system and a major sink for atmospheric carbon dioxide. Wood fuel is the most important source of energy for 70 % of the world’s population. The use of timber and subsequent timber processing safeguards income and employment. Population growth and the reduction of poverty and hunger enforce the extension of agricultural areas, and the extreme excess of the politically well-anchored demands for renewable energy and resources may complete to do the rest for the tropical forest resources. These contradictory and partially existential user interests cannot be denied but need to be addressed in a holistic approach. To create a fair balance between interest groups renders sophisticated levels of knowledge necessary, incorporating disciplinary, interdisciplinary, and transdisciplinary approaches. However, where the economic interests of a minority gain the upper hand, our expertise is helpless and powerless. We contributed our entire professional life to counter the destruction of tropical forests. We are in no doubt that population growth and economic development of any country – wherever located on the globe – will result in increasing forest utilization and conversion of forest areas to other land use. We cannot hold this vii

viii

Preface

process. However, we can contribute to minimize the collateral damage wherever tropical forests are utilized and to identify forest areas of unique value, which should be excluded from any human intervention. The Tropical Forestry Handbook has been prepared by professionals that are dedicated to the protection and the sustainable management of tropical forests. They want to make their expertise available for a wide audience. Their contributions provide valuable insight in sustainable tropical forest management. It is our hope that the Tropical Forestry Handbook will become a helpful tool for mastering the current and future challenges of tropical forests. It is our hope that this book acquires a liking for young people especially. And it is our hope that readers join us and will not be released ever again from the miscellany, uniqueness, and fascination of tropical forests. Laslo Pancel San Salvador

Michael Köhl Hamburg

Abbreviation Group

Tables and Formulas Conversion to International System Units2 See Table 1.

2

Source: Euroconsult (1989) Agricultural compendium. Elsevier, Amsterdam ix

Meter

Symbol

l

A

V

t

Area

Volume

Time

Hour Minute (60−1 h) Second (60−2 h)

Cubic meter

Square meter

ISa notation Name

Magnitude Name Space and time Length

Table 1

Year

Day

s

Cubic centimeter Liter Register ton

m3

h min

Are

cc l rt

a

24 h 86,400 s

8766 h 31,577,600 s

×10−6 ×10−3 ×2.38

×102

Foot (12 in) Yard (3 ft) Chain (22 yd) Furlong 220 yd) Mile (1760 yd) Nautical mile Square inch Square foot Square yard Acre (4840 yd2) Square mile (640 acres) Cubic inch Cubic foot Cubic yard Gallon (US) Gallon (UK) Pint (UK) Buschel (8 gal. US) Acre foot See IS and other notations

Å

Angstrom

−10−10

Inch

×10−6

μ

Micron

acre ft

in3 ft3 yd3

sq•mile

mile n mile in2 ft2 yd2

ft yd

in

Anglo-Saxon notations Name Symbol

Conversion to IS

Symbol

Other notations Name

m2

m

Symbol

×16.39×106 ×28.32×10−3 ×0.765 ×3.785×10−3 ×4.546×10−3 ×0.567×10−3 ×30.28×10−3 ×1233.5256

×25.400× 10−3 ×0.3048 ×0.9144 ×20.1168 ×201.168 ×1.6093×103 ×1.852×103 ×0.645×103 ×92.9×10−3 ×0.836 ×4.047×103 ×2.590×106

Conversion to IS

x Abbreviation Group

a

q

L

m

Volumetric rate of flow

Angle

Mechanics Mass

V

Acceleration

Radial velocity

Velocity

Kilogram

Radian per second Meter per second squared Cubic meter per second Radian

Meter per second

kg

rad

m3 s−1

m s−2

rad s−1

m s−1

g

Gon (centesimal degree) Centesimal minute

Quintal Ton

Centesimal second

″

Second

q t

cc

c

° ′

km h−1

Degree Minute

See angle

Kilometer per hour

×102 ×103

×π×π 100−1 ×200−1 ×π×π 100−2 ×200−1

×π×180−1 ×π×60−1 ×180−1 ×π×60−2 ×180−1 ×π 200−1

×3.6

Grain (7000−1 lb) Ounce (avoirdupois) Pound (avoirdupois) (16 oz) Stone (14 lb) Hundredweight (112 lb)

See other notations

Cubic foot per second (cusec)

Foot per second squared

Foot per minute Foot per second Mile per hour Knot (1 nautical mile per hour) See angle

×6.35 ×50.8 cwt

(continued)

×64.8×10−6 ×28.35×10−3 ×0.4536

×28.317× 10−3

×304.800× 10−3

×5.080×10−3 ×0.3048 ×0.4470 ×0.514

gr oz lb

ft3/s

ft/s2

ft/min ft/s mile/h kn

Abbreviation Group xi

Kilogram meter squared Newton

Newton per cubic meter

I, J

F W

γ

Moment of inertia

Force Weight

Specific weight

Kilogram per cubic meter

ρ

Symbol

ISa notation Name

Density concentration

Magnitude Name

Table 1 (continued)

N m−3

N

kg m2

kg m−3

Symbol

Kilogram force per cubic metre Kilogram force per litre

Kilogram force Dyne

Kilogram per litre Ton per cubic metre

Other notations Name

kgf l−1

kgf m−3

×9.80665 ×10−3

×9.80665

×9.80665 ×10−5

×103

t m−3

kgf dyn

×103

Conversion to IS

kg l−1

Symbol

Poundal Poundal-force (32.17 pdl) Pound force per cubic foot

Pound per cubic foot Pound per cubic inch Pound foot squared

Ounce per gallon (US) Ounce per gallon (UK)

Anglo-Saxon notations Name Short hundredweight [100 lb (US)] Barrel (160 lb) Short ton [2000 lb (US)] (Long) ton (2240 lb) Grain per cubic foot

×1.016×10−3 ×2.288×10−3 ton gr/ft3

lbf/ft3

pdl lbf

lb•ft2

×157.1

×0.1383 ×4.448

×42.141× 10−3

×16.02 ×27.68×103

×6.236

×7.490

×72.576 ×907 sh ton

oz/gall (US) (oz/ gall) (UK) lb/ft3 lb/in3

Conversion to IS ×45.36

Symbol sh cwt

xii Abbreviation Group

Newton per square meter

Pascal second Newton second per square meter Square meter per second

σ

τ

η

v

E

Normal stress

Shear stress

Viscosity

Kinematic viscosity

Energy

British thermal unit Cubic foot atmosphere Horsepower hour (British)

Ws

Square inch per second Square foot per second Foot poundal

Poundal second per square foot Pound-force second per square foot

Square inch

Poundal per square foot Pound-force per square foot

Foot pound-force

×10−7

×10−4

×0.0981× 105 ×0.9807× 105 ×105 ×1.0135× 105 ×10−1

×133.322

nm

erg

St

P

bar atm

at

mH2O

mmHg

Newton meter Watt second

erg

Stokes

Technical atmosphere Bar Normal atmosphere Poise

Millimeter of mercury Meter of water

Pound-force foot

Foot poundal

J

m2 s−1

N s m−2

Pa s

N m−2

Pa

Nm

Joule

Pascal

M T P

Bending moment Torque Pressure

Newton meter

M

Moment of force

×2.685×106 hp•h

(continued)

×1.055×103 ×2.869×103 Btu ft3 •atm

ft•lbf

×42.141× 10−3 ×1.356

×92.9×10−3 ft2/s ft•pdl

×0.645×10−3

×47.88

×1.488

in2/s

pdl•s/ ft2 lbf.s/ft2

p.s.i.

lbf/ft2

×47.88

×1.488

pdl/ft2

lbf•ft

×42.141× 10−3 ×1.356

ft•pdl

Abbreviation Group xiii

Joule per meter second Kelvin Watt per meter Kelvin

λ

C

Thermal Conductivity

Specific heat capacity

Joule per kilogram Kelvin

Joule

Q

Joule per second Newton meter per second

ISa notation Name Watt

Hertz

Symbol P

n

Frequency Heat Quantity of heat

Magnitude Name Power

Table 1 (continued)

J kg−1 K−1

W m−1 k−1

J m−1 s−1 k−1

J

HZ n s−1

N m s−1

J s−1

Symbol W

Calorie Kilocalorie

Other notations Name Metric horse power DIN horse power (vehicles) Chevaux (fiscal vehicle form)

cal Cal

–

DIN hp

Symbol Mhp

×4.1868 ×4186.8

Approx.×98

Conversion to IS ×735.489

British thermal unit inch per hour square foot degree Fahrenheit British thermal unit per hour foot degree Fahrenheit British thermal unit per second foot degree Fahrenheit British thermal unit per pound degree Fahrenheit

British thermal unit

Anglo-Saxon notations Name Foot poundal per second British thermal unit per hour Foot pound-force per second Horsepower (UK) (vehicles) Horse power SAE (US) vehicles

Approx.×660

SAE HP

×4.187×103

×6.23×103 Btu/lb• degF Btu/lb• degF

×1.731

×0.1442

Btu/h• ft degF

Btu•in/ h• ft2 degF

×1.055×103

×724.7

HP

Btu

×1.356

Conversion to IS ×42.141× 10−3 ×293.1×10−3

ft•lbf/s

Btu/h

Symbol ft/pdl/s

xiv Abbreviation Group

P

i e r G σ, EC

Electricity Current Potential Resistance Conductance Conductivity

Power Charge Capacity Composite units Mass per area

Kelvin

T

Kilogram per square meter

Ampère Volt Ohm Siemens Siemens per meter Watt Coulomb Farad

Joule per kilogram

I

Specific latent heat of evaporation Temperature

×10−2 ×10−1

q ha−1 t ha−1

(Long) ton per acre

Barley Maize Wheat Hundredweight per acre Short ton per acre

Bushel per acre:

Pound per acre

×10−4

kg m−2

kg ha−1

See IS notation See IS notation See IS notation

5/9 (degF− 32)+273.15

degF

W V A J s−1 CAs F C V−11

Add 273.15

°C

See IS notation See IS notation See IS notation mho mho per centimeter

Kilogram per hectare Quintal per hectare Ton per hectare

Degree centigrade Degree Fahrenheit

British thermal unit per pound

A V WA−1 Ω V A−1 S Ω−1 A V−1 S m−1

K

J kg−1

cwt/ acre sh ton/ acre ton/acre

lb/acre

mho mho cm−1

Btu/lb

(continued)

×25.11×10−2

×22.41×10−2

×0.538×10−2 ×0.627×10−2 ×0.672×10−2 ×1.255×10−2

×0.1121× 10−3

×1 ×0.01

×2.326×103

Abbreviation Group xv

Distance per area per unit of time= depth per unit of time

Magnitude Name Area×depth (= volume)

Symbol

Table 1 (continued)

ISa notation Name Square meter× meter (= cubic metre) Meter per second

m s−1

Symbol m2 ×m (=m3)

Millimeter per day Millimeter per hour Liter per second per hectare (8.64 mm day−1)

Other notations Name Hectare millimeter (10 cubic meters) ×11.6×10−9 ×27.8×10−8 ×10−7

mm h−1 1 s−1 ha−1

Conversion to IS ×10

mm day−1

Symbol ha mm

Acre per cusec (duty =1) (605 mm day−1) (69.97 l s−1 ha−1)

Anglo-Saxon notations Name Acre inch Acre foot Square mile inch

acre/ cusduty

Symbol acre in acre ft mile2 in

×6.997×10−6

Conversion to IS ×102.7938 ×1233.5256 ×65,785.6980

xvi Abbreviation Group

Abbreviation Group

xvii

Prefixes Denoting Decimal Multiples or Submultiples3 The following prefixes, with significance, name, and symbol as shown below, are used to denote decimal multiples or submultiples of (metric) units. These prefixes, developed in conjunction with the metric system, and are now authorized as “IS prefixes” (International System of Units). To indicate multiples tera 1012 109 giga 106 mega 103 kilo 102 hecto 10 deca

T G M k h da

To indicate submultiples 101 deci 102 centi 103 milli 106 micro 109 nano 1012 pico 1015 femto 1018 atto

d c m μ n p f a

Meaning of Million, Billion, Trillion, etc.4 Regarding the meaning of million, billion, etc., the convention shown in the table below accords with the decision of the 9th General Conference of Weights and Measures, Paris 1948, and is in use in Europe. Term Million Billion Trillion Quadrillion

Significance Thousand  thousand Million  million Million  billion Million  trillion

Corresponding decimal factor 106 1012 1018 1024

Reservoirs Carbon Exchanges5 1 teragram (Tg) = 1012 g or 106 t 1 petagram (Pg) = 1015 g or 109 t 1 gigaton (Gt) = 109 t or 1 Pg 1 Pg = 1 Gt 3

Source: British Standards Institution (1974) Conversion factors and tables. London Ídem 5 Source: FAO (1995) Climate change, forests and forest management. An overview. Rome 4

xviii

Abbreviation Group

Spacing in Plantations Triangular and squared spacing of plants per hectare Planting distance m/cm 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7

No. of seedlings Triangle 72,169 46,188 32,075 23,565 18,042 14,256 11,547 9543 8019 6833 5891 5132 4511 3996 3564 3199 2887 2618 2386 2183 2005 1848 1708 1584

Square 62,500 40,000 27,778 20,408 15,625 12,346 10,000 8265 6944 5917 5102 4444 3906 3460 3086 2770 2500 2268 2066 1890 1736 1600 1479 1372

Planting distance m/cm 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.2

No. of seedlings Triangle 1473 1373 1238 1202 1128 1060 999 943 891 843 800 759 722 687 655 625 596 570 546 523 501 48 462 427

Square 1276 1189 1111 1041 977 918 865 816 772 730 693 657 625 595 567 541 517 494 473 453 434 417 400 370

Planting distance m/cm 5.4 5.5 5.6 5.8 6.0 6.2 6.4 6.5 6.6 6.8 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0 13.0 14.0 15.0

No. of seedlings Triangle Square 396 343 382 331 368 319 343 297 321 278 300 260 282 244 273 237 265 230 250 216 236 204 205 178 180 156 160 138 143 123 128 111 115 100 105 91 95 83 87 76 80 69 68 59 58 51 51 44

Areas and Volumes Notation: a, b, c, d denote lengths, A denotes area, C denotes circumference, D denotes diameter, R denotes radius, V denotes volume, h denotes altitude, and α denotes central angle in radians.

Abbreviation Group

xix

Triangles Right triangle

c b

a

1 ab 2 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi p c ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi a2 þ b2ffi 2  b2 a ¼ pcffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi b ¼ c 2  a2 A¼

Oblique triangle

c

a

h

b

1 bh 2

A¼ Equilateral triangle

a

h

a

a

1 1 pffiffiffi A ¼ ah ¼ a2 3 2 4 1 pffiffiffi h¼ a 3 2

xx

Abbreviation Group

Four-Sided Figures Square

d

a

a ¼ a2 ; d ¼ a

pffiffiffi 2

Rectangle

d

b

a

A ¼ ab; d ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi a2 þ b2

Parallelogram (opposite sides parallel) d1

b

d2 h

a a

A ¼ ah ab sin α p¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi 2 þ b2  2ab cos α d1 ¼ paffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi d2 ¼ a2 þ b2 þ 2ab cos α

Abbreviation Group

xxi

Trapezoid (one pair of opposite sides parallel) b

h

a

1 A ¼ h ða þ bÞ 2

Circle, Ellipse, and Parabola

r

d

C ¼ πd ¼ 2π r, where r ¼ d=2 1 AðcircleÞ ¼ πr 2 ¼ π d 2 4

b a

A ¼ π ab PerimeterðsÞ  nð1:5ða þ bÞ  abÞ Parabola

xxii

Abbreviation Group

l1 d

d1

l

ffi l2 1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4d þ Length of arcðsÞ 16d2 þ l2 þ • ln 2 8d " #     2 2d 2 2 2d 4  þ ... ¼l 1þ 3 l 5 l

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi! 16d2 þ l2 l

d ðl2  l1 Þ l2 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi d1  d2 Width of segment ðl1 Þ ¼ l d

Height of segment ðd1 Þ ¼

Three-Dimensional Figures Cube

d

a

pffiffiffi V ¼ a3 ; d ¼ a 3 Total surface ¼ 6a2 Rectangular parallelepiped

Abbreviation Group

xxiii

d

c

b

a

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi V ¼ abc; d ¼ a2 þ b2 þ c2 Total surface ¼ 2ðab þ bc þ caÞ Prism or cylinder

e

h

e

V ¼ ðarea of baseÞ  length h Pyramid or cone

S

h

S

1 ðarea of baseÞ  length h 3   1 perimeter of base  ðslant heightÞ Lateral area or regular figure ¼ 2 V¼

Frustum of pyramid or cone

xxiv

Abbreviation Group

S

S

h

 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi V ¼ 13 A1 þ A2 þ A1  A2 h, where A1 and A2 are areas of bases, and h is altitude. Lateral area of regular figure = ½ (sum of perimeters of bases)  (slant height) Sphere

AðsphereÞ ¼ 4πr 2 ¼ πd2 4 1 V ðsphereÞ ¼ πr 3 ¼ πd 3 3 6 Paraboloidal segment h

d r2

r1

1 V ðsegment of one baseÞ ¼ πr 1 2 h 2  1  V ðsegment of two baseÞ ¼ π r 1 2 þ r 2 2 d 2 Neiloidal segment

Abbreviation Group

xxv

r2

r1

r3

d/2

d/2

h

 π 2 r 1 þ 4r 2 3 h 6  π 2 ¼ r 1 þ 4r 2 2 d 6

V ðsegment of one baseÞ ¼ V ðsegment of one baseÞ

The Climatic Gradients and Associated Vegetation Types (Jordan 1993) See Table 2.

Amount of annual insolation of ground (kcal/ cm2) Mean annual temperature Tm ( C) Annual variation ( C) Diurnal variation ( C) Wind

ca. 160 25

15–20

20

High velocities during summer (typhoon, hurricane, cyclone), low during dry season, strong effect of tropical convergence zone. Local frontal storms toward end to dry season (habub in Africa)

South-East Asia 140–160 Congo basin 120–130 Amazon basin 100–120 28

3

9

Predominance of tropical low pressure trough, low velocities except in conventional squalls and local tornadoes Seasonally dry tropic (harmattan) or moist tropic air (trades, monsoon), velocities moderate except during passage of tropical cyclones

20

30

ca. 180 21–32

Predominance of tropic high pressure cell, average wind speeds low to moderate, occasionally high velocities in advective storm

30

35

ca. 200 20–33

As before, dry and hot storms more frequent

ca. 220 Extreme variation

As before, dust storms common

>220 Extreme variation

Table 2 The climatic gradient from the climatic equator toward higher latitudes and the corresponding gradients of plant habitus and of the main plant formations. The habitus is expressed as the type of climatic adaptation (above the line) and the corresponding morphological adaptation type (below the line). The formation refers to formation group and formation levels of ecological-structural classification. The climatic data refer to tropical lowlands below 300 m altitude (Br€unig 1972; UNESCO 1978)

xxvi Abbreviation Group

Growing period (months)

Climate type

Precipitation (mm/a) mean, min. or max. Distribution of precipitation Location

2 or 4 seasons, 3–5 dry months Subequatorial to outer tropics with influence of trades, monsoons, and monsoon like alternating winds Tropical humid (moist), isotherm, seasonal with predominantly summer rainfall 7–10

Even, annual variation 11–12

Equatorial belt and areas with constant moist air-masses outside this belt

1300–3000; min = 25 (Tm + 12)

2000; min = 50 (Tm + 12)

2

60/80 wet, 20/50 dry season

1–2

Tropical arid

Outer tropical to subtropical belt of descending air-masses

11 dry months

100–00; min = Tm + 70

>1

Usually 50 %

Bare

(continued)

Tropical arid

Outer tropical to subtropical belt of descending air-masses

12 dry months

8

Abbreviation Group xxvii

Edaphic climax formations

Amount of annual insolation of ground (kcal/ cm2) Characteristic habitus of climatic climax formation Main climatic climax formations

Super humid to humid, ombrophilous, evergreen tropical forest and semievergreen wet forest Littoral forest Mangrove forest and woodland Freshwater swamp forest and grassland Peat swamp forest Riparian forest Single-dominant forests on certain soils Sclerophyllous forest

South-East Asia 140–160 Congo basin 120–130 Amazon basin 100–120 Megathermhydrophilous Hydromorphicmesomorphic

Table 2 (continued)

Littoral forest Mangrove forest and woodland (less luxuriant) Freshwater swamp forest and grassland Riparian forest (often relic gallery) Evergreen forest (on moister welldrained soils) Sclerophyllous forest (sandy terraces and skeletal soils)

Humid to subhumid, semi-deciduous and deciduous tropical forest

ca. 160 Megathermtropophytic Tropomorph

Littoral woodland Mangrove forest to scrub Riparian fringing forest and grassland Sclerophytic thorn woodland Deciduous moist forest

Subhumid to semiarid, deciduous tropical forest

ca. 180 Megathermtrophophytic

Littoral scrub Mangrove woodland to scrub Riparian fringing woodland and grassland Xerophytic semidesert

ca. 200 Megathermtropophyticsclerophyticxerophytic Xeromorph Semi-arid deciduous tropical thorn woodland

Littoral scrub Mangrove scrub Riparian fringing scrub and grassland

Arid thorn scrub and semidesert

ca. 220

As in the arid zone, but rarer and poorer

Per arid desert

>220 Bare

xxviii Abbreviation Group

Submontane forest (locally rich in oaks and laurels, simple structure than climatic climax) Montane forest (moist) Alto-montane forest (moist) Alto-montane moss forest (wet, misty) Alto-montane woodland and scrub (moist) Subseral secondary forest Disclimax secondary forest Disclimax pine forest to pine savanna (mostly at higher altitude) Disclimax grassland (Imperata cylindrical usually common) Disclimax karstwoodland Disclimax sclerophytic savanna

Subseral secondary forest Disclimax secondary forest Disclimax savanna Disclimax pine forest, pine woodland or pine savanna (higher altitudes) Disclimax karstwoodland Disclimax sclerophytic or xerophytic savanna

Similar to the per humid zone, except for species composition, conifers increase in southern and northern hemisphere, bamboo species become more frequent in the northern hemisphere

Disclimax xerophytic savanna Disclimax thorn scrub Disclimax semidesert scrub

Similar to the humid zone, but relative effect of exposition and barrier-effect more pronounced Similar to dry zone, very strong and noticeable effect of elevation, frost occurs regularly even at lower altitudes, particularly in hollows, and creates local dwarf vegetation Generally more open and scrub-like. Barrier-effect very pronounced Disclimax xerophytic thorn shrub Disclimax semidesert Disclimax desert

As before

Jordan CF (1993) Chapter 3: Ecology of tropical forests. In: Pancel L (ed) Tropical forestry handbook, vol 1. Springer, Berlin

Degraded formations

Physiographic climax formations

Bare

As before

Abbreviation Group xxix

Acknowledgments

Several people have assisted in the preparation of this book. We especially thank Cecilia Vides for her constant and professional support in organizing and reviewing the text and in searching for practical solutions to finalize the first drafts of this book. We are indebted to Karin Bartsch, Christina Eckey, and Susanne Friedrichsen from Springer Publishers, Heidelberg, for not losing their patience with us. Despite numerous delays and missed deadlines, they never gave up on us and continued their support of this book. Finally, special thanks go to our families. Monica, Michaela, Miriam, and Damaris braved the adverse side effects of limited availability or mental absence of the editors, and they have generously and sympathetically supported our work.

xxxi

International Forest Organizations

Research International Union of Forest Research Organizations

Center for International Forestry Research

www.iufro.org

www.cifor.org

IUFRO is a global network for forest science cooperation. It unites more than 15,000 scientists in almost 700 member organizations in over 110 countries and is a member of ICSU. Scientists cooperate in IUFRO on a voluntary basis IUFRO headquarters are located in Vienna, Austria. The Center for International Forestry Research (CIFOR) is a nonprofit, scientific facility that conducts research on the most pressing challenges of forest and landscape management around the world. With its global, multidisciplinary approach, CIFOR aims to improve human well-being, protect the environment, and increase equity. It helps policymakers, practitioners, and communities make decisions based on solid science about how they use and manage their forests and landscapes Capacity building, collaboration, and partnerships are essential to finding and implementing innovative solutions to the challenges that the globe faces. CIFOR is a member of the CGIAR Consortium and led the CGIAR Research Program on Forests, Trees and Agroforestry CIFOR headquarters are in Bogor, Indonesia. CIFOR has offices in 8 countries across Asia, Latin America, and Africa and works in more than 30 countries. (continued) xxxiii

xxxiv

Research Tropenbos International

International Forest Organizations

www.tropenbos.org

TBI’s approaches recognize that good decisions depend on credible, relevant, and practical knowledge and information. Good decisions also require strong individual and organizational capacity and institutional governance across the forest sector, as well as strong networks and platforms for the sharing of knowledge. Knowledge must be targeted to the right audiences. Sound information that supports societal processes has the capacity to resolve and prevent conflicts and dilemmas TBI’s approach has several important components: TBI enables comprehensive, evidencebased multistakeholder dialogues that are supported by the generation of critical knowledge, development of local capacities, and strengthening of institutions TBI is an intermediary, linking policymakers, practitioners, and knowledge providers; formal, professional, and traditional knowledge; and northern and southern actors TBI is a knowledge broker, helping people make their own choices by developing scenarios and mapping their consequences TBI supports partnerships by engaging scientists, policymakers, the private sector, local communities, and civil society at the local, national, regional, and global level in collaborative action TBI achieves local impacts by creating lasting and dependable local policy and knowledge networks that are supported by effective country offices in a range of partner countries TBI supports communication by adapting knowledge into practical information for policymakers and forest practitioners TBI’s target groups Forest-dependent people Policymakers and regulators Practitioners, forest managers, forest owners, and forest users Nongovernmental organizations and civil society Researchers and educators (continued)

International Forest Organizations

Research Centro agronómico tropical de investigación y enseñanza (CATIE)

www.catie.ac.cr

International Institute for Environment and Development (IIED)

www.iied.org

Consultative Group on Agricultural Research (CIGAR)

www.cigar.org

xxxv

CATIE, as an international entity with a unique combination of science, graduate education, and innovation for development, has its bases well grounded and a clear action plan for creating professionals with a distinct perspective who can also contribute to the sustainable growth of the communities IIED builds bridges: between local and global, policy and practice, rich and poor, government and private sector, and across diverse interest groups. Its strength lies in the combination of research and action – generating robust evidence and know-how that is informed by a practical perspective acquired through hands-on research with grassroots partners. IIED has four research groups: Climate change Human settlements National resources Sustainable markets CGIAR is a global partnership that unites organizations engaged in research for a food-secure future We are no longer the “Consultative Group on International Agricultural Research.” In 2008, we underwent a major transformation; to reflect this and yet retain our roots, we are now known simply as CGIAR CGIAR research is dedicated to reducing rural poverty, increasing food security, improving human health and nutrition, and ensuring sustainable management of natural resources. It is carried out by 15 Centers that are members of the CGIAR Consortium, in close collaboration with hundreds of partners, including national and regional research institutes, civil society organizations, academia, development organizations, and the private sector The 15 research centers generate and disseminate knowledge, technologies, and policies for agricultural development through the CGIAR Research Programs. The CGIAR Fund provides reliable and predictable multiyear funding to enable research planning over the long term, (continued)

xxxvi

International Forest Organizations

Research

Food and Agriculture Organization of the United Nations

www.fao.org

Centre for Agricultural Research in Suriname (CELOS)

www.celos.sr.org

Forestry Research, Development and Innovation Agency (FORDA), Indonesia

http://www.fordamof.org

resource allocation based on agreed priorities, and the timely and predictable disbursement of funds. The multidonor trust fund finances research carried out by the Centers through the CGIAR Research Programs An intergovernmental organization, FAO, has 194 Member Nations, two associate members, and one member organization, the European Union. Its employees come from various cultural backgrounds and are experts in the multiple fields of activity FAO engages in. FAO’s staff capacity allows it to support improved governance inter alia; generate, develop, and adapt existing tools and guidelines; and provide targeted governance support as a resource to country- and regional-level FAO offices. Headquartered in Rome, Italy, FAO is present in over 130 countries CELOS fosters applied scientific research in the agricultural and forestry sector in Suriname and the region. In addition to conducting research, CELOS also makes its services and facilities available to university scientists and students, namely, those faculty members and students affiliated with the Department of Agriculture of the Faculty of Technology. Since the expertise of the various agricultural and forestry fields exists within the staff of CELOS and the Faculty of Technology, these institutions are able to adequately adapt their programs to target many agricultural and forestry subareas Fields of specialization are agriculture and animal science, forestry, wood technology, tissue culture, soil science, biodiversity, agronomy, agroforestry, GIS and remote sensing (NARENA), aquaculture, and fish ecology FORDA aims at improving governance and forestry science and technology expediency Confirming research and development–supporting elements Conducting research and development relevant and appropriate to the needs of users and be able to respond to the challenges and problems of the forestry (continued)

International Forest Organizations

xxxvii

Research

Forestry Research Institute of Ghana

http://csir-forig. org.gh

Institute of Tropical Forest Conservation

http://itfc.must.ac.ug

The Center for Tropical Forest Science – Forest Global Earth Observatories (CTFS-ForestGEO)

http://www. forestgeo.si.edu/

sector Improving the usefulness and applicability of the results of research and development Confirming research and development–supporting elements Forestry Research Institute of Ghana is one of the 13 institutes of the Council for Scientific and Industrial Research (CSIR). It is located at Fumesua near Kumasi in the Ashanti Region of Ghana. It started as a research unit within the Forestry Department in 1962. Forestry Research Institute of Ghana’s mission is to conduct forest and forest product research for social, economic, and environmental benefits of society ITFC is a postgraduate institute established in 1991 under Mbarara University of Science and Technology focused on research, training, and monitoring for conservation management. We are located just inside Bwindi Impenetrable National Park in South West Uganda, a world heritage site and the homeland of Uganda’s mountain gorillas The Center for Tropical Forest Science – Forest Global Earth Observatories (CTFS-ForestGEO) is a global network of forest research plots and scientists dedicated to the study of tropical and temperate forest function and diversity. The multiinstitutional network comprises over 60 forest research plots across the Americas, Africa, Asia, and Europe, with a strong focus on tropical regions. CTFSForestGEO monitors the growth and survival of approximately six million trees and 10,000 species CTFS-ForestGEO conducts long-term, large-scale research on forests around the world to: Increase scientific understanding of forest ecosystems Guide sustainable forest management and natural-resource policies Monitor the impacts of global climate change Build capacity in forest science (continued)

xxxviii

Research European Tropical Forest Research Network

International Forest Organizations

http://www.etfrn. org/

The Oxford Centre for Tropical Forests (OCTF)

http://www. tropicalforests.ox. ac.uk/

Forest Research Institute Malaysia (FRIM)

http://www.frim. gov.my

Established in 1991, the European Tropical Forest Research Network (ETFRN) aims to ensure that European research contributes to conservation and sustainable use of forest and tree resources in tropical and subtropical countries The Oxford Centre for Tropical Forests (OCTF) is a network of Oxford University departments and neighboring NGOs, consultancies, and businesses in the Oxford area. OCTF facilitates collaborative research and communication among its members on issues related to forest governance, management, and conservation and also serves as a platform for broader collaboration between Oxford-area institutions and the global forest community Forest Research Institute Malaysia (FRIM) is one of the leading institutions in tropical forestry research in the world. Founded in 1929, the former Forest Research Institute became a full-fledged statutory body, governed by the Malaysian Forestry Research and Development Board (MFRDB) under the Ministry of Primary Industries, in 1985. Main research areas are: Forestry and environment Forestry biotechnology Forest products Forest biodiversity Economic and strategic analysis

Introduction to Tropical Forestry and to This Handbook

Tropical forests are one of the oldest biomes on our planet. Over time, they have developed fascinating diversity, and not only in terms of species numbers. They are an essential component of the global climate system, and they furnish habitats for a myriad of plant and animal species, contribute to the livelihood of indigenous peoples, provide goods and services, and include awe-inspiring locations that provoke admiration and amazement among visitors to the forest. They offer enough potential for future generations of scientists to carry out scientific work, as most of the biogeochemical functions and relationships that underpin tropical forest ecosystems are still beyond our knowledge and control. The current lack of knowledge is of particular concern when tropical forests are considered as a common good with unrestricted access for utilization and conversion to other land uses. The world’s forests are in an alarming state. According to Mitchard (2015), the results of a country-by-country analysis of forest loss are of great concern: In the period between 2000 and 2012, 13.4 % of the forest cover of Malaysia has been lost, 7.9 % in Indonesia, and 4.0 % in Brazil. Mitchard goes on to explain that “these figures are not a proportion of forest loss, but a proportion of the whole country that has been reported as undergoing forest loss over that period”! A similarly pessimistic impression arises from FAO’s State of the World’s Forests report (FAO 2014), whose illusive Key Messages1 seem to suggest that the forest sector in the tropics will have a long road ahead. Unhampered population growth, poverty, the increasing demand for timber, the extension of agricultural land use driven by the demand for food and energy, and the reckless search for economic profits put tropical forests in a weak position in land-use conflicts.

1

Key Messages State of the World’s Forests (FAO 2014) – To measure the socioeconomic benefits from forests, data collection must focus on people, not only trees. – Forest policies must explicitly address forests’ role in providing food, energy, and shelter. Recognition of the value of forest services, such as erosion protection and pollination, is essential to sound decision-making. – To meet rising and changing demands, sustainable forest management must include more efficient production. – Providing people with access to forest resources and markets is a powerful way to enhance socioeconomic benefits. xxxix

xl

Introduction to Tropical Forestry and to This Handbook 442 Mha 327 Mha 247 Mha 190 Mha 70 Mha 120 Mha

2005 Non-tropical

255 Mha 166 Mha

108 Mha 139 Mha

2010

187 Mha

161 Mha

2015

2020

Tropical

Fig. 1 Estimated potential development of forest plantations (Tomaselli 2007)

In light of these doomsday scenarios, can we as foresters respond to these threats and challenges? What are our prospects to halt or even reverse these developments, as the FAO (2006) asserts that we face a rate of forest area destruction of over 13 million hectares per year? We undertook the work on the current Handbook to raise the understanding of tropical forest ecosystems and to pass on long-term experience in the protection and sustainable management of tropical forests. There are nearly 1.65 billion hectares of natural tropical forests, of which 41 % are rain forests, 33 % are moist deciduous forests, 10 % are dry deciduous forests, 12 % are hill and mountain forests, and 4 % can be classified as other tropical forest formations. The diversity of tropical forests requires a similar diversity of approaches to halt their destruction and to safeguard their role as providers of multiple ecosystem services and functions. Utilizing tropical forests without understanding their underlying functional relationships will open a Pandora’s box. The Tropical Forestry Handbook aims to narrow the gulf between forest protectionists and forest users by providing scientifically sound evidence of unique tropical forest ecosystems and proven experience in their sustainable utilization. The land-use cascade in the tropics often runs from natural forests to degraded forests used for agriculture or pasture, and finally to fully degraded sites. According to Zomer et al. (2008), the areas in the tropics with potential for afforestation amount to 750 million hectares (Fig. 1). Thus, even after the destruction of forests, there is a glimmer of hope. There are still potential areas that may be planted and have a significant effect in providing forest goods and services by simultaneously protecting remaining natural forests from depletion and combating global warming to offset the annual global increase in CO2 (Paul et al. 2009). But are we professionally prepared to manage these resources adequately? Tropical forestry has evolved over the last two millennia in the Americas and Asia, but written traces of this progress are almost inexistent, mainly due to the very same tropical environmental conditions which allow an exuberant environment (Dawkins and Philip 1998). The editors of this Handbook are conscious of the fact that forestry in the tropics did not start with the arrival of European-style forestry, but we were reluctant to

Introduction to Tropical Forestry and to This Handbook

xli

equate a few documented historical data on forestry activities with a more comprehensive historical basis for the tropics. We respect and admire the accrued experiences of former tropical forest managers and the invaluable interdisciplinary knowledge on cautious and thoughtful forest utilization led by indigenous peoples. Today’s tropical foresters do not start from scratch; the scientific basis of modern tropical forestry was established in the nineteenth century with the start of systematic forest management documented and introduced by Sir Dietrich Brandis (1824–1906). Brandis also established forest research in India, among other areas. In the long line of excellent professionals dealing with the subject of tropical forestry, one has to mention Andre Aubreville (1897–1982), a French forestry expert whose work on botany and silviculture has influenced the orientation of tropical forestry today. Carlos Flinta and later Julian Evans have also contributed substantially to the subject of plantation forestry in the tropics. Hans Lamprecht (1919–2012) systematized tropical silviculture and contributed to the understanding of stand dynamics. Lamberto Golfari helped to understand the precise precipitation requirements for the selection of forestry species for plantations, and Paul Westmacott Richards with his outstanding work on tropical forest ecology contributed essentially to the understanding of functioning of tropical ecosystems. The systematic work of Heinz Ellenberg (1913–1997) allowed an ecological classification of forest types worldwide, and P.K. Ramachandran Nair helped to put agroforestry on a firm scientific ground. In addition to these outstanding professionals, tropical forestry was nourished by the establishment of institutions which contributed to a fuller understanding of the subject. FAO’s Forestry Department, ITTO, ICRAF, CIFOR, IUFRO, and CATIE have all channeled international research on tropical forestry and committed themselves to publishing all of the relevant advances in an understandable and professional way. Is a handbook format still topic driven? How does the present Handbook on Tropical Forestry fit into the vast landscape of information that is now available with a click of our fingertips? A handbook should provide proven methodologies and concepts to resolve specific issues. Furthermore, our philosophy was to invite authors whose experience and/or publications ensure that state-of-the-art knowledge is presented. We, the editors, are convinced that a periodic and systematic review of a subject helps to galvanize the existing advances in science and that this can be achieved by a handbook. The scope of this Tropical Forestry Handbook is broader than that of the first edition from 1993: It consciously includes the aspect of a globalized forest sector of which tropical forestry is a part. Therefore, tropical resources and international processes are dealt with separately and in detail. Basic forest sciences are unimaginable without the subjects of geology and soils, climatology, and forest ecology. Since the last edition, the World Reference Base for Soil Resources has partially reclassified the existing soil classes, which as a task had to be completed. Climatology and forest ecology received a complete overhaul both in terms of content as well as orientation. The Handbook would have not been complete without the generally undervalued subject on species identification. Resource monitoring and

xlii

Introduction to Tropical Forestry and to This Handbook

assessment is covered by remote sensing and forest inventory; these chapters are not new, but they received a completely new orientation due to the considerable advances within the last 20 years. Silviculture is and will be a science vital to the successful management of tropical forest resources. The chapters on the following topics have been included in this edition: genetics, forest seeds, nursery management, plant nutrition, species selection, plantation forestry, tropical silviculture, and mangrove management. This set of chapters provides enough material to be a standalone publication, but in the context of the Handbook, each chapter offers an important complement to assure that all technical aspects are covered for forest management. Resource planning combines all of the subjects that are necessary to ensure consideration for not only forest-centered views but also a wider spectrum of subjects as well that are essential for the management of forests in the tropics: forestry management planning, land evaluation for forestry, watershed management, range management, agroforestry, community forestry, wildlife management, and nature conservation. The engineering chapters on forestry road construction and harvesting ensure that the operations related to forestry management planning are complete and supplement the instructions. The subjects of pest and fire management are of increasing importance given the effects of climate change, in which alterations in air humidity affect the population dynamics of pests and create optimal conditions for forest fires. The increasing complexity of the framework conditions of a modern society require ongoing monitoring of the instruments and mechanisms that govern the interaction of markets, social demands, and the necessary safeguards imposed by the society. For this purpose, the Handbook has established three subject areas. The first deals with economics and management: Non-timber Forestry Products, Grading Forestry Products, Forestry Economics, Compensation Payments, Certification, and Forestry Project Management. The second one refers to human resource management, discussing Ergonomics, Forest Target Groups/Indigenous People, and Forestry Extension. The third and last subject area deals with topics related to governance: Forestry Policy, Legislation, the Rights of Target Groups, and Forestry Law Enforcement. Research in tropical forestry presents a summary from all previous chapters on the needs and orientation for future research to be carried out for the successful development of forestry resource management. The present book was written by experts dedicated to the fascinating world of tropical forests. It intends to contribute to the understanding of tropical forests for the protection of these unique and widely unknown ecosystems and to present guidelines for their sustainable and careful utilization. With this Handbook and its chapters, we have the knowledge to allow us to address difficult questions in natural forest and plantation management in the tropics. But how much of a challenge are the framework conditions, and where are the true limitations for possible action? Our goal as foresters with an eye toward ongoing climate change is to offer alternatives for the production of raw materials and to provide all of the possible potential functions of our tropical forests. We believe that our Handbook is a step in the right direction to achieve this. A taste of possible targets, challenges, and limitations is given in Table 1.

Introduction to Tropical Forestry and to This Handbook

xliii

Table 1 A sample of concrete targets, challenges, and limitations for tropical forest management to contribute to a better world Targets Certification of sustainable management of all relevant natural tropical forests Manage at least 600 million ha of tropical forests sustainably

Principal challenges Not enough political pressure for certification (e.g., FSC)

Close to nature forest plantations on all potential areas in the tropics Increase the plantation area by 2020 to at least 220 million ha

Many tropical countries are missing the necessary rule of law to establish plantation enterprises

Village/community plantation on all relevant communal land in the tropics

Population pressure impedes successful establishment of village/community plantations

All urban spaces in the tropics have at least 30 % green areas planted with trees

Urban area planning surrenders to population pressure and business planning interests

Limitations Production cycle only for a limited number of native species/species known Growth and yield data only for a limited number of species known Climatic plasticity of most tropical species unknown From an estimated total of approximately 51,000 tree species, only 215 species account for 93 % of all tree plantations in the tropics. This selection denotes the use of just 0.42 % of the available tree gene pool of the tropics (Pancel 2015) Site specificity only known for approximately 60 species Mixture patterns of plantation species almost unknown Attractiveness of forest products (growth and value) cannot compete with cash crops, land hunger, and criminal activities Only approximately 70 species are utilized in 90 % of all urban greening projects in the tropics Air and soil pollution resistance to only approximately 20 % of tropical species known

It is our desire to make sure that future generations can still experience the fascination with tropical forests that we do today.

References Dawkins HC, Philip MS (1998) Tropical moist forest silviculture and management. CAB International, Wallingford FAO (2006) Global forest resources assessment 2005 – progress towards sustainable forest management. FAO Forestry Paper 147. FAO, Rome

xliv

Introduction to Tropical Forestry and to This Handbook

FAO (2014) State of the world’s forests; enhancing the socioeconomic benefits from forests. FAO, Rome Hansen MC, Potapov PV, Moore R, Hancher M, Turubanova SA, Tyukavina A, Thau D et al (2013) High-resolution global maps of 21st-century forest cover change. Science 342:850–853 Mitchard E (2015) The world seen in deforestation. http://www.globalforestwatch. org/map/4/-6.48/-69.49/ALL/grayscale/loss?begin=2001-01-01&end=2014-01-01& threshold=30. Accessed 23 Jun 2015 Paul C, Weber M, Mosandl R (2009) Kohlenstoffbindung junger Aufforstungsflächen. Karl Gayer Institut, Lehrstuhl f€ur Waldbau der Technischen Universität M€ unchen, Freising Tomaselli I (2007) Global wood and products flows – trends and perspectives. FAO/stcp, Shanghai Zomer RJ, Trabucco A, Bossio DA, Verchot LV (2008) Climate change mitigation: a spatial analysis of global land suitability for clean development mechanism afforestation and reforestation. Agric Ecosyst Environ 67–80

Contents

Volume 1 Part I

The Global Forestry Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

Tropical Forest Resources: Facts and Tables . . . . . . . . . . . . . . . . . . . . . Jutta Poker and Kenneth MacDicken

3

International Processes: Framework Conditions for Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thomas Schneider and Prem Raj Neupane

47

Part II

Basic Sciences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

91

Geology and Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wolfgang Zech

93

Climate Aspects of the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Richter

285

Classifications of Climates in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . Michael Richter

293

The Atmospheric Circulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thorsten Peters and Michael Richter

303

Radiation and Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thorsten Peters

333

Temperatures in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Richter

343

..................................

363

............................

391

Precipitation in the Tropics Michael Richter

Water Balance in Tropical Regions Thorsten Peters

xlv

xlvi

Contents

Climatic Types of Water Balances in the Tropics . . . . . . . . . . . . . . . . . . Thorsten Peters

405

Microclimate in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Richter

413

Climate Change in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thorsten Peters

429

Design of Data Collection Related to the Climate in the Tropics . . . . . . Thorsten Peters

445

Tree Species Identification in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . Michelle Szejner and Patricio Emanuelli

451

Tropical Forest Ecology in the Anthropocene . . . . . . . . . . . . . . . . . . . . Richard T. Corlett

471

Classifying Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Richard T. Corlett

479

Tropical Forest Ecosystem Ecology: Water, Energy, Carbon, and Nutrients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Richard T. Corlett

491

..................

503

Applied Ecology of Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . Richard T. Corlett

511

Scope and Extent of Wood Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jörg Fromm, Gerald Koch, and Silke Lautner

519

Part III

543

Ecological Roles of Animals in Tropical Forests Richard T. Corlett

Resource Monitoring and Assessment . . . . . . . . . . . . . . . .

Fundamentals and Applications of Remote Sensing in Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thomas Baldauf and Abner Josue Jimenez Galo

545

Acquisition, Characteristics, and Preprocessing of Passive Remote Sensing Images in Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . Abner Josue Jimenez Galo

571

Interpretation and Processing of Passive Sensor Images in Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Abner Josue Jimenez Galo

607

.........

635

Image Processing of Radar and Lidar in Tropical Forestry Thomas Baldauf and Mariano Garcia

Contents

xlvii

Monitoring of Tropical Forest Cover with Remote Sensing . . . . . . . . . . Abner Josue Jimenez Galo

663

Measurements and Assessments on Field Plots . . . . . . . . . . . . . . . . . . . Michael Köhl and Marco Marchetti

687

...................

749

Sampling in Forest Inventories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Köhl and Steen Magnussen

777

Objectives and Planning of Forest Inventories Michael Köhl and Marco Marchetti

Measurement, Reporting, and Verifications Systems in Forest Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Daniel Plugge, Daniel K€ubler, Prem Raj Neupane, Konstantin Olschofsky, and Laura Prill

839

Volume 2 Part IV

Silviculture

......................................

883

Genetics and Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bernd Degen and Alexandre Magno Sebbenn

885

Genetics and Forest Seed Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lars Schmidt

921

Forest Seed Collection, Processing, and Testing . . . . . . . . . . . . . . . . . . . Lars Schmidt

959

Trade and Transfer of Tree Seed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lars Schmidt

995

Tropical Nursery Concepts and Practices . . . . . . . . . . . . . . . . . . . . . . . 1005 Diane L. Haase, R. Kasten Dumroese, Kim M. Wilkinson, and Thomas D. Landis Planning and Managing a Tropical Nursery . . . . . . . . . . . . . . . . . . . . . 1043 Kim M. Wilkinson, Thomas D. Landis, Diane L. Haase, and R. Kasten Dumroese Collecting, Processing, and Treating Propagules for Seed and Vegetative Propagation in Nurseries . . . . . . . . . . . . . . . . . . . . . . . . . . . 1079 R. Kasten Dumroese, Diane L. Haase, Kim M. Wilkinson, and Thomas D. Landis Plant Nutrition in Tropical Forestry Alfredo Alvarado

. . . . . . . . . . . . . . . . . . . . . . . . . . . 1113

Species Selection in Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . 1203 Laslo Pancel

xlviii

Contents

Species Files in Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1221 Laslo Pancel Basic Outline of Tree Plantations in the Tropics . . . . . . . . . . . . . . . . . . 1441 Laslo Pancel Reforestation Incentives Systems for Tree Plantations in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1531 Laslo Pancel Mixed Tree Plantations in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . 1549 Laslo Pancel . . . . . . . . . . . . . . 1561

Forest Restoration and Rehabilitation in the Tropics Laslo Pancel

21st Century Viewpoint on Tropical Silviculture . . . . . . . . . . . . . . . . . . 1605 Bryan Finegan Technical Orientation of Silviculture in the Tropics Laslo Pancel

. . . . . . . . . . . . . . . 1639

Mangroves: Unusual Forests at the Seas’ Edge . . . . . . . . . . . . . . . . . . . 1693 Norman C. Duke and Klaus Schmitt Mangrove Management, Assessment, and Monitoring Klaus Schmitt and Norman C. Duke

. . . . . . . . . . . . . 1725

Volume 3 Part V

Forest Resources Planning . . . . . . . . . . . . . . . . . . . . . . . . . .

Forest Management Thomas Knoke

1761

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1763

Tropical Forest Management Planning . . . . . . . . . . . . . . . . . . . . . . . . . 1793 Laslo Pancel, G€ unther Haase, Werner Schindele, and Michael Köhl Land Evaluation and Forestry Management . . . . . . . . . . . . . . . . . . . . . 1835 Anthony Young Introduction to Watershed Management . . . . . . . . . . . . . . . . . . . . . . . . 1869 Hosea M. Mwangi, Stefan Julich, and Karl-Heinz Feger Watershed Management Practices in the Tropics . . . . . . . . . . . . . . . . . 1897 Hosea M. Mwangi, Stefan Julich, and Karl-Heinz Feger Forest Hydrology in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1917 Stefan Julich, Hosea M. Mwangi, and Karl-Heinz Feger

Contents

xlix

Range Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1941 A. Swenne Agroforestry: Essential for Sustainable and Climate-Smart Land Use? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2013 Reinhold G. Muschler Community Forestry Carsten Schusser

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2117

Wildlife Management in the Tropics: An Overview . . . . . . . . . . . . . . . . 2145 Johannes Bauer How Environmental and Societal Changes Affect Wildlife in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2177 Johannes Bauer The Development of Wildlife Governance, Science, and Management Capacity in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2197 Johannes Bauer Modern Adverse Trends Which Affect the Wildlife Management Efforts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2223 Johannes Bauer Re-imagining Wildlife Management for the Tropics . . . . . . . . . . . . . . . 2239 Johannes Bauer Nature Conservation in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2255 Laslo Pancel Part VI

Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2299

Forest Road Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2301 John Sessions, Rudolf Heinrich, and Héctor Castaneda-Langlois Harvesting Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2363 Leif Nutto, Jorge R. Malinovski, Gustavo Pereira Castro, and Rafael A. Malinovski Machinery and Equipment in Harvesting . . . . . . . . . . . . . . . . . . . . . . . 2395 Gustavo Pereira Castro, Jorge R. Malinovski, Leif Nutto, and Ricardo A. Malinovski Harvesting Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2445 Gustavo Pereira Castro, Leif Nutto, Jorge R. Malinovski, and Ricardo A. Malinovski Harvesting Costing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2487 Rafael A. Malinovski, Jorge R. Malinovski, Leif Nutto, and Éllen C. Bianchi

l

Contents

Safety and Training in Harvesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2521 Ricardo A. Malinovski, Jorge R. Malinovski, Leif Nutto, and Nathan S. Sanches Pest Management in Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . 2561 Martin R. Speight and Stephen Woodward Insects and Other Animals in Tropical Forests . . . . . . . . . . . . . . . . . . . 2607 Martin R. Speight Fire Management in Tropical Forests Johann Georg Goldammer

. . . . . . . . . . . . . . . . . . . . . . . . . . 2659

Volume 4 Part VII

Economics/Management . . . . . . . . . . . . . . . . . . . . . . . . . . .

2711

From Lifelines to Livelihoods: Non-timber Forest Products into the 21st Century . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2713 Patricia Shanley, Alan R. Pierce, Sarah A. Laird, Citlalli López Binnq€ uist, and Manuel R. Guariguata The Principles of Wood Characteristic Formation Christoph Richter

. . . . . . . . . . . . . . . . 2761

Wood Characteristics Inherent in a Tree’s Natural Growth . . . . . . . . . 2785 Christoph Richter Biotically and Abiotically Induced Wood Characteristics; Cracks – Forms and Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2839 Christoph Richter Financial and Economic Evaluation Guidelines for International Forestry Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2875 Frederick Cubbage, Robert Davis, Gregory Frey, Diji Chandrasekharan Behr, and Erin Sills Bioeconomic Approaches to Sustainable Management of Natural Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2897 Thomas Holmes and Erin Sills Timber Production Cost and Profit Functions for Community Forests in Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2923 Frederick Cubbage, Robert Davis, Diana Rodríguez Paredes, Yoanna Kraus Elsin, Ramon Mollenhauer, and Gregory Frey Financial Analysis of Community-Based Forest Enterprises with the Green Value Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2949 Shoana Humphries and Thomas Holmes

Contents

li

Financial and Economic Analysis of Reduced Impact Logging . . . . . . . 2967 Thomas Holmes Identifying the Causes of Tropical Deforestation: Meta-analysis to Test and Develop Economic Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2987 Stibniati Atmadja and Erin Sills Concept of Compensation Payments and Ecosystems . . . . . . . . . . . . . . 3019 Julian Michel, Kay Kallweit, and Evy von Pfeil Compensation Payment Scheme Requisites and Financial Arrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3027 Julian Michel, Kay Kallweit, and Evy von Pfeil Payments for Ecosystem Services (PES) . . . . . . . . . . . . . . . . . . . . . . . . 3039 Julian Michel, Kay Kallweit, and Evy von Pfeil The Clean Development Mechanism (CDM) . . . . . . . . . . . . . . . . . . . . . 3057 Julian Michel, Kay Kallweit, and Evy von Pfeil Reducing Emissions from Deforestation and Forest Degradation (REDD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3065 Julian Michel, Kay Kallweit, and Evy von Pfeil Compensation Payments: Opportunities, Risks, and Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3093 Julian Michel, Kay Kallweit, and Evy von Pfeil Forestry Project Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3097 Steve Sepp and Stefan Mann Forest Market Policy Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3117 Frances Maplesden and Steven Johnson Forest Product Market Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3139 Frances Maplesden and Steven Johnson Forest Market Standards and Transport . . . . . . . . . . . . . . . . . . . . . . . . 3169 Frances Maplesden and Steven Johnson Forest Market Strategy Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3199 Frances Maplesden and Steven Johnson Part VIII

Human Resource

................................

3209

Ergonomics and Labor in Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3211 E. Apud, F. Meyer, J. Espinoza, E. Oñate, J. Freire, and F. Maureira Indigenous People and Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3311 Angelika Kandzior

lii

Contents

Application of a Participatory Approach to Forestry Extension Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3345 Angelika Kandzior and Esteban Rivas Forestry Extension as a Work Approach for Forestry Programs and Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3361 Angelika Kandzior and Esteban Rivas Part IX

Governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3387

Tropical Forest Policy and Legislation . . . . . . . . . . . . . . . . . . . . . . . . . . 3389 Arnoldo Contreras-Hermosilla Legal Recognition of Forest Rights of Indigenous Peoples and Local Communities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3449 Fernanda Almeida Evaluating Formal Recognition of Forest Rights of Indigenous Peoples and Local Communities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3485 Fernanda Almeida Forest Crime in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3525 J€urgen Blaser and Astrid Zabel Changing Trends of Forestry Research Demand . . . . . . . . . . . . . . . . . . 3559 Don Koo Lee, Mohammed N. Salleh, Wai Mun Ho, and Marilyn S. Combalicer Introduction to Forest Certification Schemes . . . . . . . . . . . . . . . . . . . . . 3571 Jörn Struwe and Thorsten Specht Research Areas in Tropical Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . 3593 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3603

About the Editors

Laslo Pancel Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), GIZ El Salvador Education: 1977: Diploma in Forestry at the Albert-Ludwigs University, Freiburg 1982: Doctorate at the Chair of World Forestry, University of Hamburg; Degree: Dr. rer. nat. Professional Experience: 2010: GIZ, Chief Technical Advisor REDD regional programs Central America 2002: GTZ, Vietnam, Chief Technical Advisor, GIZ Forestry Programs, Forest Admin. Reform, Natural Forest Management 1996: GTZ, Chile Chief Technical Advisor, Natural Forest Management 1990: GTZ, Amman Jordan, Project manager Forestry Project 1980: International Forestry Consultant: Africa, Near East, Latin America 1977: Latin America, Forestry Officer for Forestry Ecology, FAO

Michael Köhl Center for Wood Sciences, Institute of World Forestry, University of Hamburg, Hamburg, Germany Education: 1983: Diploma in Forestry at the Albert-Ludwigs University, Freiburg, Germany 1986: Doctorate in Forest Sciences, University of Freiburg, Germany; Degree: Dr. rer. nat. 1993: Habilitation, Swiss Federal Institute for Technology (ETH) Zürich, Switzerland Professional Experience: 2004: University of Hamburg, Professor for World Forestry 2004–2013: Institute Head, Institute for World Forestry, Federal Research Center for Forestry and Forest Product, Hamburg, Germany 1997–2004: Professor for Forest Biometrics and Computer Sciences, Technical University of Dresden, Tharandt, Germany liii

liv

About the Editors

1988–1997: Research Group Leader, Swiss Federal Institute for Forest, Snow and Landscape Research, National Forest Inventory, Birmensdorf, Switzerland 1987–1988: Biometrician, Pfizer Inc., Medical Research, Karlsruhe, Germany 1983–1987: Researcher, University of Freiburg, Department of Forest Biometrics, Freiburg, Germany

Contributors

Fernanda Almeida Almeida Dohrn Consulting Ltda, Berlin, Germany Alfredo Alvarado Centro Investigaciones Agronómicas, Universidad de Costa Rica, San Pedro Montes de Oca, Costa Rica E. Apud Unidad de Ergonomía, Facultad de Ciencias Biológicas, Universidad de Concepción, Concepción, Chile Stibniati Atmadja Forests and Livelihoods, Center for International Forestry Research, Bogor, Jawa Barat, Indonesia Thomas Baldauf European and International Forest Policy, Federal Ministry of Food and Agriculture, Bonn, Germany Formerly Institute for World Forestry, Johann Heinrich von Th€unen-Institute, Hamburg, Germany Johannes Bauer Australian Carbon Co-operative Ltd., Bathurst, NSW, Australia Diji Chandrasekharan Behr World Bank, Washington, DC, USA Éllen C. Bianchi Malinovski Florestal, Curitiba, Brazil Citlalli López Binnq€ uist Center for Latin American Studies, University of Florida, Gainesville, FL, USA J€ urgen Blaser School of Agricultural, Forest and Food Sciences, Bern University of Applied Sciences, Bern, Switzerland Héctor Castaneda-Langlois San Salvador, El Salvador Gustavo Pereira Castro Malinovski Florestal, Curitiba, Brazil Marilyn S. Combalicer Department of Forest Biological Sciences, College of Forestry and Natural Resources, University of the Philippines Los Baños, Laguna, Philippines Arnoldo Contreras-Hermosilla Labro, Rieti, Italy lv

lvi

Contributors

Richard T. Corlett Center for Integrative Conservation, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun, Yunnan, China Frederick Cubbage North Carolina State University, Raleigh, NC, USA Robert Davis World Bank, Latin America and the Caribbean Division, Washington, DC, USA Bernd Degen Th€unen Institute of Forest Genetics, Grosshansdorf, Germany Norman C. Duke TropWATER – Centre for Tropical Water and Aquatic Ecosystem Research, James Cook University, Townsville, QLD, Australia R. Kasten Dumroese USDA Forest Service, Rocky Mountain Research Station, Moscow, ID, USA Yoanna Kraus Elsin World Bank, Washington, DC, USA Patricio Emanuelli Sud-Austral Consulting SpA, Santiago, Chile J. Espinoza Unidad de Ergonomía, Facultad de Ciencias Biológicas, Universidad de Concepción, Concepción, Chile Karl-Heinz Feger Institute of Soil Science and Site Ecology, Technische Universität Dresden, Dresden, Germany Bryan Finegan Production and Conservation in Forests Programme, CATIE, Turrialba, Costa Rica J. Freire Unidad de Ergonomía, Facultad de Ciencias Biológicas, Universidad de Concepción, Concepción, Chile Gregory Frey USDA Forest Service, Washington, DC, USA Jörg Fromm Institute for Wood Biology, University of Hamburg, Hamburg, Germany Abner Josue Jimenez Galo Geographic Information System and Remote Sensing (GIS/RS), Deutsche Gesellschaft f€ur Internationale Zusammenarbeit (GIZ) GmbH, La Libertad, El Salvador, CA University of Alcalá (UAH), Madrid, Spain Faculty of Spatial Sciences, Universidad Nacional Autónoma de Honduras (UNAH), Tegucigalpa, Honduras Mariano Garcia Centre for Landscape and Climate Research, University of Leicester, Leicester, UK Johann Georg Goldammer Global Fire Monitoring Center (GFMC), c/o Freiburg University/United Nations University (UNU), Freiburg, Germany Manuel R. Guariguata CIFOR C/O Centro Internacional de la Papa (CIP), La Molina, Lima, Peru

Contributors

lvii

Diane L. Haase State and Private Forestry, USDA Forest Service, Portland, OR, USA G€ unther Haase University of Hamburg, Hamburg, Germany Rudolf Heinrich Logging and Roads Branch, Food and Agriculture Organization, Rome, Italy Wai Mun Ho Forest Biotechnology Division, Forest Research Institute Malaysia, Kepong, Selangor, Malaysia Thomas Holmes USDA Forest Service, Southern Research Station, Research Triangle Park, Asheville, NC, USA Shoana Humphries Earth Innovation Institute, San Francisco, CA, USA Steven Johnson International Tropical Timber Organization, Yokohama, Japan Stefan Julich Institute of Soil Science and Site Ecology, Technische Universität Dresden, Dresden, Germany Kay Kallweit Deutsche Gesellschaft f€ur Internationale Zusammenarbeit (GIZ) GmbH, Eschborn, Germany Angelika Kandzior Puerto Montt, Chile Thomas Knoke Institute of Forest Management, TUM School of Life Sciences Weihenstephan, Department of Ecology and Ecosystem Management, Technische Universität M€ unchen, Freising, Germany Gerald Koch Th€ unen Institute of Wood Research, Hamburg, Germany Sarah A. Laird People and Plants International, Bristol, VT, USA Thomas D. Landis Native Plant Nursery Consulting, Medford, OR, USA Silke Lautner Eberswalde University for Sustainable Development, Eberswalde, Germany Michael Köhl Center for Wood Sciences, Institute of World Forestry, University of Hamburg, Hamburg, Germany Daniel K€ ubler Institute for Worldforestry, University of Hamburg, Hamburg, Germany Don Koo Lee Park Chung Hee School of Policy and Saemaul, Yeungnam University, Daegu, Republic of Korea Steen Magnussen Natural Resources Canada, Victoria, Canada Kenneth MacDicken Formaly FAO, Rome, Italy Jorge R. Malinovski Malinovski Florestal, Curitiba, Brazil Rafael A. Malinovski Malinovski Florestal, Curitiba, Brazil

lviii

Contributors

Ricardo A. Malinovski Federal University of Paraná, Curitiba, Brazil Stefan Mann ECO Consult Sepp & Busacker Partnership, Oberaula, Germany Frances Maplesden Maplesden Consulting, Rotorua, New Zealand Marco Marchetti Department DIBT - BioScienses and Territory, University of Molise, Pesche, IS, Italy F. Maureira Unidad de Ergonomía, Facultad de Ciencias Biológicas, Universidad de Concepción, Concepción, Chile F. Meyer Unidad de Ergonomía, Facultad de Ciencias Biológicas, Universidad de Concepción, Concepción, Chile Julian Michel Frankfurt am Main, Germany Ramon Mollenhauer World Bank, Washington, DC, USA Reinhold G. Muschler Agroecology and Agrobiodiversity, Agroforestry and Sustainable Agriculture Program, Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), Turrialba, Costa Rica Hosea M. Mwangi Institute of Soil Science and Site Ecology, Technische Universität Dresden, Dresden, Germany Biomechanical and Environmental Engineering Department, Jomo Kenyatta University of Agriculture and Technology, Juja, Kenya Prem Raj Neupane University of Hamburg, World Forestry, EFI (European Forest Institute) Project Centre SURF (Supporting the Global Implementation of REDD+ and FLEGT), Hamburg, Germany Leif Nutto Malinovski Florestal, Curitiba, Brazil Konstantin Olschofsky Institute for Worldforestry, University of Hamburg, Hamburg, Germany E. Oñate Unidad de Ergonomía, Facultad de Ciencias Biológicas, Universidad de Concepción, Concepción, Chile Laslo Pancel Deutsche Gesellschaft f€ur Internationale Zusammenarbeit (GIZ) GmbH, La Libertad, El Salvador Diana Rodríguez Paredes World Bank, Latin America and the Caribbean Division, Washington, DC, USA Jutta Poker Formerly Institute for World Forestry, Hamburg, Germany Thorsten Peters Institute of Geography, Friedrich-Alexander-University ErlangenNuremberg, Erlangen, Germany Alan R. Pierce People and Plants International, Duxbury, VT, USA

Contributors

lix

Daniel Plugge SPC/GIZ Regional Program, REDD+ - Forest Conservation in Pacific Island Countries II, Deutsche Gesellschaft f€ur Internationale Zusammenarbeit (GIZ), Suva, Fiji Formerly Institute for Worldforestry, University of Hamburg, Hamburg, Germany Laura Prill Institute for Worldforestry, University of Hamburg, Hamburg, Germany Christoph Richter Opitzer Weg 20, Tharandt, Germany Michael Richter Institute of Geography, Erlangen-Nuremberg, Erlangen, Germany

Friedrich-Alexander-University

Esteban Rivas Valdivia, Chile Mohammed N. Salleh Academy of Sciences Malaysia, Kuala Lumpur, Malaysia Nathan S. Sanches Federal University of Paraná, Curitiba, Brazil Werner Schindele Forest management expert and Independent Consultant, Hamburg, Germany Lars Schmidt University of Copenhagen, Copenhagen, Denmark Klaus Schmitt Department of Environment and Natural Resources, Deutsche Gesellschaft f€ ur Internationale Zusammenarbeit (GIZ) GmbH, Quezon City, Philippines Thomas Schneider University of Hamburg, World Forestry, Hamburg, Germany Carsten Schusser Forest Policy Expert, associated with Chair of Forest and Nature Conservation Policy, Georg-August-Universität Göttingen, Göttingen, Germany Alexandre Magno Sebbenn Instituto Florestal do Governo do Estado de Sao Paulo, Piracicaba, SP, Brazil Steve Sepp ECO Consult Sepp & Busacker Partnership, Oberaula, Germany John Sessions College of Forestry, Oregon State University, Corvallis, OR, USA Patricia Shanley Woods & Wayside International, Hopewell, NJ, USA Erin Sills Department of Forestry and Environmental Resources, North Carolina State University, Raleigh, NC, USA Thorsten Specht GFA Certification Ltd., Buchholz, Germany Martin R. Speight St Anne’s College, University of Oxford, Oxford, UK Jörn Struwe Thuenen Institute of International Forestry and Forest Economics, Hamburg, Germany A. Swenne Brussels, Belgium Michelle Szejner German Cooperation Agency (GIZ), Panamá, Panamá

lx

Contributors

Evy von Pfeil Deutsche Gesellschaft f€ur Internationale Zusammenarbeit (GIZ) GmbH, Eschborn, Germany Kim M. Wilkinson Gibsons, BC, Canada Stephen Woodward Department of Plant and Soil Science, Institute of Biological and Environmental Sciences, University of Aberdeen, Aberdeen, UK Anthony Young School of Environmental Sciences, University of East Anglia, Norwich, UK Astrid Zabel School of Agricultural, Forest and Food Sciences, Bern University of Applied Sciences, Bern, Switzerland Wolfgang Zech Bodenkunde und Bodengeographie, Universitaet Bayreuth, Bayreuth, Germany

Part I The Global Forestry Sector

Tropical Forest Resources: Facts and Tables Jutta Poker and Kenneth MacDicken

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Extent of the Tropical Forest Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Extent, Naturalness, and Designation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annual Change Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Growing Stock and Carbon Stocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biodiversity in Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Area of Primary Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Forest Area Designated for Conservation of Biological Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . Tropical Forests in Protected Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Health and Vitality of Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Forest Fires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pests and Diseases, Natural Disasters, and Invasive Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Human-Induced Disturbances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Climate Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Productive Functions of Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Areas Designated for Productive Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Planted Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Removals of Wood Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Removals of Non-wood Forest Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Protective Functions of Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Forest Area Designated for Soil and Water Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Socioeconomic Functions of Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ownership and Management Rights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Public Expenditure and Revenue Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Value of Wood and Non-wood Forest Product Removals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Area of Forest Designated for Social Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 6 6 9 10 10 10 11 11 11 11 12 12 12 13 13 13 13 15 16 16 16 16 17 17 17 18

J. Poker (*) Formerly Institute for World Forestry, Hamburg, Germany e-mail: [email protected] K. MacDicken Formaly FAO, Rome, Italy # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Köhl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_7

3

4

J. Poker and K. MacDicken

Annex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18 18 18 18 18 18 18 43

Abstract

More than 40 % of the world’s 4 billion hectares forests are located in tropical regions and cover 1.73 billion hectares which corresponds to nearly half of the tropical land area. Deforestation – mainly the conversion of tropical forests to agricultural land – shows signs of decreasing in several countries but continues at a high rate in others. Around 8 million hectares of tropical forest were converted to other uses or lost through natural causes each year in the last decade compared to more than 10 million hectares per year in the 1990s. Fifteen tropical countries loose more than 1 % of their forests per year, in five countries forest area is stable, and in nine countries forest area is slightly increasing by a total of 0.3 million hectares per year. Half of the world’s growing stock is located in tropical forests. In terms of carbon content, they have a share of about 60 %. On average, tropical forests in Africa and Latin America/Caribbean store 100 t carbon per ha, in Asia/Pacific 75 t carbon per ha. Primary forest, i.e., forest of native species where there are no clearly visible indications of human activities and the ecological processes have not been significantly disturbed, includes the most species-rich, diverse terrestrial ecosystems. In Africa and Asia/Pacific, the share of primary forests on total tropical forest area is 42 %, while in Latin America/Caribbean still 74 % are primary. Overall, the area of primary forests is decreasing in all tropical regions with about 3.7 million hectares per year, but the situation seems to be improving especially in Asia/Pacific, while the rates of conversion show an increasing trend in Latin America/Caribbean. About 15 % of tropical forests are designated as primary function for the conservation of biodiversity. National parks, game reserves, wilderness areas, and other legally established protected areas also cover about 15 % of the total tropical forest area. The primary function of these forests may be the conservation of biological diversity, the protection of soil and water resources, or the conservation of cultural heritage. Half of all tropical countries declare forest fires as severe problem. Severe storms, flooding, and earthquakes have also damaged areas of forests. Nearly all countries in the tropics face at least forest degradation as result of the impact of human interventions in production forests, protected areas and parks. In many tropical countries, the climate appears to be changing. Recent data provide evidence of, for example, increasing temperatures and prolonged dry periods

Tropical Forest Resources: Facts and Tables

5

in some regions and increased rainfall and more frequent tropical storms in others. Half of the tropical forest is designated as permanent forest estate (PFE). Again half of these, about 400 million hectares, serve production purposes. Due to accessibility problems, only parts of the production forests are available for harvest. About 3 % of the permanent forest estate is planted forest. Reported wood removals amount to 1.3 billion cubic meters annually and equivalent to 0.5 % of the total growing stock. By far the most important product is fuelwood, although the statistics on this product are neither complete nor precise. Only few tropical countries are able to report on amount and value of non-timber forest products. Keywords

Biodiversity • Carbon content • Climate change • Conversion of tropical forests • Deforestation • Primary tropical forests • Tropical forests • Tropical forest resources

Introduction During the last decade, the information on tropical forests improved considerably. Though still many information gaps exist, an attempt is made to summarize current knowledge on state of the forests and forestry in tropical countries. The following analysis is based on data (see Annex) compiled from: 1. Global Forest Resources Assessment 2010 and associated remote sensing analyses (FAO 2010) 2. ITTO (Blaser et al. 2011): ITTO producer countries (33 countries representing more than 80 % of the total tropical forest area) 3. FCPF (Country Readiness Preparation Proposals: http://www. forestcarbonpartnership.org): all participating countries as supplement 4. Country data presented at official websites: all countries with low information status as supplement. Considered are all countries situated in the tropical regions as listed by ITTO and FAO (65 countries) as well as Nepal which is listed by FAO only. The descriptions follow the structure of the Forest Resources Assessment (FAO 2010). A fundamental difficulty in reporting tropical forest area is that many countries have more than one climatic domain. For example, China and the United States both have tropical forest but they are a fraction of forest area. Likewise, while Peru has substantial tropical forest and is an ITTO producer country, they also have significant forest that is not tropical. Thus, one must take care in interpreting forest area based on country alone unless the country has reported forest area by forest type. Of the analyses presented in this chapter, only the remote sensing work of the Global Forest Resources Assessment (FRA) reports forest area and change based on climatic ecozones (Table 1).

6

J. Poker and K. MacDicken

Table 1 Forest area (million hectares,  95 % confidence interval) by region and climatic domain. Forest area figures are presented rounded to the nearest significant digit FRA region Africa Asia

Europe

North and Central America

Oceania

South America

World

World

Climatic domain Subtropical Tropical Boreal Subtropical

Samples 122 2,415 31 769

Temperate Tropical Boreal Subtropical Temperate Boreal Subtropical Temperate Tropical Subtropical Temperate Tropical Subtropical Temperate Tropical Boreal

1,273 911 294 94 531 2,777 368 1,593 127 429 51 300 177 96 1,217 3,102

Subtropical Temperate Tropical

1,959 3,544 4,970 13,575

1990 4  51 % 590  6 % 16  16 % 130  12 % 70  16 % 310  8 % 800  5 % 18  26 % 270  9 % 380  2 % 90  13 % 260  6 % 70  12 % 30  25 % 21  20 % 70  19 % 20  26 % 13  33 % 820  4 % 1,200  3 % 300  7 % 630  5 % 1,860  3 % 4,000  3 %

2000 5  51 % 580  7 % 17  15 % 150  11 % 80  15 % 290  8 % 800  5 % 18  25 % 270  9 % 390  2 % 90  13 % 260  6 % 70  12 % 30  25 % 21  20 % 70  19 % 20  25 % 13  33 % 790  4 % 1,200  3 % 320  7 % 640  5 % 1,790  3 % 3,950  3 %

2010 4  52 % 560  7 % 18  16 % 160  11 % 90  15 % 280  9 % 790  5 % 18  25 % 260  9 % 380  2 % 90  12 % 250  6 % 70  12 % 30  25 % 20  20 % 70  19 % 20  25 % 13  33 % 760  4 % 1,190  3 % 330  7 % 630  5 % 1,730  3 % 3,890  3 %

Source: FAO Global Forest Resources Assessment remote sensing analysis (2014)

Extent of the Tropical Forest Resource Extent, Naturalness, and Designation Tropical forests form a variety of unique ecosystems leading to the rich diversity of the tropics. Tropical rainforests merge into other types of forest depending on the altitude, latitude, and various soil, flooding, and climate conditions. They occur in the equatorial zone, within the area bounded by latitudes 23.5 N (Tropic of Cancer) and 23.5 S (Tropic of Capricorn). One of the major characteristics of tropical forests is their distinct seasonality: winter is absent, and only two seasons may occur. The

Tropical Forest Resources: Facts and Tables

7

length of daylight is 12 h and varies little. The seasonal distribution of rainfalls determines the subdivision in: • Evergreen rainforest: no dry season. • Seasonal rainforest: short dry period in a very wet tropical region (the forest exhibits definite seasonal changes as trees undergo developmental changes simultaneously, but the general character of vegetation remains the same as in evergreen rainforests). • Semievergreen forest: longer dry season (the upper tree story consists of deciduous trees, while the lower story is still evergreen). • Moist/dry deciduous forest (monsoon): the length of the dry season increases further as rainfall decreases (all trees are deciduous). • http://www.srl.caltech.edu/personnel/krubal/rainforest/Edit560s6/www/where.html

Arctic Ocean

Atlantic Ocean

Pacific Ocean

equator Indian Ocean

Pacific Ocean

Rainforests of the world

The bulk of the world’s tropical rainforest occurs in the Amazon Basin in South America. The Congo Basin and Southeast Asia, respectively, have the second and third largest areas of tropical rainforest. Rainforests also exist on some the Caribbean islands, in Central America, in India, on scattered islands in the South Pacific, in Madagascar, in West and East Africa outside the Congo Basin, in Central America and Mexico, and in parts of South America outside the Amazon. Brazil has the largest extent of rainforest of any country on Earth.

8

J. Poker and K. MacDicken

3000 2000

'000 ha/year

1000 0 –1000 –2000 Gains

–3000 Losses

–4000 Net change

South America

Africa

Asia

Europe

North and Central America

tropical

temperate

subtropical

tropical

temperate

subtropical

boreal

temperate

boreal

subtropical

tropical

temperate

boreal

subtropical

tropical

subtropical

tropical

temperate

subtropical

–5000

Oceania

Fig. 1 Annual change in forest land-use area (1990–2010) by region and climatic domain

According to FAO, tropical forests extend on 1.70 billion hectares in 2010 based on Landsat image analysis (Table 5, Table 1). The world’s forests are distributed unevenly with just under half the world’s forests in the tropical domain (45 % of total area), about one third in boreal (31 %) and smaller amounts in temperate (16 %) and subtropical (8 %) domains. Figure 1 shows regional differences in the rate of change in forest area. The highest rate of forest conversion to other land uses was in South America, followed by Africa and Asia. Net forest loss in the tropical domain was reasonably constant from 1990 to 2010, going from 6 million hectares per year in the 1990s to 7 million hectares per year in the 2000s. About half of the land area in the tropics is covered by forests. Forest coverage is highest in Latin America/Caribbean (56 %), followed by Africa (48 %) and Asia/ Pacific (39 %). On country level, the highest coverage (85–98 %) is found in Gabon, Suriname, and French Guyana. Only few forests (4–7 % coverage) exist in Togo, Burundi, Kenya, and Haiti. ITTO producer countries are covered by 1.42 billion hectares tropical forests following FAO, but ITTO estimates the extent in a range between 1.30 and 1.39 billion hectares. While FAO includes the total forest area of India and Mexico (133 million hectares), ITTO estimates the area of tropical forests only (69 million hectares). Only ten of the 33 ITTO producer countries correspond to FAO figures. Seven countries conduct no forest inventory, 2 countries prepare for their first inventory, 10 countries request inventories only within the forest management units (FMUs), 5 countries rely on inventories conducted before 2000 and 8 countries accomplished their last inventory during the previous decade (Table 5, Annex 2).

Tropical Forest Resources: Facts and Tables

9

Table 2 Forest area and permanent forest estate in tropical countries and subdivision of the PFE in ITTO producer countries (Blaser et al. 2011) complemented by FAO 2010 Forest area

Permanent forest estate (PFE)

PFE for production natural

PFE for production planted

Region Total tropical Total ITTO

'000 ha 1,730,831 1,420,513

‘000 ha (% of total forest area) 881,081 (51) 783,101 (55)

‘000 ha (% of PFE)

‘000 ha (% of PFE)

403,196 (52)

22,371 (3)

Africa

270,067

112,751 (42)

68,244 (62)

950 (1)

Asia/Pacific

282,006

178,627 (63)

108,219 (61)

12,038 (7)

Latin America/ Caribbean

868,440

491,723 (57)

226,706 (46)

9,383 (2)

PFE for protection ‘000 ha (% of PFE) 357,755 (45) 43,210 (38) 58,370 (33) 255,687 (52)

The Permanent Forest Estate in Tropical Countries ITTO reported that some 910 million hectares are primary forests, of these 870 million hectare are in ITTO producer countries. Half of the forest area serves no designated purpose. In tropical Africa 60 % of the forest area has no defined status, in Asia/Pacific its 44 % like in Latin America/Caribbean (45 %). 880 million hectares are designated as permanent forest estate (PFE), of these 780 million hectares are in ITTO producer countries. Nearly 3 % of the PFE are planted forests, i.e., 0.4 % in Africa, 7 % in Asia/ Pacific, and 2 % in Latin America/Caribbean. The PFE serves production purposes (55 % of the area) as well as protection services (45 %). In Latin America/Caribbean, the area of forests for protection exceeds that of production forests (Table 2). More than half of the tropical forest is closed forest whose tree canopy covers 60 % or more of the ground surface, when viewed from above. In Africa, on average 60 % of the forest area is closed. The highest ranking is found in Liberia and Gabon, but the canopy covers less than 20 % in Ghana, Côte d’Ivoire, Nigeria, and Togo. In Asia/Pacific, on average 51 % of the forests are closed namely in Vanuatu, Papua New Guinea, and Malaysia. In contrast, India has a low proportion of closed forests. In Latin America/Caribbean, more than half of the forests (55 %) are closed in all countries except Mexico. Suriname and Guyana show the highest ranking of all ITTO producer countries in terms of closed forests.

Annual Change Rates Annual change rates in tropical forest area vary slightly between FAO and ITTO estimates (Table 4, Annex 1). Greatest discrepancies exist in Nigeria, Cameroon, Mexico, and Peru.

10

J. Poker and K. MacDicken

Annual change rates range from 5, to 75 % in Togo to +1.1 % in Viet Nam. The total gross annually deforested area in the tropics between 2005 and 2010 is 8.2 million hectares, when considering new plantations, the annually affected net area is reduced to 7.9 million hectares. The highest annual losses are observed in Brazil (2.2 million hectares), Indonesia (0.7 million hectares), Nigeria and Tanzania (0.4 million hectares), and Cameroon, Democratic Republic of Congo Zimbabwe, Bolivia, and Venezuela (about 0.3 million hectares). In all tropical regions, deforestation is driven primarily by conversion to agricultural land use. Additionally, in Africa fuelwood gathering and charcoal production play an important role – but one that is not well quantified. The Asia/Pacific regions suffer periodically from destruction by fires. In Latin America/Caribbean, mining and infrastructure development are also important drivers. In some tropical countries, the forest area is extending namely in India and Viet Nam as well as, though on lower level, in Costa Rica and Cuba. Still, reafforestation in tropical regions reduces tropical forest losses only by about 0.3 million hectares per year.

Growing Stock and Carbon Stocks Half of the world’s growing stock is located in tropical forests. The majority is stocking in Latin America/Caribbean (62 % with 48 % in Brazil) followed by Africa (27 %) and Asia/Pacific (11 %) (Table 5, Annex 2). FRA 2010 estimates that the world’s forests store 289 gigatonnes (Gt) of carbon in their biomass alone. Tropical forests have a share of about 60 %. Carbon in tropical forests is again concentrated in Latin America/Caribbean (55 % with 37 % in Brazil) followed by Africa (29 %) and Asia/Pacific (16 %). On average. tropical forests in Africa and Latin America/Caribbean store 100 t carbon per ha, in Asia/ Pacific 75 t carbon per ha. While sustainable management, planting, and rehabilitation of forests can conserve or increase forest carbon stocks, deforestation, degradation, and poor forest management reduce them. Information on changes in carbon stocks is scarce. For reporting period 2005–2010, most countries report not significant changes, only Indonesia (1.7 t/ha/year) and Malaysia (0.8 t/ha/year) provided data.

Biodiversity in Tropical Forests Area of Primary Forests Forests of native species where there are no clearly visible indications of human activities and the ecological processes have not been significantly disturbed are considered as primary forests. They include the most species-rich, diverse terrestrial ecosystems. More than half of the tropical forests worldwide, i.e. 0.91 billion hectares, are primary forests (Table 6, Annex 3). In Africa and Asia/Pacific, the share of primary forests on total tropical forest area is 42 %, while in Latin America/

Tropical Forest Resources: Facts and Tables

11

Caribbean still 74 % are primary. The decrease of primary forests during the last decades is largely due to reclassification of primary forest to "other naturally regenerated forest" because of selective logging, shifting cultivation, and other human interventions. Overall, the area of primary forests is decreasing in all tropical regions at a rate of about 3.7 million hectares per year, but the situation seems to be improving especially in Asia/Pacific, while the rates of conversion show an increasing trend in Latin America/Caribbean. More than 70 % of all losses of primary tropical forests occur in Brazil although this seems to be slowing in recent years. Relatively high conversion rates are also observed in Papua New Guinea and Gabon.

Forest Area Designated for Conservation of Biological Diversity About 15 % of tropical forests are designated as primary function for the conservation of biodiversity (Table 6, Annex 3). This is more than the global average of about 13 %. Only five countries in the tropics were not able to report on biodiversity conservation areas though for instants countries like Kenya and the Dominican Republic are known for their nature reserves. The highest share of biological diversity conservation areas which are tropical forests is found in the Asia/Pacific region. Most, but by far not all of these areas are legally established protected areas. This is especially true for Latin America/Caribbean.

Tropical Forests in Protected Areas National parks, game reserves, wilderness areas, and other legally established protected areas also cover about 15 % of the total tropical forest area (Table 6, Annex 3). The primary function of these forests may be the conservation of biological diversity, the protection of soil and water resources, or the conservation of cultural heritage. In Africa and Latin America/Caribbean, the share of legally protected area is about 12 % of the total tropical forest area, while in Asia/Pacific the share amounts to 28 %. The situation varies widely between countries. The highest shares with more than half of the total forest area in a legally protected status are found in Thailand, Nicaragua, and Panama.

Health and Vitality of Tropical Forests Forest Fires While some forest ecosystems depend on fire for their regeneration, forest fires can be devastating to others and also frequently cause loss of property and human life. In tropical forests, less than 1 % of all forests were reported to be significantly affected each year by forest fires. However, the area of forest affected by fires was severely

12

J. Poker and K. MacDicken

underreported, with information missing from many countries. Still, half of all tropical countries declare forest fires as severe problem. The greatest damaged areas are reported from India, Ghana, Cameroon, and Myanmar (Table 7, Annex 4). Less than 10 % of all forest fires are prescribed burning; most are classified as wildfires.

Pests and Diseases, Natural Disasters, and Invasive Species Information availability and quality continues to be poor for most of these disturbances. Outbreaks of forest insect pests are reported from India, Mexico, El Salvador, Guatemala, Honduras, and Peru. Severe storms, flooding, and earthquakes have also damaged large areas of forests. During the last 15 years, hurricanes hit especially Myanmar, Guatemala, Honduras, Cuba, Haiti, Nicaragua, and Jamaica. Mozambique, Indonesia, Myanmar, and Thailand suffered from severe flooding. Earthquakes destroyed parts of Indonesia, Papua New Guinea, El Salvador, and Haiti. Woody invasive species are of particular concern in small island developing states, where they can threaten the habitat of endemic species.

Human-Induced Disturbances Healthy biological functioning of forest ecosystems can be affected by a variety of human actions such as encroachment, illegal harvesting, human-induced fire and pollution, grazing, mining, poaching, etc. Nearly all countries in the tropics face at least forest degradation as result of the impact of human interventions in production forests, protected areas, as well as in parks.

Climate Change ITTO producer countries were asked to specify their expectations concerning the vulnerability of their forests to climate change (Table 7, Annex 4). Blaser et al. (2011) concluded: “Climate change and climate variability could be among the most serious threats to sustainable development, with potential adverse impacts on natural resources, physical infrastructure, human health, food security and economic activity. Forests and rural landscapes in the tropics may be particularly vulnerable to the effects of climate variability, for example extreme weather events such as droughts (and associated wildfires), flooding and storms. At the same time, forests have the capability to reduce both environmental and social vulnerability. In many tropical countries the climate appears to be changing. Recent data provide evidence of, for example, increasing temperatures and prolonged dry periods in some regions, and increased rainfall and more frequent tropical storms in others. In Mexico, there has been an increase in mean annual temperature

Tropical Forest Resources: Facts and Tables

13

of 0.6  C in the past four decades. In Peru, average annual temperature has increased by 0.3  C in the last 50 years. In Ghana, average annual temperature has increased by 1.0  C since 1960, thus damaging the integrity of forest ecosystems. Adaptive approaches to forest management will become increasingly important in the face of climate change. Regardless of the pace of such change, healthy forests maintained under SFM will be better able to cope than those weakened and/or degraded by overexploitation.”

Productive Functions of Tropical Forests Areas Designated for Productive Functions Half of the tropical forest is designated as permanent forest estate (PFE). Again half of these, about 400 million hectares, serve production purposes. In Asia/Pacific, production forests have a share of more than one third of the total forests. ITTO producer countries report on their production forests in more detail (Table 3). Due to accessibility problems, only parts of the production forests are available for harvest. In Latin America/Caribbean, only one fourth of these forests can be exploited, while in Africa nearly two thirds are accessible. In Asia/Pacific, half of the production forests are covered by management plans. This share is with 20 % lowest in Latin America/Caribbean. Certification also plays a minor role in Latin America/Caribbean. Still, up to now the area of certified forests is slightly increasing throughout the tropics but especially some countries in Latin America/ Caribbean observe nonrenewals of certificates because demand for certified timber is lacking.

Planted Forests About 3 % of the permanent forest estate is planted forest. During the decade 2000–2010, there is a decreasing trend in forest plantations in Angola, Burundi, Papua New Guinea, and Sri Lanka. In half of the tropical countries, the plantation area did not change significantly, but 28 countries show an increasing trend, especially Brazil, Viet Nam, Malaysia, Peru, Myanmar, Ghana, Colombia, and Ecuador.

Removals of Wood Products Reported wood removals amount to 1.3 billion cubic meters annually and equivalent to 0.5 % of the total growing stock (Table 8, Annex 5). Most countries have a stable timber production level. By far the most important product is fuelwood. Since some countries regard fuelwood as non-timber forest product (NTFP) and do not include this wood in their statistics, the actual amount of wood removals is undoubtedly

Region Total ITTO Africa Asia/Pacific Latin America/ Caribbean

‘000 ha 403,169 68,244 108,219 226,706

PFE for production

PFE available for harvest % ‘000 ha PFE prod. 165,332 41 45,714 67 62,766 58 56,852 25

Forest area with management plans % ‘000 ha PFE prod. 129,062 32 26,359 39 58,013 54 44,690 20

Table 3 Status of the PFE for production in ITTO producer countries (Blaser et al. 2011) Certified forest area ‘000 ha % PFE 2010 total 17,617 2 4,628 4 6,367 4 6,622 1

‘000 ha 7/2012 24,179 5,699 7,170 11,310

% PFE total 3 5 4 2

14 J. Poker and K. MacDicken

Tropical Forest Resources: Facts and Tables

15

higher than reported. There is also no estimate on informally and illegally removed wood. About 1.5 % of the harvested wood is exported.

Forest Management for Production More than half of all tropical countries developed forestry guidelines, six of them have none (Table 5, Annex 2). Twenty out of 65 countries conducted a national forest inventory, 16 countries conduct inventories in their forest management units (FMUs), 7 countries definitely have no inventory information, the situation in the remaining counties is unknown. The monitoring capacity is low in most countries; high capacities are reported by Côte d’Ivoire, India, Malaysia, Brazil, Guyana, and Mexico. Seventeen countries contract out concessions which differ considerably in size and duration between countries (Table 8, Annex 6). Thirteen countries offer shortterm harvest permits. Usually, standards for harvest are set and minimum diameter rules for species or species groups are prescribed. Ten countries are committed to reduced impact logging systems (RIL), but chainsaw logging and high grading are still widespread. Most countries rely on successful natural regeneration, but 12 countries also use enrichment planting.

Removals of Non-wood Forest Products Only few tropical countries are able to report on amount and value of non-timber forest products (NTFPs) such as Brazil, Colombia, India, Malaysia, Mexico, Costa Rica, El Salvador, Tanzania, and The Philippines. The major categories of NWFP removals about which countries provided the most information are (in descending order of importance): 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Food Exudates Other plant products Wild honey and beeswax Ornamental plants Raw materials for medicine and aromatic products Wild meat Raw materials for utensils, handicrafts and construction Living animals Hides, skins, and trophies

Some countries, especially in Latin America/Caribbean, introduced or are introducing markets to facilitate payments for environmental services (PES) such as water catchment protection, biodiversity conservation, and carbon sequestration. At the international level, the volume and value of payments is still low, but it is expected that there is substantial potential for an increase, especially for carbon sequestration.

16

J. Poker and K. MacDicken

Protective Functions of Tropical Forests Forest Area Designated for Soil and Water Conservation One of the most important protective function of forests is related to soil and water resources. Forests conserve water by increasing infiltration, reducing runoff velocity and surface erosion, and decreasing sedimentation. Forests play a role in filtering water pollutants, regulating water yield and flow, moderating floods, enhancing precipitation, and mitigating salinity. The forest area with “protection of soil and water as the primary designated function” refers specifically to the area of forests that have been set aside for the purposes of soil and water conservation, either by legal prescription or by decision of the landowner or manager. More specifically, the variable refers to soil and water conservation, avalanche control, sand dune stabilization, desertification control, and coastal protection. It does not include forests that have a protective function in terms of biodiversity conservation or those in protected areas, unless the main purpose is soil and water conservation. Following FAO, about 133 million hectares or nearly 8 % of the tropical forests have soil and water conservation as their primary objective (Table 6, Annex 3). The quantification of the protection forest area remains difficult. ITTO producer countries report much greater areas especially in Latin America. Brazil reported 43 million hectares forest designated for soil and water protection to FAO. The ITTO report states: “The Amazon Basin produces 20 % of the world’s freshwater; it is therefore vital that its soil and water resources are properly protected. An estimated 243 million hectares of forest in Brazil are managed primarily for soil and water protection.” In Africa, the greatest protective forests are located in Mozambique, Central African Republic, and the Republic of Congo. In Asia Pacific, Indonesia, Myanmar, Lao PDR, and Vietnam have the greatest protective forests. In Latin America/ Caribbean, they are found in Brazil, Venezuela, and Colombia.

Socioeconomic Functions of Tropical Forests Ownership and Management Rights In African tropical countries, most of the forests are in public ownership. Significant private ownership exists in Sierra Leone (86 % belong communities), Togo (73 % belong individuals), Uganda (68 %), Kenya (61 % belong mainly communities), Zimbabwe (32 %), and Central African Republic (9 %). The holder of management rights in public forests is usually public administrations or in few cases communities. In those countries where concessions for timber harvest are granted, business entities hold management rights for a given period. In Asia/Pacific as well as in Latin America/Caribbean, private forest ownership is much more spread especially in Papua New Guinea (97 %), Fiji (95 %), Timor Leste (67 % belong communities), El Salvador (69 %), Colombia (67 %), Jamaica (65 %), Paraguay (61 %), and Guatemala (52 %). Still, the holder of management rights are mainly public administrations.

Tropical Forest Resources: Facts and Tables

17

Public Expenditure and Revenue Collection Thirty-one of 65 tropical countries reported on revenues from forestry and public expenditure for forestry measures in 2005 (FAO 2010). On average, total forest revenue collection was about US$4.4 per hectare, ranging from US$0.3 per hectare in tropical Africa to US$6.6 per hectare in tropical Asia/Pacific. Public expenditures range from US$0.7 per hectare in tropical Africa to US$2.5 per hectare in Asia/ Pacific. In Latin America/Caribbean, the situation is dominated by Brazil. Here, revenue collection is relatively high with more than US$5 per hectare, and public expenditures are low. Without Brazil, the relation of revenues (US$0.7 per hectare) and expenditures (US$1.9 per hectare) are similar to the African situation. Only in Asia/Pacific, namely, Malaysia and Papua New Guinea, and Brazil revenues are higher than expenditures.

Value of Wood and Non-wood Forest Product Removals Forty-four tropical countries report on values of wood and non-wood forest removals in 2005 (FAO 2010). Wood removals valued just over US$25.7 billion annually in the period 2003–2007, accounted for by industrial roundwood (60 %) and woodfuel (40 %). In Liberia, Burundi, Madagascar, Rwanda, Tanzania, India, and Myanmar, the value of woodfuel trade exceeds that of industrial roundwood. The reported value of non-wood forest product removals amounts to about US$0.8 billion for 2005. Food products account for the greatest share. However, information is still missing from many countries in which non-wood forest products are highly important, and the true value of subsistence use is rarely captured. As a result, the reported statistics probably cover only a fraction of the true total value of harvested non-wood forest products. High values are reported by Brazil, Colombia, and India. Millions of people depend on food, medicine, and products from the forest in their daily life. Some ITTO producer countries estimated the number of depending people such as 45 million in the Democratic Republic of Congo, 48 million in Nigeria, 115.5 million in Cambodia, 38 million in Myanmar, more than 200 million in India, more than 5 million in Papua New Guinea, 25 million in the Philippines as well as in Thailand, and 12 million in Mexico. Payment for environmental services (PES) may generate additional income, but are not fully established yet. PES mainly for the management of water catchments is practiced in Kenya, Fiji, Brazil, Colombia, Ecuador, Guatemala, Guyana, Mexico, Costa Rica, Dominican Republic, and Paraguay. Regional initiatives or pilot projects are conducted in Madagascar, Indonesia, Vietnam, Panama, and El Salvador.

Employment During the last decade, reported employment in forest establishment, management, and use employment increased in 14 countries especially in Malaysia, Vietnam, and Paraguay – probably because roundwood production has increased faster than gains

18

J. Poker and K. MacDicken

in labor productivity. Employment decreased in 9 countries, especially in Indonesia and Jamaica. Some countries reported increased employment in management of protected areas such as Nigeria, Zimbabwe, and Vietnam. Given that much forestry employment is outside the formal sector, forest work is surely much more important for rural livelihoods and national economies than the reported figures suggest.

Area of Forest Designated for Social Services The forest area designated for recreation, tourism, education, or conservation of cultural and spiritual heritage is expanding in the tropics. Roughly about 0.17 million hectares or about 10 % of the tropical forest are designated for the provision of social services. Brazil has designated more than one fifth of its forest area for the protection of the culture and way of life of forest-dependent people.

Annex Annex 1 See Table 4

Annex 2 See Table 5

Annex 3 See Table 6

Annex 4 See Table 7

Annex 5 See Table 8

Annex 6 See Table 9

68

66

33 85 22 45

10 5 48

154,135

22,411

10,403 22,000 4,940 4,329

9,041 287 270,067

58,480 4,561 172 1,626 480

Country Cameroon Central African Republic Congo, Democratic Republic Congo, Republic of Côte d’Ivoire Gabon Ghana Liberia

Nigeria Togo Subtotal Africa

Angola Benin Burundi Equatorial Guinea Gambia

47 41 7 58 48

% 42 36

Total forest 000 ha FAO 19,916 22,605

% of total area

Table 4 Tropical forest area, change in area

112,000–154,000

22,400 10,400 21,700–24,600 4,680 3,330–4,390 9,000 500–1,700 226,140–282,370

0.2

0.05 – 0 2.19 0.68 4.0 5.75

0.21 1.06 1.01 0.71 +0.38

000 ha ITTO 19,700–21,200 22,700–>30,000

Total forest

Change 05–10 %/a FAO 1.07 0.13

n.a. 5.75

n.s. 0.12 2.2 0.351.0

0.03

0.2

%/a ITTO 0.14 0.19

Change 05–10

125 50 2 10 +2

410 20 1,310

60 %

Tropical Forest Resources: Facts and Tables 19

Cambodia Fiji India Indonesia Malaysia Myanmar Papua New Guinea

Country Guinea Guinea Bissau Kenya Madagascar Mozambique Rwanda Sierra Leone Tanzania, United Rep. of Uganda Zambia Zimbabwe Total Africa

Table 4 (continued)

% 27 72 6 22 50 18 38 38

15 67 40 44

57 56 23 52 63 48 63

2,988 49,468 15,624 503,663

10,094 1,014 68,434 94,432 20,456 31,773 28,726

% of total area

Total forest 000 ha FAO 6,544 2,022 3,467 12,553 39,022 435 2,726 33,428

1.2 +0.34 0.21 0.7 0.42 0.95 0.470.9

127 0 30–40 684 90 310 140–300

10,000–10,700 1,000 37,800a 94,400–98,500 18,400–18,600 30,800–35,400 28,600–33,000

000 ha/a 40 10 10 60 210 +10 20 400

1.22 +0.34 +0.21 0.71 0.42 0.95 0.49

%/a ITTO

Deforestation

90 170 330 2,825

000 ha ITTO

Total forest

Change 05–10

2.27 0.33 1.97

Change 05–10 %/a FAO 0.54 0.49 0.31 0.45 0.53 +2.47 0.7 1.16 0 0 5 27 (118,410) 3 48 42 50 21 10 91

Primary forest % (000 ha) 1 0 19 24 0 0 4 0

8,300 219 36,300 68,400 14,400 22,000 10,500

1,900 3,240 910 200,721

000 ha 1,190 – 1,260 3,260 – – 290 13,000

Permanent F. Estate

39 n.a. 13 69 79 58 79

%

Canopy cover > 60 %

20 J. Poker and K. MacDicken

Bolivia, Plurinational State of Brazil Colombia Ecuador Guatemala Guyana Honduras Mexico

Philippines Thailand Vanuatu Subtotal Asia/ Pacific Brunei Darussalam Lao People’s Democratic Republic Nepal Solomon Islands Sri Lanka Timor-Leste Vietnam, Socialist Republic of Total Asia/Pacific

26 37 36 39

72

68

25 79 29 50 44

40

53

62 55 36 34 77 46 33

7,665 18,972 440 282,006

380

15,751

3,636 2,213 1,860 742 13,797

320,385

57,196

519,522 60,499 9,865 3,657 15,205 5,192 64,802

52,400–57,200

519,000 56,900–61,500 9,900–11,200 3,700–4,600 15,200 5,800 31,400a

0.53

0.42 0.17 1.89 1.47 0 2.16 0.24

0.42 0.17 1.89 1.47 0.6–0 2.16 0.49

2,200 101 198 56 9–0 120 155

270

92 14 40 43 45 8 53

38 (120,940) 67

1,325 0.5

14 50 9 0 1

84 5 15 10 +144

0 0.25 0.77 1.44 +1.08

9

11 35 – 42 (117,840) 69

80

+30 +15 2 1,353

0.49

+0.7 +0.08 0.3

2

7,100–7,700 15,900–19,000 440 244,440–262,140

0.47

+0.73 +0.08 0

316,650 15,240 8,700 2,500 12,200 3,600 12,200

38,300

178,949

– 0 – – –

–

320

6,350 12,160 0 178,629

(continued)

51 60 59 51 89 51 35a

64

42 32 89 51

Tropical Forest Resources: Facts and Tables 21

Belize Costa Rica Cuba Dominican Republic El Salvador French Guiana Haiti Jamaica Nicaragua Paraguay

Country Panama Peru Suriname Trinidad & Tobago Venezuela, Bolivarian Republic of Subtotal LAC

Table 4 (continued)

14 98 4 31 26 44

287 8,082 101 337 3,114 17,582

56

868,440

61 51 26 41

52

46,275

1,393 2,605 2,870 1,972

% 44 53 95 44

% of total area

Total forest 000 ha FAO 3,251 67,992 14,758 226 46,700

0.61

4 4 1 n.s. 70 180

2 95 0 26 38 11

74 (648,000) 43 24 0 –

3,559

1.47 0.04 0.77 0.12 2.11 0.99

45

Primary forest % (000 ha) 22 60 93 28

288

000 ha/a 12 150 4 1

Deforestation

10 +23 +35 0

0.6

%/a ITTO 0.36 0.1 0.1 0.32

Change 05–10

0.68 +0.9 +1.25 0

827,026–842,926

000 ha ITTO 3,000–4,300 68,000–71,000 14,800 226

Total forest

Change 05–10 %/a FAO 0.36 0.22 0.02 0.32

– 6,600 – 120 – –

– – 2,870 –

491,690

33,400

000 ha 2,300 38,900 7,500 200

Permanent F. Estate

55

55

% 49 81 96 66

Canopy cover > 60 %

22 J. Poker and K. MacDicken

270,067

282,006

868,449

Africa ITTO

A/P ITTO

LAC ITTO

Bold: ITTO producer countries a Tropical forest only

48

1,730,831

56

39

48

50

142,0513

Total ITTO countries Total

55

906,783

Total LAC

827,026–842,926

244,440–262,140

226,410–282,370

1,297,876–1,387,436

6,222 7,920 1,310 1,325 3,559 8,214

0.43 0.45 0.48 0.48 0.39 Total gross

3,770 73 (660,110) 61 (873,570) 53 (909,460) 42 (113,730) 42 (117,840) 74 (648,000) 491,690

178,629

112,751

880,950

783,070

501,280

55

51

60

56

Tropical Forest Resources: Facts and Tables 23

75 58 96 125

2,710 381 585 1,085 – 4,385 263 17 203

4,895 291 684 1,161

– 59,592 2,266 161 20 268

18 506 61 629

Gabon Ghana Liberia Nigeria

Togo Subtotal Africa Angola Benin Burundi Equatorial Guinea

Gambia Guinea Guinea Bissau Kenya

32 96 96 525

–

3,438 1,842

4,539 2,632

66 47 47 137

123 77 135 120

153 177

Tonnes 135 127 127

Million t 2,969 2,861 19,639

Million m3 6,141 3,776 35,473

Country Cameroon Central African Republic Congo, Democratic Republic of Congo, Republic of Côte d’Ivoire

Carbon/ ha

Carbon

Growing stock

Table 5 Stocks, carbon

1998 1989 1991 2005

1955 1993 1985 1997

2000 65, in prep. 2001 1998 2006 1937 in prep 2008

Low

Low Low Low

Low

Medium Partly strong Low Low

Low Low

Institutional framework Forest Law enforcement law capacity 1994 Low 2008 Low 2002 Low

In work

Yes

None

1998 2006 1996

2005 2010

Forestry guidelines 1998 None

1992 Partly, FMUs 200810

None

In FMUs None, ghg In FMUs 1985–92 None None

Inventory 2004 1991–93 In FMUs

Low

Low

Improved Medium Low Low

Low High

Monitoring capacity Low Insufficient Low

24 J. Poker and K. MacDicken

Brunei Darussalam Lao People’s Democratic Republic Nepal Solomon Islands

Madagascar Mozambique Rwanda Sierra Leone Tanzania, United Republic of Uganda Zambia Zimbabwe Total Africa Cambodia Fiji India Indonesia Malaysia Myanmar Papua New Guinea Philippines Thailand Vanuatu Subtotal A/P

1,626 1,692 39 216 2,019 109 2,416 492 49,736 464 – 2,800 13,017 3,212 1,654 2,306 663 880 –

72 1,074 485 182

2,146 1,420 79 109 1,237

131 2,755 596 71,994 956 – 5,489 11,343 4,239 1,430 2,726 1,278 783 – Min 28,244 72 –

– 208 133 82

188 68

36 49 49 99 46 – 41 138 157 52 80 87 46 –

130 43 91 79 60

1993 2004

1984 2007

Weak Low

Improving

Low Low Low Low Low

Inadequate Low

Weak

2002 1992 1927 1999 1984 1992 1991 1975 2007 2001

1995 1996

Yes Yes

None

1999 1990 in work Several 2009, 10 Several 2000 1993 Several

Yes

No Yes

Low Low Low Low

2000

Low Low

2003 1973 1996

1997 1999/12 1988 1988 2002

1999 In FMUs

In FMUs 2006–08 Yes Yes 2007 None None 2003–05 None 1989–92

In progress In FMUs

2005

Low

(continued)

Low None High Medium High Low Low Medium Low Low

Medium Low Low Inadequate

Tropical Forest Resources: Facts and Tables 25

Colombia Ecuador Guatemala Guyana Honduras Mexico Panama Peru Suriname Trinidad and Tobago Venezuela, Bolivarian Republic of

Bolivia, Plurinational State of Brazil

Country Sri Lanka Timor-Leste Vietnam, Socialist Republic of Total A/P

Table 5 (continued)

75 78

24,116 4,442 62,607 6,805 – 281 1,629 330 2,043 367 8,560 3,165 19 –

Min 29,394 4,242

126,221

8,982 – 596 2,206 629 2,870 664 8,159 3,389 24 –

112 – 77 107 64 32 113 126 214 85 –

121

Tonnes 33 – 72

Million t 61 – 992

Million m3 – – 870

Carbon/ ha

Carbon

Growing stock

1974 In work 1996 2009 2008 2003 1994 2001 1992 In work 2008

1965

In work

Partially

Partially Improving High Low Improving Low Improving Medium

Strengthened

Weak

Institutional framework Forest Law enforcement law capacity 1995 Low 2000 Low 1992 Low

2003 Yes Yes None Yes

None 2004 Yes Yes 1996

2006

1997, 2006

Forestry guidelines

1980, fmus In prep. In fmus 2002–03 In fmus 2006 2004–07 In fmus In fmus In fmus 1969 In work

in FMUs

2008–10 2000–05

Inventory

Medium Medium Low High Low High Medium Medium Medium None Medium

High

low

Monitoring capacity

26 J. Poker and K. MacDicken

258 122 – 2,829 7 52 461 – Min 162,209 Min 263,597

Cuba Dominican Republic El Salvador French Guiana Haiti Jamaica Nicaragua Paraguay Total LAC

FAO 2010

Total

Total

Belize Costa Rica

Min 157,982 226 272

Subtotal LAC

170,648

226 114 – 1,651 5 48 349 – Min 93,050

171 238

289 Gt

79 58 – 204 54 141 112 – 103

123 91

Low Improving Improving

Low

Low

Yes

Yes Yes

Estimate for f r a total biomass carbon

527 billion m3 worldwide = 50 %

1998 1999 2002 2001 1926 1996 2003 1973/04

2000 1996

2003 2007–08 In fmus

In fmus

Pilot study

Low Low

Low

Low

Tropical Forest Resources: Facts and Tables 27

Country Cameroon Central African Republic Congo, Democratic Republic of Congo, Republic of Côte d’Ivoire Gabon Ghana Liberia Nigeria Togo Subtotal Africa Angola Benin Burundi Equatorial Guinea Gambia Guinea Guinea Bissau Kenya Madagascar Mozambique

% (‘000 ha) 18 10 51 33 6 90 5 4(FAO)–56 0 0 42(113,730) 0 0 23 0 n.s. 1 0 19 24 0

22,411 10,403 22,000 4,940 4,329 9,041 287 270,067 58,480 4,561 172 1,626 480 6,544 2,022 3,467 12,553 39,022

Primary forest

Total forest 000 ha (FAO) 19,916 22,605 154,135

Table 6 Biodiversity: primary forest

– – 0 – n.s. 0 – 0.3 / = 0.65 /+ –

0.08 /= 0 2.1 / + 0 0 n.s / ++ –

% / trend – 2.9/+ –

Change 90–10

990 810 3,430 43 190 2,510 – 33,623 1,860 1,260 40 590 40 240 – – 4,750 4,140 38 11

4 8 16 1 4 28 – 12 3 28 23 37 8 4

Protected areas FAO % total 000 ha forest 9,100 46 250 1 16,300 11 (0)3,660 374 (0) 350 (0) (0)57 (45)6 (7,070)10,145 0 0 0 0 60 590 240 3,260 1,250 8,580

000 ha (600) 5,700 (0)

Soil/water primary function

896 832 3,960 49 173 2,531 46 35,800 1,754 1,277 0 585 43 3,010 1,112 0 4,770 4,292

000 ha 884 226 26,203

Biol.div. prim. func

28 J. Poker and K. MacDicken

Rwanda Sierra Leone Tanzania, United Republic of Uganda Zambia Zimbabwe Total Africa Cambodia Fiji India Indonesia Malaysia Myanmar Papua New Guinea Philippines Thailand Vanuatu Subtotal A/P Brunei Darussalam Lao People’s Democratic Republic Nepal * Solomon Islands Sri Lanka Timor-Leste Vietnam, Socialist Republic of Total A/P 42(117,840) 69 9 14 50 9 0 1 38(120,940)

3,636 2,213 1,860 742 13,797 320,385

2 4 0 0 0 5 27(118,410) 3 48 42 50 21 10 91 11 35

435 2,726 33,428 2,988 49,468 15,624 503,663 10,094 1,014 68,434 94,432 20,456 31,773 28,726 7,665 18,972 440 282,006 380 15,751 0/+ 0 0/++ – 1.2/+

0.9/= 0

0/++ n.s./++ 0 0.2/+ 0 0 1.5/ 0 0 –

0 3.3 / = – – – 0

526 0 – 495 – 80,061

60 190 2,000 730 10,680 800 61,003 3,090 90 19,770 37,810 4,640 2,080 310 1,800 9,430 – 79,020 20 –

19

14 0

14 7 6 24 22 5 14 31 9 29 40 23 7 1 24 50 – 28 5

436 620 190 310 5,100 55,690

50 0 0 0 0 470 21,570 550 304 (10,700)4,540 (22,660)26,400 (2,660)5,200 (1,270)21,100 0 (613)1,500 1,330 – (39,870)60,920 20 9,140

(continued)

509 487 558 185 2,237 61,855

0 191 2,006 1,076 10,883 781 67,580 3,937 91 19,820 15,109 2,046 2,224 1,436 1,226 8,917 – 54,806 80 2,993

Tropical Forest Resources: Facts and Tables 29

Country Bolivia, Plurinational State of Brazil Colombia Ecuador Guatemala Guyana Honduras Mexico Panama Peru Suriname Trinidad & Tobago Venezuela, Bolivarian Republic of Subtotal LAC Belize

Table 6 (continued)

% (‘000 ha) 67 92 14 40 43 45 8 53 22 60 93 28 45 74(648,000) 43

868,440 1,393

Primary forest

Total forest 000 ha (FAO) 57,196 519,522 60,499 9,865 3,657 15,205 5,192 64,802 3,251 67,992 14,758 226 46,275

0

% / trend 0.5/ 0.48/+ 0.17/= 0.26/= 3.7/– 0 0 0.1/++ – 0.3 = 0.1 0 –

Change 90–10

115,185 –

13

Protected areas FAO % total 000 ha forest 10,680 19 89,540 17 – – – – – – – – 2,340 45 8,490 13 2,120 65 – – 2,015 14 – – – –

(55,040)266,750 0

000 ha 0 (43,000)243,000 (605)3,800 2,300 (0)950 0 (1,140)1,000 0 (65)406 (n.s.)756 0 (60)37 (7,870)14,500

Soil/water primary function

121,752 599

000 ha 10,867 46,757 8,470 4,834 2,304 152 2,285 8,424 1,333 18,358 2,214 20 15,734

Biol.div. prim. func

30 J. Poker and K. MacDicken

2,605 2,870 1,972 287 8,082 101 337 3,114 17,582 906,783

24 0 – 2 95 0 26 38 11 73(660,110)

0 – – 0 0.1/+ – 0.07/= 2.1 0

Countries in bold: 1. Blaser et al.(ITTO) 2011 all: 2. FAO 2010 (in brackets)

Costa Rica Cuba Dominican Republic El Salvador French Guiana Haiti Jamaica Nicaragua Paraguay Total LAC

– 630 – 30 2,420 5 120 2,020 – 120,410 29

10 30 5 36 65

22

290 1,350 – 10 0 0 10 190 n.s. 55,450

625 603 – 32 2,425 4 71 2,024 1,934 130,069

Tropical Forest Resources: Facts and Tables 31

Togo Angola Benin Burundi Equatorial Guinea Gambia Guinea Guinea Bissau Kenya Madagascar

Nigeria

Liberia

Ghana

Congo, Democratic Rep. of Congo, Republic of Côte d’Ivoire Gabon

Central African Republic

Country Cameroon

2 16

47

500

100 100

100 100

40

80

Burned 2003–07 000 % Wild ha/year fire 497 93

Table 7 Forest fires/climate change

x

x x

x

x

x

x

Fire reported as problem x

Incr.

Increased Increased

Expected trend Partly increased Increased

– Increased Increased of 0.14  C/ decade Increased of 0.21  C/ decade Increased of 0.18  C/ decade Increased of 0.03  C/ decade Increased

Expected trend Increased of 0.15  C/ decade Increased

Mean temperature

Decreased

Decreased

Decreased of 2.6 %/ decade Decreased of

–

Expected trend Decreased of 2.2 %/ decade Decreased of 2.2 %/ decade

Mean rainfall

32 J. Poker and K. MacDicken

Fiji India Indonesia Malaysia Myanmar Papua New Guinea Philippines Thailand Vanuatu Brunei Darussalam Lao People’s D R Nepal Solomon Islands Sri Lanka Timor-Leste Vietnam, Socialist Republic Bolivia, Plurinational State of Brazil

Mozambique Rwanda Sierra Leone Tanzania, United Republic of Uganda Zambia Zimbabwe Cambodia

2 21

1,605 5 2 218

20

15

100

100 100

100 100 100

x x

90 100

x x x x

x x x

xx x x

x x xx

x x

x

100

Increased

Increased

Increased

Increased Increased Increased incr.

Increased of 0.18  C/ decade Increased Increased Increased Increased

Decreased

Decreased Decreased

(continued)

Increased No long-term trend

No change

Tropical Forest Resources: Facts and Tables 33

63

100

100

7 9 3

0

x x

100

3

x

x x

x

95 92

x x x x

Fire reported as problem x

3 12

23 38

Burned 2003–07 000 % Wild ha/year fire

Countries in bold: 1. Blaser et al.(ITTO) 2011 all: 2. FAO 2010

Panama Peru Suriname Trinidad & Tobago Venezuela, Bolivarian Republic Belize Costa Rica Cuba Dominican Republic El Salvador French Guiana Haiti Jamaica Nicaragua Paraguay

Country Colombia Ecuador Guatemala Guyana Honduras Mexico

Table 7 (continued)

Increased

Increased

Increased

Expected trend Increased Increased

Increased Increased Increased

Increased of 0.13  C/ decade

Increased

Expected trend

Mean temperature

Changing patterns Changing patterns

No trend

Expected trend Changing patterns

Mean rainfall

34 J. Poker and K. MacDicken

x in prep.

2 29 6 26

39 30 0 25 4 31 9 5

1,700 2,720 0 68,244 2,340 1,410 15 80

Liberia Nigeria Togo Subtotal Africa* Angola Benin Burundi Equatorial Guinea n.s. 130 590 210 3,260

68 19 48 16

15,200 1,950 10,600 774

Gambia Guinea Guinea Bissau Kenya Madagascar

15

PFE production nat. % Total 000 ha forest 7,600 38 5,200 23

22,500

Fee

PES

Country Cameroon Central African Republic Congo, Democratic Republic Congo, Republic of Côte d’Ivoire Gabon Ghana

Table 8 Timber production

NTFP 90 50 75 90 95 45 80 95 85 95 90

5.10 = 6.70 = 10.70 + 1.00 = 0.80 = 12.60 = 2.70 = 27.60 = 13.30 =

NTFP 90 45 NTFP

2.60 = 21.50 = 3.40 = 1.32 + 0.36 + 77.00 = 6.00 =

85

% 67–85 85

Fuelwood

80.00 +

Production Mio m3/ trend 14.00 = 3.00 =

n.s. n.s. n.s. no logs2008 n.s. 0.17 0.01 n.s. n.s.

few 0.22 0.10

0.80 0.50 1.90 0,25

0,22

Mio m3 1.00 0.08

Export

(continued)

Low-quality timber, forests are difficult to assess High transport costs, no port Policy revision 2010, low political will Ban on unprocessed timber 2010 Log export banned since 1997, chainsaw lumber is illegal, but traded 2 of 4 ports work again >1/2 of log volume harvested by chainsaw No commercially exploitable forests left

High transport costs, no port; artisanal timber

Tropical Forest Resources: Facts and Tables 35

30 61 1 0 38 58

8,700

4,700 251 0 108,219 220

Philippines Thailand Vanuatu Subtotal A/P* Brunei Darussalam

12

Papua New Guinea

Regionally

x

360

PFE production nat. % Total 000 ha forest 26,170 67 320 74 240 9 23,730 71

24 10 28 37 0 38 41 50 50

Fee

PES

11,870 1,560 140,529 3,710 0 26,160 38,600 10,298 15,800

Zambia Zimbabwe Total Africa Cambodia Fiji India Indonesia Malaysia Myanmar

Country Mozambique Rwanda Sierra Leone Tanzania, United Republic Uganda

Table 8 (continued)

90 90

10.40 = 9.50 =

30 90 75 n.s.

0.10 =

NTFP

0.85 = 45.00 = 0.14 =

2.90 +

20 NTFP 85 86 NTFP 91

90

43.70 =

0.10  0.47 = 307.00 = 101.00 = 18.00  43.10 =

% 95 95 95 90

Fuelwood

Production Mio m3/ trend 18.10 = 6.20 = 5.70 = 25.00 =

banned

0.00 1.60 few

1.90

0.02 0.01 0.00 3.00 4.40 1.40

no logs 1,999 n.s. n.s.

Mio m3 0,01 – 0.02 0.01

Export

Logging in natural forest banned since 1988 Remaining forest difficult to access 50 % of wood supply from non-forest resources Illegal logging equals official harvest FSC + PEFC certified, harvest from plantations Government controls teak, limited profit for others 1,8 Mio m3/a by clearance authorities for agriculture, difficult access 1988 ban on old-growth logging Logging ban in natural forests since 1988 All land is owned by individuals or clans

36 J. Poker and K. MacDicken

21 13 11 28 36 56 28

1,964 1,140 11,090 1,096 8,400 350

18,700 5,319 127 12,920

x x x

x

(x)

Ecuador Guatemala Guyana

Honduras Mexico

Panama

Peru Suriname Trinidad & Tobago Venezuela, Bolivarian Republic

20 31 73

5,500

x

Colombia

9

26

37 44

119,587 25,100

135,000

17 9 33 47

380 170 247 6,480

x

pilot p.

23

3,620

Lao People’ Democratic Republic Nepal Solomon Islands Sri Lanka Timor-Leste Vietnam, Socialist Republic Total A/P Bolivia, Plurinational State of Brazil

90 NTFP 1 NTFP NTFP

2.40 = 0.20 = 0.05 = 2.40 =

90 NTFP

10.80  2.40 = 1.50 

80 40 5

85

13.00 = 4.80 + 16.00 + 0.30 =

50

NTFP

25 NTFP 90 100 90

95

247.00 =

2.70 +

0.20 = 1.60 = 5.80  0.10 = 27.80 =

6.20 =

0.50 0.05 0.00 n.s.

n.s.

0.07 0.00

0.20 0.01 0.15

n.s.

1.10

0.40

banned 1.4 n.s. n.s. no logs

no logs

(continued)

FSC + PEFC certified, 166 Mio m3 from plantations Wood is abundant, prices are low, no incentives for management Harvest in planted forests is greater 30–50 % of official production is illegal Overmature stands, industry sector underdeveloped Illegal production is three to four times higher Nonrenewals of certificates, lack of price premium Forests are considered as common goods, no political priority Log export not permitted Lack of interest, gold-mining has priority Needs imports No demand for certified timber

No demand for certified timber

Resource exhausted by 2014 Logging ban

Tropical Forest Resources: Facts and Tables 37

a

none x

x pilot p

x

Fee

PES

PFE production nat. % Total 000 ha forest 226,706 26 0 0 360 14 890 31 – 70 24 0 0 50 50 10 3 620 20 0 0 2,000 25

Subtotals contain estimates; thus cross totals do not match exactly

Country Subtotal LACa Belize Costa Rica Cuba Dominican Republic El Salvador French Guiana Haiti Jamaica Nicaragua Paraguay Total LAC

Table 8 (continued)

% – 30 70 – 85 0 – – 15 0

0.20 = 4.70 = 1.90  0.90 = 4.90 = 0.20 = 2.20 = 0.70  6.10 = 10.60 +

Fuelwood

Production Mio m3/ trend n.s. 0.20 – n.s. 0.02 n.s. – – n.s. 0.02

Mio m3

Export

38 J. Poker and K. MacDicken

Yes

50,000–400,000 Small None

Cancelled

Liberia Nigeria Togo Subtotal Africa Angola Benin Burundi Equatorial Guinea Gambia Guinea

Guinea Bissau

Small

50,000–600,000 Abolished

Gabon Ghana 25 1–3

40

15–25 15–20

Mean 400,000 >25,000 Small Small

200,000 42,000–475,000

1

1 1

1

Harvest permits Size (ha) Years

Country Cameroon Central African Rep. Congo, Democratic Repblic Congo, Republic of Côte d’Ivoire

Concessions

Table 9 Management in production forests

Yes

60–90 No

Yes Yes

60–80

Yes

Min dia/sp Yes Yes

No

Yes

No

Yes Yes, seed trees

Partly, seed trees Yes Partly

Standards Yes Yes

Planned

RIL

Harvest system

High grading

Chainsaw Chainsaw

Chainsaw

High grading

Other

No

Yes

Enrichment planting Yes

(continued)

In pilot phase Yes

Timber tracking system

Tropical Forest Resources: Facts and Tables 39

30–100 30

Min 35

None

Up to 2,555,000

Max 5,000

Malaysia Myanmar

Papua New Guinea

Philippines

25

35

20–30 10–30

25

3–5 50

Years

None In revision

Suspended

20

5–20

>24,000

>250 1,000– > 5,000

Guyana Honduras

Mexico Panama

30–40 7 cm, including branches Total aboveground volume of a tree without branches Stem volume without branches Thin branches and twigs 4 times palm of hand Crack 1 m Insect damage Cancer < ½ of stem circumference Cancer > ½ of stem circumference Fungi Broken crown Brittle branches Lianas Strangler fig

Location of defect Root Stem Trunk Crown

Cause of defect Harvesting Skidding Human induced Game Insects Fungi Fire Rockfall Landslides Wind, storm Sun

Age Where management reports of tree planting or sowing are available, tree age can easily be determined. In temperate zones, information on tree age can be achieved through counting annual rings. Seasonal patterns in the number of cells, cell size, and cell shape laid down in the wood make it relatively easy to discern annual rings in the wood. Several difficulties can arise in the determination of tree age. Seedlings and young trees may have endured suppression for periods of time which can vary greatly in length, from a year to several decades, where young trees are suppressed. An increased illumination caused by the opening of the canopy results in rapid growth in height until the tree reaches the intermediate canopy. During this phase, growth in diameter remains relatively modest. Only after the tree has developed a full crown does growth in diameter begin to accelerate. For trees with this growth pattern, an age determination based on size would not only be extremely difficult but in all likelihood also biased. Consequently age is often limited to indicate the period of time a tree has been part of the upper canopy layer and how long it has taken to grow from 1 diameter class into the next. In tropical forests, an estimate of age can be provided by repeatedly measuring the same trees (Vanclay 1996). Reliable and useful estimates require selection of a sufficient number of trees in each size class. The diameter is measured accurately to the nearest millimeter with a tape. It has proved useful to mark the level at which the diameter was last measured with paint, so as to minimize measurement errors. The measurements should always be conducted at the same time of year and should cover a period of at least 5 years. It is widely thought that cambial growth in tropical trees is continuous and generally not sensitive to climatic variation (Richards 1996); in general, tropical

Measurements and Assessments on Field Plots

729

trees should not produce reliable ring chronologies. Worbes (1992), Worbes and Junk (1989), and Devall et al. (1995) provide a historic perspective of this viewpoint. A corollary to this viewpoint has been that due to seasonal variation in temperature and soil water, many woody temperate species interrupt cambial activity and add a morphologically distinct xylem layer which results in a discontinuity between annual growth increments. Even though the seasonal variation is not distinct, most tropical locations show intra-annual variation with respect to humidity, rainfall, and solar radiation (Richards 1996). Eckstein et al. (1981) discuss the problems of dendrochronology and the prospects for tree-ring analyses. Tropical tree-ring chronologies have been described among others by Pumijumnong et al. (1995) for Teak (Tectona grandis) in northern Thailand; Vetter and Botosso (1989) for Amazonian trees; Devall et al. (1995) for Equator laurel (Cordia alliodora), Barrigon (Pseudobombax septenatum), and Cherimoya (Annona spraguei) in central Panama; Palmer and Murphy (1993) for Teak in Java, or Groenendijk et al. (2014) for several tree species in Central Africa. However, Xing et al. (2012) report on tree species that did not exhibit tree-ring structures. Worbes and Junk (1989) used radioactive and stable isotopes in estimating age. Bomb-produced C14 seemed to be the most promising method for dating between 1955 and the present.

Tree Crown The tree crown consists of the mass of branches and foliage growing outward of the stem of a tree. Common crown measurements are the crown diameter, crown length, and crown ratio. Crown width is generally measured by any two perpendicular diameters in circular crowns or the longest width and the longest perpendicular width (Fig. 33). The crown length is the distance from the tree’s tip to base crown. The crown ratio is given by dividing the crown length by the total tree height (Fig. 34). In addition to the measures of the size of crown, attributes can be assessed that describe crown form and crown level (Figs. 35 and 36).

Attributes Describing Forest Stands The forest stand concept is originating from sustainable forest management and is used to divide a forest district into several management units. A forest stand is defined as a group of trees that are stocked on a given area and share common characteristics such as age, species, or dimension, are subject to a specific silvicultural treatment, and thus display homogeneous structures. When describing natural forest ecosystems, the concept of forest stands applies only to a limited extent. Due to interrelated processes and patterns between abiotic and biotic factors, functional levels, material and energy fluxes, or community diversity and complexity, natural forests show heterogeneous structures with

M. Ko¨hl and M. Marchetti

730

Longest diameter

Longest cross diameter

Fig. 33 Crown diameter measurements

Fig. 34 Crown length

Crown length Height to base of crown

Total tree height

Crown width

small-scale variability (Lamprecht 1989). Hence, the assessment of natural forests is focusing on attributes describing the ecological functioning and integrity, while the assessment of managed forests gives a priority to timber production and sustainable yield.

Measurements and Assessments on Field Plots

731

1: Dominant – crown extended above general level of crown cover – Receive light form every side 2: Codominant – General level of the crown cover – Receive full light from above,only partial from side

1

2

2 3 4 4

3: Intermediate – Shorter than dominant and codominant – Penetrating the crown cover – Receive direct light only from above 4: Suppressed – Crown entirely below general level of crown cover – No direct light from above

Fig. 35 Tree crown level

Symmetric

Slightly one-sided

One-sided

Fig. 36 Crown form

Geographic Location and Topography The assessment of geophysical attributes provides auxiliary data that can be used for the categorization of inventory results and advanced analyses on ecological and economic aspects.

Geographic Location The geographic location is used to identify a spot on the Earth’s surface. It can be given in relative (i.e., 6.3 km north of Manaus) or absolute terms (2 420 55.270800 S, 60 50 32.884800 W). An absolute location uses a paired value in a Cartesian coordinate grid such as the Universal Transverse Mercator (UTM) or the

732

M. Ko¨hl and M. Marchetti

World Geodetic System (ESRI). The geographic location is an important interface to other data sets, such as digital terrain models, road infrastructure, topography, climatic data, or growth zones. It is of uttermost importance in any analysis utilizing spatial interrelationships such as forest industries feasibility studies, accessibility, transport, hydrology, or climatic zoning. Nowadays it is good practice to assess a geographical position by using global satellite navigation. The Global Positioning System (GPS) is a satellite-based navigation system made up of a network of 24 satellites placed into orbit by the US Department of Defense. GPS was originally intended for military applications but was made available for civilian use in the 1980s. GPS works in any weather conditions, anywhere in the world, 24 h a day. GPS satellites circle the earth twice a day in a very precise orbit and transmit signal information to earth. GPS receivers take this information and use triangulation to calculate the user’s exact location. There is an increasing demand for use of the GPS in the forest environment. The main applications are for surveying, managing, harvesting, and fire control. The GPS represents a very powerful tool to support forest owners and forest practitioners. Nevertheless, many factors can cause errors in position calculation. They are (Elfick et al. 1994, pp. 335–336) satellite and receiver clock errors, satellite ephemeris errors, errors due to atmospheric conditions, receiver errors, multipath errors, and errors in centering the antenna over a station and measuring its height. Multipath propagation of GPS signals is one of the main sources of GPS errors in the forest as well as the tree canopies may block satellite signals. Canopy density varies from location to location, and its impact on GPS performance is different from location to location. The tree canopy is one of the most important obstacle for using GPS in the tropical forest. Dense forest canopy posed a significant physical barrier to quality GPS signal reception (Dominy and Duncan 2001). Dominy and Duncan (2001) suggest to use a satellite subscription service for producing differential GPS measurements, rather than using a base station. A base station is an additional GPS receiver with an antenna stationed in an area that has an unobstructed view of the sky and is relatively close to the study site (at least within 1,500 km or 250 km for the highest accuracies). A base station operates simultaneously, while the rover unit is collecting field data, and allows to remove unwanted error inherent in the GPS signal. This is referred to as post-processed differential correction, because the corrections are applied after field data collection is completed. Remote sensing, GPS, and geographic information system (GIS) are the three most important spatial information technologies (Apan 1999). Combining remote sensing and GIS often produces synergistic effects. Remote sensing and GPS can be viewed as data capture or data acquisition facilities, while GIS is more of a data manipulation and analysis tool. GIS data could be used in improving satellite image classification (through the use of DEM and other ancillary data), while satellite images could be used as a primary information source. GIS, when combined with up-to-date data from a remote sensing system, can assist in automation of interpretation, change detection, map compilation, and map revision functions (Ehlers 1990).

Measurements and Assessments on Field Plots

733

Table 7 Relief (After Keller 2005) Category Flat terrain

Explanation Slope 10 %: plateau, terrace, valley plain

Terrace

Ridge

Convex form, edge, rim

Hillside

Inclined plane

Foot of the slope

Transition area of slope and plane

Topographic Situation The topographic situation describes the shape of the land and is captured by relief, aspect, and slope. Relief are elevations and depressions of the surface. It can be represented by digital elevation models or by contour lines and shading in maps. The assessment in the field is done by categories (see Table 7 as an example). Aspect is the compass direction that a topographic slope faces. It is usually measured in degrees from north with a compass. The slope is the rise or fall of the land surface (Fig. 37) and is expressed as a ratio or in percent. It is measured with a clinometer. Slope ¼ Slope in % ¼

height difference ½m horizontal distance ½m

height difference ½m  100 horizontal distance ½m

(38)

(39)

Elevation is the vertical distance between two points. Most countries have established a national network of bench marks with officially registered elevations. All bench mark heights are given in relationship to the one national datum plane which is generally the mean sea level. The altitude is the vertical distance to the national datum (i.e., sea level).

M. Ko¨hl and M. Marchetti

734 Fig. 37 Topography

p Slo e altitude

elevation

Sea level

Stand Structure Stand structure is the distribution of species and tree sizes within a forest area (Hush et al. 2003). The most important attributes to characterize stand structure include tree height, dbh, basal area, age, or volume and density aggregated to the stand level. In managed forests, a broad classification of stands is according to species composition (single-species vs. mixed species stands) and age (evenaged vs. uneven-aged stands) and their regeneration form (coppice vs. high forest). These stands can be characterized by a single or by vertically structured canopy layers. The maintenance and enhancement of species diversity and other aspects of structural diversity became a major objective in sustainable forest management and biodiversity studies (Otto 1994). Species composition has three different components (Hush et al. 2003): Abundance, the number of species in a forest area Frequency, the number of sampling units in which a species can be found Dominance, the size of individuals in a population Diversity indices combine information on frequency, abundance, and dominance and offer a tool for describing species variability on a site (Ricotta et al. 2003). Neumann and Starlinger (2001) comment that “methods applied in assessing different types of diversity are as manifold as the ways of calculating measures of diversity.” Most diversity indices are based on neighborhood methods (Pielou 1975). They can easily be applied in field surveys and provide measures that allow comparisons of different sites or the development over time. Examples of those indices are the Shannon index, Simpson index, or Pielou index (Magurran 1988), or the more recently published forest variability indices (Das and Nautiyal 2004), the uniform angle index (Gadow et al. 1998) and the tree species spatial diversity (Hui et al. 2011).

Measurements and Assessments on Field Plots

735

Where decisions are guided by forest variability indices, the understanding of those indices becomes crucial. Indices for species diversity and stand structure have been compared in several studies (Franc 1998; Gleichmar and Gerold 1998; Ko¨hl and Zingg 1996; Magnussen and Boyle 1995; Neumann and Starlinger 2001). Ko¨hl and Zingg (1996) found inconsistent estimates for changes of species diversity when indices are applied at successive occasions. Neumann and Starlinger (2001) compare species diversity and stand diversity indices in a wide range of ecological conditions in Austria. They found many correlations between the indices and arrive at the conclusion “that the method of calculation is not so important” given that a consistent set of elements of a forest is studied. Beside the spatial structure of a forest ecosystem, monitoring species diversity becomes an important objective in forest ecosystems. Random or systematic allocation of sampling units involves the risk of missing rare species. Thompson (1992) describes adaptive cluster sampling as an efficient method for the assessment of species with low abundance. Adaptive cluster sampling allocates sampling units in two steps: 1. An initially fixed number of sample plots are randomly or systematically distributed over the sampling area. 2. In each plot where rare species are found, the neighboring plots are measured. In any of these additional plots where the species of interest is found, another set of neighboring plots is established. The procedure is continued until no further plots are found where the species of interest occurs. The procedure is an efficient approach to sample rare species (Roesch 1993), but operational applications of adaptive cluster sampling are few due to unknown a priori sample sizes and cumbersome field assessments. In natural tropical forests, the procedure might not be efficient due to the pronounced small-scale variability of forest species.

Disturbances Forests are subject to a wide range of damages which can be of biotic or abiotic origin and natural or human induced. Damages shape landscapes and forest systems by influencing their composition, structure, and functional processes. Disturbance like storm, fire, or pests are major agents of forest dynamics, and apart from competition and senescence, they are the main causes of tree death (Marchetti 2004; Oliver and Larson 1996). Damages reduce forests health and vitality, leading to environmental as well as economic losses. Biotic agents include insect, phytopathogens, and wild animals. Insects can be grouped into defoliators (sucking, mining, galls) or borers (stem, bark, branch, twig, fruit, and root) and phytopathogens into bacteria, viruses, and fungi. Both groups show long latent periods, which might convert into severe outbreaks due to changing environmental conditions. Wildlife induces mechanical damages (e.g., rodents,

736

M. Ko¨hl and M. Marchetti

browsing) or defoliation. For an overview on tropical forest and their pests, see Nair (2007) and Coley and Kursar (2014). Abiotic agents can be assigned to the following five categories (Moore and Allard 2011): (I) Meteorological: cyclones, storms (wind, snow, ice and hail, dust, and sand), tornadoes, and thunderstorms and lightning (II) Climatological: drought (III) Hydrological: floods and flash floods, avalanches, landslides, and mudslides (IV) Geophysical: tsunamis, earthquakes, and volcanic eruptions (V) Anthropogenic: fire, oil spills, air pollution, and radioactive contamination All the abovementioned disturbances affect environment differently, changing the provisioning of ecosystem goods and services offered by the forests. Biotic and abiotic agents have considerable impacts on forests: they can significantly reduce wildlife habitat, thereby reducing local biodiversity and species richness. Economic and social impacts on the forest sector are also remarkable; indeed, damages can reduce tree growth and the yield of wood and non-wood products. Anthropogenic and natural disturbances are strictly related: they influence each other. The distinction between anthropogenic and natural disturbances is very blurred (Chazdon 2003). Anyway, it is stated that the most relevant human-induced causes of damages are improper logging activity and climate change. These factors have effects on species composition and forest structure. In last decades, damages dynamics have been adversely affected by climate change; it strongly influences disturbances by altering the frequency, intensity, and timing of fire events, hurricanes, storms, landslides, and insect and disease outbreaks. Damages indirectly affect management practices; some pests have necessitated changes in management regimes often forcing forest managers to switch to alternative tree species in plantations. Adaptive forest management involving all sectors and stakeholders is therefore essential to protect the world’s forest resources. And since such disturbances do not respect borders, regional and international cooperation is urgently required (Moore and Allard 2011). The assessment of disturbances has to address a symptom description, the damage cause or diagnosis, and the quantification of damage extent (Table 8).

Deadwood Deadwood (coarse woody debris) is recognized as an important component of forest ecosystems (Delaney et al. 1998) and represents an important part of the carbon pool in many forests. Deadwood in form of snags (dead standing trees) and logs (dead lying trees) is a proxy indicator for invertebrate biodiversity, since it is a habitat for a wide array of organisms. Behind the humidification, deadwood became an important component of forest soil. Deadwood decomposition plays a key role in the recycling of nutrients and organic matter as well as the creation of a wide variety

Measurements and Assessments on Field Plots

737

Table 8 Assessment of disturbances Symptom description

Damage causes (diagnosis)

Quantification of extent

Affected parts Leaves Branches Bark Fruits Root Symptom Devoured Holes Game and grazing Insects Fungi Abiotic agents Human induced Fire Atmospheric pollutants Other factors Unidentified Quantity (%, area) of the affected part of the tree Frequency

of micro sites for the regeneration of plant species and the creation of a wide variety of habitats for other organisms. The amount of deadwood is an excellent indicator of the conservation value of a forest, and its occurrence together with the forest structure is an indicator of old-growth forest conditions (Lombardi et al. 2012). Nevertheless, the amount of deadwood in tropical forests is poorly quantified and extremely variable depending upon forest age and climatic regime (Saldarriaga et al. 1986; Uhl et al. 1988; Uhl and Kauffman 1990). Lack of data on this significant forest component obviously can lead to underestimates of the total amount of biomass in a forest. According to Global Forest Resources Assessment (FAO 2010), the amount of deadwood in tropical forest can be quantified as 22 t/ha (Fig. 38). Deadwood includes all nonliving woody biomass not contained in the litter, either standing, bending, or lying on the ground or in the soil (FAO 2004). Deadwood includes wood lying on the surface, dead roots, and stumps larger than or equal to 10 cm in diameter and greater than 1 m in length. Deadwood can be measured through the following instruments: (i) GPS or handheld compass, (ii) height measurement tool, and (iii) measuring tape. The main parameters to measure are the dbh and tree height of each type of deadwood (standing, bending, lying, etc.). All deadwood elements have to be classified according their status density (e.g., sound, intermediate, rotten). To determine the density class of a piece of deadwood, strike each piece with a machete.

M. Ko¨hl and M. Marchetti

738

Fig. 38 Deadwood (left, logs; right, snag)

Plot centre

1 lying, brittle; measure dbh 2 dead standing tree, measure dbh 3 stump of felled tree 4 dead standing tree, measure dbh 5 lying, brittle; measure dbh 6 stump of felled tree 7 lying, brittle; measure dbh 8 lying, brittle; measure dbh 9 brittle branches, dbh ≥ 7cm 10 brittle branches, dbh ≥ 7cm 11 dead standing tree, measure dbh 12 log, dbh ≥ 7cm

Fig. 39 Assessment of deadwood on a sample plot (After Keller 2005)

Deadwood information is collected in the field either by assessing woody debris on plots (Bo¨hl and Bra¨ndli 2007) (Ligot et al. 2012) or by linear transects (Ringvall and Sta˚hl 1999; Sta˚hl 1998). Deadwood on the plot level is defined as the total volume of coarse woody debris as well as of all the standing and laying dead trees. For the estimation of deadwood assessed by transects, see Sta˚hl (1998), Figs. 39 and 40. Rondeux et al. (2012) presented a system of nomenclature for dead woody debris that includes different stages of decay (Table 9). Where the assessment of deadwood is associated with an assessment of the causing agent for mortality, the inventory can provide significant insight in

Measurements and Assessments on Field Plots

739

Fig. 40 Assessment of deadwood on a transect (After Keller 2005)

transect

diameter lengt

h

Table 9 Deadwood elements (From Rondeux et al. 2012) Deadwood elements Living and dead stems Standing and lying stems Decay classes

Stem volume of dead trees

Piece of coarse woody debris

Volume of coarse woody debris

Definition A living stem has active or dormant cambium; otherwise the stem is dead A lying stem is the main stem that is not self-supporting with the majority of its length lying on the ground; otherwise it is a standing stem Four decay classes (A, B, C, and D) are considered on the basis of the percentage of hard texture wood present in the deadwood volume. Wood is considered “hard texture” if a knife cannot penetrate more than 2 cm Class A: hard texture 90 % (not decayed, completely hard) Class B: hard texture 90–60 % (slightly decayed, most part still hard) Class C: hard texture 60–30 % (decayed, most part soft) Class D: hard texture 30 % (very decayed, completely soft) The stem volume of dead trees is the aggregated aboveground volume of all dead stems, standing or lying, over a specified area. Included are volumes – from the stump height to a top diameter of 10 cm – of dead stems with a dbh of 10 cm. Branches are excluded A piece of coarse woody debris is a downed (not suspended) piece of deadwood lying on ground, with sections coarser than 10 cm (over bark) of at least 1 m in length. Lying dead stems, including attached branches, are excluded The volume of coarse woody debris is the aggregated aboveground volume of all pieces of coarse woody debris over a specified land area. Included are those sections of the coarse woody debris pieces that are coarser than 10 cm (over bark) on a length of at least 1 m

damages due to logging activities. In addition to the assessment, it is advisable to conduct a stump inventory. The diameter of stumps can be related to diameters in respective height of standing trees and provides information on extracted timber volumes (Loetsch et al. 1973).

740

M. Ko¨hl and M. Marchetti

Assessment of Forest Area Information about the forest area and forest area proportion presents a major result of forest inventories for the following reasons: 1. Information about the forested area itself is of major interest for, e.g., forest policy, management planning, and nature conservation. Detailed accounts of subareas are frequently requested for location-specific management purposes, i.e., the absolute size and proportions of forest types, ownership categories, or age classes. 2. Many attributes are presented in terms of unit area, such as the growing stock per hectare. Area-related ratios serve to standardize and facilitate comparisons over time and between different units of reference. 3. Quantifying forest area changes is especially important in regions with strong land-use changes and forest area dynamics. Information on area can be expressed in two different ways: (1) the total forest area and (2) the proportion of forest area within a given region. Data sources for area information are analog or digital maps, aerial photographs, digital remote sensing data, or terrestrial surveys. Information on the size of an area can be provided by two techniques: (1) the measurement of an area or (2) the estimation of an area by sampling. Where digital maps or wall-to-wall remote sensing imagery is available in digital format, GIS offers an efficient technique for calculating areas (Bolstad 2012). Where analog maps or aerial photographs are available, areas may be measured by means of area calculations, weighing, planimetry, or counting the number of cells in a square or parallel grid that covers the area of interest. Area measurements can be very time-consuming and impractical for large-scale inventories. Especially where large-scale inventories are carried out, areas of concern should not be delineated on the basis of the aerial photographs. The reliability of area estimates provided by measurement techniques depends on the errors given by the method, the instruments applied, the process of execution, and the particularity of the staff involved. Knowledge of the positional error of a line (boundary) and the process involved in the delineation of objects of interest allows a model-based estimation of the error of a polygon drawn by this process. Magnussen (1994) provides a model for estimating the area of a forest stand when the stand is delineated by photo interpretation. Sampling techniques for area estimation may be based on maps, aerial photographs, digital classification, or field surveys. Where maps and aerial photographs are used, it must be ensured not only that the latest data are verified but it must also be recognized that precision will depend on the scale at which the sampling is done. As a general rule: the larger the scale, the better is the accuracy. Forested areas are often only found on maps when they exceed a certain area. Small patches of forest area may consequently not appear which leads to underestimation of the total forest

Measurements and Assessments on Field Plots

741

area. Only maps with a high dimensional stability should be employed or, even better, recent aerial photographs, if available. Point sampling for area estimation is based on the concept of random points: a point is chosen at random from the possible locations, and a value is assigned to the random point, e.g., it is either forest or non-forest. In practical applications, point sampling is realized by applying dot grids, where the dots are considered to be a realization of a random process vis-a-vis the map. The localization of one point corresponds to a Bernoulli experiment with the possible values of non-forest and forest. The binomial distribution describes the probability of all possible outcomes of the random sampling design completely. In the estimation of the total area, it is helpful to note that the number of sample locations with forest (or non-forest) is asymptotically (when both sample size n and population size N goes toward infinity) a realization of a Poisson process with a density λ where each point represents on average an area of 1/λ. An estimator for the total forest area AF is AF ¼ N=λ

(40)

If nF out of n random points are found in the forest, λ = nF/n is an unbiased estimator of the true forest density (viz., proportion p). The forest area proportion is estimated according to Cochran (1977) by: ^λ ¼ ^p ¼ nF n ^p^q n pffiffiffiffiffiffiffiffiffi ^sp ¼ vð^pÞ

vð^pÞ ¼ s2p ffi

(41) (42) (43)

where ^ p ¼ forest area proportion ^ q ¼ 1  ^p vð ^ pÞ ¼ variance of ^p: Sp ¼ standard error of p^: n ¼ number of all dots on the dor grid inside the area A of interest nF ¼ number of forest dots on the grid: The area proportion always has to be seen in relation to the total area, e.g., a forest patch of 100 ha leads to a forest area proportion, p^ ¼ 0:1 in a total area of 1,000 ha and p^ ¼ 0:5 in a total area of 200 ha. In practical applications, the total forest area AF is often estimated by multiplying the (known) total area A with the estimated forest area proportion,

M. Ko¨hl and M. Marchetti

742

^ F ¼ nF A ¼ A^p A n

(44)

ˆ F) and the standard error s(A ˆ F) with the variance v(A   ^ F ¼ A 2 sp 2 v A

(45)

    ^ F ¼ √v A ^F s A

(46)

If the estimation equations shown above are used for a systematic dot grid, as opposed to a random sampling of dot locations, the standard error is generally overestimated. The form and the spatial distribution pattern of the forest areas influence the amount of overestimation.

References Aiba S, Kitayama K (1999) Structure, composition and species diversity in an altitude-substrate matrix of rain forest tree communities on Mount Kinabalu. Borneo Plant Ecol 140:139–157 Apan AA (1999) GIS applications in tropical forestry. Faculty of Engineering and Surveying, University of Southern Queensland, Toowoomba, p 132 Bachmann P, Ko¨hl M, Pa¨ivinnen R (1998) Assessment of biodiversity for improved forest planning. Kluwer, Dordrecht Bailey RL (1979) The potential of Weibull-type functions as flexible growth curves: discussion. Can J Forest Res 10:117–118 Batalha MA, Aragaki S, Mantovani W (1998) Chave de identificacao das especies vasculares do cerrado em emas (pirassununga, sp) baseada em caracteres vegetativos. Bol Bot Univ Sao Paulo 17:85–108 Batista JL, Couto HTZ, Marquesini M (2001) Performance of height-diameter relationship models: analysis in three forest types. Scientia Forestalis 60:149–163 Bitterlich W (1962) Genauere Umrechnungsfaktoren Raummaß-Festmaß durch Winkelza¨hlproben. Holzkurier 48, p 11 Bo¨hl J, Bra¨ndli U-B (2007) Deadwood volume assessment in the third Swiss National Forest Inventory: methods and first results. Eur J For Res 126(3):449–457 Bo¨hm M, Lo¨w R, Haag J, Kerwer A, L€ uttge U, Rausch T (1993) Evaluation of comparative DNA amplification fingerprinting for rapid species identification within the genus Clusia. Bot Acta 106(5):448–453 Bolstad PV (2012) GIS fundamentals: a first textbook on geographic information systems, 4th edn. St. Paul, Eider Press Brown S (1997) Estimating biomass and biomass change of tropical forests: a primer. FAO forestry paper, Rome Bruce D, Max TA (1990) Use of profile equations in tree volume estimation. In: LaBau J, Cunia T (eds) USDA Forest Service, Portland, pp 213–220 Br€unig EF (1973) Species richness and stand diversity in relation to site and succession of forests in Sarawak and Brunei (Borneo). Amazoniana 4(3):293–320 Chave J (2006) Measuring wood density for tropical forest trees a field manual. U.P.S. Lab. Evolution et Diversité Biologique (ed) http://www.eci.ox.ac.uk/research/ecodynamics/ panamazonia/wood_density_english.pdf (online publication) Chave J, Andalo C, Brown S, Cairns MA, Chambers JQ, Eamus D, Fo¨lster H, Fromard F, Higuchi N, Kira T, Lescure J-P, Nelson BW, Ogawa H, Puig H, Riéra B et al (2005) Tree

Measurements and Assessments on Field Plots

743

allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145:87–99 Chazdon RL (2003) Tropical forest recovery: legacies of human impact and natural disturbances. Perspect Plant Ecol Evol Syst 6(1–2):51–71 Cochran WG (1977) Sampling techniques, 3rd edn. Wiley, New York Coley PD, Kursar TA (2014) On tropical forests and their pests. Science 343:35–36 Comvalius LB (2001) Surinamese timber species: characteristics and utilization. Celos, Paramaribo Connell JH (1971) On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. In: Boer P, Gradwell G (eds) Dynamics of populations. PUDOC, Wageningen, pp 298–312 Curtis RO (1967) Height-diameter and height-diameter-age equations for second-growth Douglasfir. Fort Sci 13:365–375 Czaplewski RL (1989) Graphical analysis of stem taper in model building. Can J For Res 19:522–524 Czaplewski RL, Brown AS, Walker RC (1989) Profile models for estimating log end diameters in the Rocky Mountain Region. USDA Forest Service, Research Paper RM-284, Ft. Collins p 7 da Silva Scaranello MA, Ferreira Alves L, Aparecida Vieira S, Barbosa de Camargo P, Joly CA, Martinelli LA (2011) Height-diameter relationships of tropical Atlantic moist forest trees in southeastern Brazil. Sci Afr 69(1):26–37 Das JK, Nautiyal J (2004) Forest variability index: a vector quantifying forest stand diversity and forest compactness. For Policy Econ 6:271–288 Delaney M, Brown S, Lugo AE, Torres-Lezama A, Quintero NB (1998) The quantity and turnover of dead wood in permanent forest plots in six life zones of Venezuela. Biotropica 30(1):2–11 Devall MS, Parresol BR, Wright SJ (1995) Dendrochronological analysis of Cordia alliodora, Pseudobombax septenatum and Annona spraguei in central Panama. IAWA J 16:411–424 Dominy NJ, Duncan B (2001) GPS and GIS methods in an African rain forest: applications to tropical ecology and conservation. Conserv Ecol 5(2) http://www.consecol.org/vol5/iss2/art6/ UN ECE/FAO (2000) Forest resources of Europe, CIS, North America, Australia, Japan and New Zealand. Geneva Eckstein D, Ogden J, Jacoby GC, Ash J (1981) Age and growth rate determination in tropical trees: the application of dendrochronological methods. In: Age and growth rate of tropical trees, new directions for research. Yale University, New Heaven, pp 63–106 Ehlers M (1990) Remote sensing and geographical information systems: towards integrated spatial information processing. EEE Trans Geosci Remote Sens 28(4):763–766 Eichhorn F (1904) Beziehunen zwischen Bestandesho¨he und Bestandesmasse. Allg Forst- und Jagdz 80:45–49 Ella AB, Escobin RP (1993) Taxonomy and wood anatomy of the manggasinoro species (shorea spp.): Dipterocarpaceae. Philipp J Sci 122(3):205–232 Elfick M, Fryer J, Brinker RC, Wolf PR (1994) Elementary Surveying: SI Adaptation, Pearson Education Ltd, Harlow, p. 528 ESRI (2014) Coordinate systems, map projections, and geographic (datum) transformations. http:// resources.esri.com/help/9.3/arcgisengine/dotnet/89b720a5-7339-44b0-8b58-0f5bf2843393.htm. Accessed 12 June 2014 Fang Z, Bailey RL (1998) Height-diameter models for tropical forests on Hainan Island in southern China. For Ecol Manage 315–327(110):315–327 FAO (2004) Global forest resources assessment update 2005: terms and definitions. Working papers 83/E, Rome FAO (2010) Global forest resources assessment 2010: main report. Food and Agriculture Organization of the United Nations, Rome Feldpausch TR, Banin L, Phillips OL, Baker TR, Lewis SL, Quesada CA, Affum-Baffoe K, Arets EJMM, Berry NJ, Bird M, Brondizio ES, de Camargo P, Chave J, Djagbletey G, Domingues

744

M. Ko¨hl and M. Marchetti

TF et al (2011) Height-diameter allometry of tropical forest trees. Biogeosciences 8:1081–1106 Franc A (1998) Some mathematical remarks on forest biodiversity. In: Bachmann P, Ko¨hl M, Pa¨ivinen R (eds) Assessment of biodiversity for improved forest planning. Kluwer, Dordrecht, pp 159–169 Freese F (1974) A collection of log rules. Madison, Wisconsin, p 65 FVA-Waldnutzung (1997) Stichprobenverfahren zur Rundholzvermessung, 6 June 2014 Gadow KV, Hui GY, Albert M (1998) Das Winkelmaß – ein Strukturparameter zur Beschreibung der Individualverteilung in Waldbesta¨nden. Cbl Ges Forstw 155:1–10 Gallant MN (1939) Classification of teak logs. FRI, India, p 28 Gifford R (2000) Carbon content of woody roots: revised analysis and a comparison with woody shoot components. n. National carbon accounting system technical report. Canberra Giordano G (1976) Tecnologia del legno. UTET, Torino Gleichmar W, Gerold D (1998) Ndizes zur Charakterisierung der horizontalen Baumverteilung. Forstwiss Centralbl 117:69–80 Groenendijk P, Sass-Klaassen U, Bongers F, Zuidema PA (2014) Potential of tree-ring analysis in a wet tropical forest: a case study on 22 commercial tree species in Central Africa. For Ecol Manage 323:65–78 Grubb PJ (1977) Control of forest growth and distribution on wet tropical mountains: with special reference to mineral nutrition. Annu Rev Ecol Syst 8:83–107 Hohenadel W (1924) Der Aufbau der Baumscha¨fte. Forstwiss Centralbl Hohenadel W (1936) Die Bestandesmessung. Forstwiss Centralbl 58:51–61, 69–86, 114–127 Huang S, Titus SJ (1992) Comparison of nonlinear height-diameter functions for major Alberta tree species. Can J Forest Res 22:1297–1304 Hui G, Zhao X, Zhao Z, Gadow KV (2011) Evaluating tree species spatial diversity based on neighborhood relationships. For Sci 57(4):292–300 Hush B, Beers TW, Kershaw JA (2003) Forest mensuration. Wiley, New York IUFRO (1959) The standardization of symbols in forest mensuration. University of Maine, Orono, p 32 Janzen DH (1970) Herbivores and the number of tree species in tropical forests. Am Nat 104:501–528 Kaufmann E (2002) Estimation of standing timber, growth and Cut. Swiss National Forest Inventory: methods and models of the second assessment. Swiss Federal Research Institute WSL, Birmensdorf, pp 162–196 Keller R (1996) Identification of tropical woody plants in the absence of flowers and fruits. A field guide. Birkha¨user Verlag, Basel/Boston/Berlin Keller M (2005) Schweizerisches Landesforstinventar: Anleitung f€ ur die Feldaufnahmen der Erhebung 2004–2007. Birmensdorf, p 393 Kera¨nen M, Aro E-M, Tyystja¨rvi E (2003) Automatic plant identification with chlorophyll fluorescence fingerprinting. Precis Agric 4:53–67 King DA (1996) Allometry and life history of tropical trees. J Trop Ecol 12:25–44 Ko¨hl M, Zingg A (1996) Eignung von Diversita¨tsindizes bei Langzeituntersuchungen zur Biodiversita¨t in Waldbesta¨nden. Allg Forst- u J-Zeitung 167(4):76–85 Ko¨hl M, Magnussen S, Marchetti M (2006) Sampling methods, remote sensing and GIS multiresource forest inventory. Springer, Berlin/Heidelberg Kollmann FFP, Cote` WA (1968) Principles of wood sciences. Springer, New York Kublin E, Breidenbach J (2013) Package “TapeR” http://cran.r-project.org/web/packages/TapeR/ TapeR.pdf (online publication) Kublin E, Breidenbach J, Kaendler G (2013) A flexible stem taper and volume prediction method based on mixed-effects B-spline regression. Eur J For Res 132:983–997 Kuzelka K, Marusàk R (2014) Comparison of selected splines for stem form modeling: a case study in Norway spruce. Ann For Res 57(1):137–148

Measurements and Assessments on Field Plots

745

Lamprecht H (1989) Silviculture in the tropics: tropical forest ecosystems and their tree species. Possibilities and methods for their long term utilization. GTZ Eschborn Ligot G, Lejeune P, Rondeux J, Hébert J (2012) Assessing and harmonizing lying deadwood volume with regional forest inventory data in Wallonia (Southern region of Belgium). Open Forest Sci J 3, 15–22 Loetsch F, Zo¨hrer F, Haller KE (1973) Forest inventory. BLV Verlagsanstalt, M€ unchen/Bern/ Wien, p 469 Lombardi F, Lasserre B, Chirici G, Tognetti R, Marchetti M (2012) Deadwood occurrence and forest structure as indicators of old-growth forest conditions in Mediterranean mountainous ecosystems. Ecoscience 19(4):244–355 Lorenzi H (2002) Árvores Brasileiras, Manual de identificac¸a˜o e cultivo de Plantas Arbo´reas Nativas do Brasil. Instituto Plantarum, Nova Odessa Luxmi C, Raturi RD, Rao RV, Dayal R (1995) Wood anatomy of Indian Flacourtiaceae. Indian Forester 121(9) Luxmi C, Raturi RD, Rao RV, Dayal R (1998) Identification of Indien bamboos using culm epidermal features – an overview. In: The proceeding of national seminar on processing and utilisation of plantation timbers and bamboo. IPIRTI, Bangalore, pp 66–73 Magnussen S (1994) A coordinate-free area variance estimator for forest stands with a fuzzy outline. For Sci 42:76–85 Magnussen S, Boyle TJB (1995) Estimating sample size for inference about Shannon  Weaver and the Simpson indices of species diversity. For Ecol Manage 78:71–84 Magurran AE (1988) Ecological diversity and its measurement. Croom Helm, London, p 179 Marchetti M (2004) Monitoring and indicators of forest biodiversity in Europe – form ideas to operationality. In: EFI proceedings 51, Joensuu, p 526 Max TA, Burkhart HE (1976) Segmented polynomial regression applied to taper equations. For Sci 22:283–289 Maydell VHJ (1983) Arbres et arbustes du Sahel, leurs caracteristiques et leurs utilisations. GTZ, Eschborn/Ts Moore BA, Allard G (2011) Abiotic disturbances and their influence on forest health. A review. Working paper, Rome Nair KSS (2007) Tropical forest insect pests – ecology, impact and management. Cambridge University Press, New York, p 404 Neumann M, Starlinger F (2001) The significance of different indices for stand structure and diversity forests. For Ecol Manage 145(91):91–108 Noack D (1971) Evaluation of properties of tropical timbers. J Inst Wood Sci 5(5):17–23 Oliver CD, Larson BC (1996) Forest stand dynamics. Wiley, New York Otto HJ (1994) Walso¨kologie. Ulmer, Stuttgart Paine CET, Stahl C, Courtois EA, Patino S, Sarmiento C, Baraloto C (2010) Functional explanations for variation in bark thickness in tropical rain forest trees. Funct Ecol 24:1202–1210 Palmer JG, Murphy JO (1993) An extended tree-ring chronology (teak) from Java. Proc Koninklijke Nederlandse Akademie van Wetenschappen-Biol Chem Geol Phys Med Sci 96:27–41 Parmentier I, Duminil J, Kuzmina M, Philippe M, Thomas DW, Kenfack D, Chuyong G, Cruaud C, Hardy OJ (2013) How effective are DNA barcodes in the identification of African rainforest trees? PLoS One 84(4) Pielou EC (1975) Ecological diversity. Wiley, New York Prodan M (1965) Holzmesslehre. Sauerla¨nders, Frankfurt Pumijumnong N, Eckstein D, Sass U (1995) Tree-ring research on Tectona grandis in northern Thailand. IAWA J 16:385–392 Rajora OP, Zsuffa O (1991) Screening populus deltoides Marsh. Selections by allozymes to assure species identity. Scand J For Res 6(1–4):471–478

746

M. Ko¨hl and M. Marchetti

Rejmánek M, Brewer SW (2001) Vegetative identification of tropical woody plants: state of the art and annotated bibliography. Biotropica 33(2):214–228 Richards PW (1996) The tropical rain forest, 2nd edn. Cambridge University Press, Cambridge Richter C (2014) Wood characteristics. Springer, Heidelberg Ricotta C, Corona P, Marchetti M, Chirici G, Innamorati S (2003) LaDy: software for assessing local landscape diversity profiles of raster land cover maps using geographic windows. Environ Model Software 18:373–378 Ringvall A, Sta˚hl G (1999) Field aspects of line intersect sampling for assessing coarse woody debris. For Ecol Manage 118:163–170 Roesch FA (1993) Adaptive cluster sampling for forest inventories. For Sci 39:655–669 Rondeux J, Bertini R, Bastrup-Birk A, Corona P, Latte N, McRoberts RE, Sta˚hl G, Winter S, Chirici G (2012) Assessing deadwood using harmonized national forest inventory data. For Sci 58(3):269–283 Roth I (1981) Structural patterns of tropical barks. Borntraeger, Berlin Ryan MG, Phillips N, Bond BJ (2006) The hydraulic limitation hypothesis revisited. Plant Cell Environ 29:267–281 Saldarriaga JG, West DC, Thorp ML (1986) Forest succession in the Upper Rio Negro of Colombia and Venezuela. Environmental Sciences Division, Oak Ridge Santander Flores C, Albertin W (1974) Performance of Dalbergia retusa Hemsl. in the humid tropics. Turrialba 24(1):76–83 Schmid-Haas P, Werner J, Baumann E (1978) Kontrollstichproben: Aufnahmeinstruktion. Bericht 186. Eidg. Anst. Forstl. Versuchswes, Birmensdorf, p 57 Sharma M, Zhang SY (2001) Variable-exponent taper equations for jack pine, black spruce, and balsam fir in eastern Canada. For Ecol Manage 198(1–3):39–53 Smaltschinski T (1983) Individuelle Baumschaftform und cubische Spline Interpolation [Individual taper curve of trees and cubic spline interpolation]. Allgemeine Forstund Jagdzeitung 155:193–197 Smith AP (1979) Buttressing of tropical trees in relation to bark thickness in Dominica. Biotropica 11(2):159–160 Sta˚hl G (1998) Transect relascope sampling – a method for the quantification of coarse woody debris. For Sci 44:58–63 Sterba H (1980) Stem-curves: a review of literature. For Abs 41:141–145 Tang S (1994) Self-adjusted height-diameter curves and one entry volume model. For Res 7:512–518 Thompson SK (1992) Sampling. Wiley, New York Uhl C, Kauffman JB (1990) Deforestation, fire susceptibility, and potential tree responses to fire in the eastern Amazon. Ecology 71:437–449 Uhl C, Buschbacher R, Serrao EAS (1988) Abandoned pastures in eastern Amazonia. I. Pattern of plant succession. J Ecol 76:663–681 Vanclay JK (1995) Growth models for tropical forests: a synthesis of models and methods. For Sci 41:7–42 Vanclay JK (1996) Estimating sustainable timber production from tropical forests. CIFOR working paper, p 17 Vetter RE, Botosso RC (1989) Remarks on age and growth rate determination of Amazonian trees. IAWA J 10:133–145 West PE (2009) Trees and forest measurement. Springer, Heidelberg Wilent S (2014) A look through the TruPulse 200L rangefinder. For Sour 19:15 Wilson AM, Bollandsa˚s OM, Eid T (2013) Relationships between diameter and height of trees in natural tropical forest in Tanzania. South For J For Sci 75(4):221–237 Worbes M (1992) Occurrence of seasonal climate and tree-ring research in the tropics. Lundqua Rep 34:338–342 Worbes M, Junk WJ (1989) Dating tropical trees by means of C-14 from bomb tests. Ecol Lett 70:503–507

Measurements and Assessments on Field Plots

747

Wright SJ (2002) Plant diversity in tropical forests: a review of mechanisms of species coexistence. Oecologia 130:1–14 Xing P, Zhang Q-B, Baker PJ (2012) Age and radial growth pattern of four tree species in a subtropical forest of China. Trees 26:283–290 Zingg A (1988) Schweizerisches Landesforstinventar, Anleitung f€ ur die Erstaufnahme 1982–1986. Eidgen. AnstaltForstl. Versuchswes., Birmensdorf

Objectives and Planning of Forest Inventories Michael Ko¨hl and Marco Marchetti

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives of Forest Inventories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Typology of Forest Inventories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inventory Planning Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of the Optimal Inventory Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sample Design Alternatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Defining the Sample Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Forest Area Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Target Population and Sample Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality Assurance and Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . QA Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . QC Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

750 752 756 759 762 763 765 765 766 766 767 767 770 770 771 773 775

M. Ko¨hl (*) Center for Wood Sciences, Institute of World Forestry, University of Hamburg, Hamburg, Germany e-mail: [email protected] M. Marchetti Department DIBT - BioScienses and Territory, University of Molise, Pesche, IS, Italy e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_70

749

750

M. Ko¨hl and M. Marchetti

Introduction Forest inventories, also called forest resources assessments, can be understood as “the procedure for obtaining information on the quantity and condition of the forest resource, associated vegetation and components and many of the characteristics of the land area on which the forest is located” (Hush et al. 2003). The term “forest inventory” refers to both, a catalog of information on forests and the measurement and assessment of data on which the information is based. Forest inventory forms the foundation of forest planning and forest policy. While early designs of sustainable forest management and forest inventory focused on timber production (Hartig 1819; Cotta 1804), modern forest inventory designs support a holistic view of forest ecosystems addressing not only timber production but the multiple functions of forests as well as the need to understand the functioning mechanisms of forest ecosystems (Lund 1998; Corona et al. 2003; Ko¨hl et al. 2006; Corona and Marchetti 2007). Forests have to be managed judiciously not only for environmental protection and other services but also for various products and industrial raw material. In some parts of the world, biological resources are being depleted faster than they can regenerate mainly caused by unsustainable management and illegal logging. Following the 1992 UNCED conference in Rio de Janeiro, considerable progress has been made in the area of sustainable forest management. Among others, the International Tropical Timber Organization (ITTO) and the Forest Stewardship Council (FSC) developed criteria and indicators for sustainable forest management and certification. The Kyoto Protocol of the United Nations Framework Convention on Climate Change (UNFCCC) described measures to mitigate greenhouse gasses and addressed in particular the impact of deforestation and afforestation on global climate change. The Convention on Biological Diversity (CBD) that was ratified in 1994 deals with the protection and maintenance of biodiversity (CBD 1995). Forest inventories facilitate a multifaceted analysis and study of forests not only as important source of subsistence, employment, revenue earnings, and raw materials to a number of industries but also for their vital role in ecological balance, environmental stability, biodiversity conservation, food security, and sustainable development of countries and entire biosphere (Corona and Marchetti 2007). Despite the fact that forest inventories are carried out for different purposes and in different environments, there are several general implications to be considered for any assessment: 1.

The information provided has to satisfy user needs. Therefore, the inventory objectives need to be defined by the joined efforts of inventory specialists and the parties involved (e.g., forest authorities, forest owners, wood-processing industries, land-use planning and environmental protection agencies, consumers of secondary forest products, wildlife organizations, or local societies).

Objectives and Planning of Forest Inventories

751

2. The information has to be objective. Objective information requires the objective assessment of data and a clear reference to the population sampled. Any interpretation or evaluation of the content of information provided should be omitted. 3. Transparency: the assumptions and methodologies utilized need to be clearly explained and documented. 4. The methodologies as well as the terms and definitions applied need to be consistent over time. Otherwise one cannot distinguish between true change and change due to changes in methodologies or terms and definitions. Changes should be only applied when it can be argued that the benefits outweigh the problems introduced. 5. The information has to be reliable. Any forest inventory is subject to a certain degree of uncertainty, introduced by sampling, measurements, and models applied. It is good practice to specify the degree of uncertainty of any figure provided. As sample-based results are always subject to sampling errors, it is necessary to accompany any statistical estimate with estimates of sampling error or confidence interval. Measures for quality control and assurance need to be introduced in order to improve reliability. 6. The information must be assessed in a cost-efficient way. For a given budget, the inventory design resulting in the most reliable estimates needs to be selected among different design alternatives. Alternatives can be based on different sampling designs, sampling intensities, or data sources. 7. The results of an inventory should be intuitively clear for potential users. Users are normally not very familiar with sampling statistics, and thus the results should not require a PhD in statistics for any immediate and basic interpretation. Users will have confidence only in information that they can understand. 8. Planning of a forest inventory is a complex task that involves the expertise from many fields. Thus experts from silviculture, forest management planning, economy, policy, ecology, or timber products need to be consulted at an early stage. Information on forest resources in a defined area can be obtained by either (1) a total tally of all trees or (2) sample surveys that utilize representative samples to draw inference for the area of interest. As total tallies are time consuming, they qualify only for small, forested patches. Therefore, forest resources assessments generally use sample-based techniques. A sample survey consists of standardized approaches to collect information and utilizes four components (Wright and Marsden 2010): 1. Sampling: the selection of representative samples from populations, whose observed characteristics provide unbiased information of the characteristics of those populations

M. Ko¨hl and M. Marchetti

752

2. Inference: the generalization of sample statistics to estimate population parameters within calculable error margins 3. Assessment: strategies to collect reliable and valid information on individual members of the population 4. Analysis: (multivariate) data analysis techniques for the identification of complex statistical relationships among many characteristics of the population

Objectives of Forest Inventories Inventory objectives define the specific results that are to be achieved within a time frame and with available resources. They are the basic reference for planning, activities, and the evaluation of performance. Thus the objectives of an inventory have to be laid down in a very early phase of inventory planning. Three specific aspects should be considered when determining inventory objectives (FAO 1998): 1. Objectives need to be determined jointly by the people who will use the results, including forest managers, planners, and decision makers, as well as by inventory specialists. Inventory objectives should not be determined by inventory specialists alone. 2. Not all inventory objectives have the same level of importance. Some have higher priority than others and it is the objectives having highest priority that should determine the inventory design and the presentation of results. 3. Inventory objectives should reflect the physical effort that will be required to conduct an inventory, the organization, estimated costs and time, the existing knowledge of resources, the availability of specific aspects of inventory technologies, and institutional capability. All have a direct bearing upon the implementation of an inventory. An overriding consideration is that an inventory must be practicable and achievable. The value of an incomplete inventory that lacks important information and thus limits the possibilities for inference could be zero or close to zero. All objectives should be SMART: Specific Measurable Agreed upon Realistic Time framed

Well defined They are clear to anyone that has a basic knowledge of the project They provide quantifiable measures of achievement and variance from set objectives Have agreement between the users and the project team on what objectives should be Looking at the resources, knowledge, and time available, can the objective be accomplished? How much time is needed to accomplish the objective? Having too much time can affect the project performance

Objectives and Planning of Forest Inventories

753

As the thematic scope of forest inventories can vary considerably, it is advisable to review global initiatives and obligations in order to get a broad view on potential information topics to be covered by a forest inventory. The United Nations Conference on Environment and Development, Rio de Janeiro, 1992, produced a nonlegally binding document with recommendations for the conservation and sustainable management of forests (United Nations 1992). This document is known as the “Forest Principles” and may serve as a framework for the definition of inventory objectives.

Box 1: Forest Principles (Excerpt)

Forestry issues and opportunities should be examined in a holistic and balanced manner within the overall context of environment and development, taking into consideration the multiple functions and uses of forests, including traditional uses, and the likely economic and social stress when these uses are constrained or restricted, as well as the potential for development that sustainable forest management can offer. Source: United Nations (1992)

ITTO, TARA, CIFOR, ATO, and CCAD as well as the Tarapoto Proposal of Criteria and Indicators for Sustainability of the Amazon Forest, UNEP/FAO Expert Meeting on Criteria and Indicators for Sustainable Forest Management in Dry-Zone Africa, or the Lepaterique Process of Central America expanded on the Forest Principles and developed systems of criteria and indicators for sustainable forest management for managed forests, natural forests, and plantations, which cover administrative, economic, legal, social, technical, and scientific issues. Criteria define the essential factors of forest management against which forest sustainability may be assessed. Each criterion relates to a key management factor, which may be described by one or more qualitative, quantitative, or descriptive indicators. Through measurement and monitoring of selected indicators, the effects of forest management action, or inaction, can be assessed and evaluated and action adjusted to ensure that forest management objectives are more likely to be achieved. Table 1 (after FAO 1998) summarizes the criteria and indicators identified by the processes and initiatives and facilitates the definition of inventory objectives. As not all objectives have the same importance, the priority of inventory objectives has to be assessed before designing the inventory. Before a final decision on the inventory objectives, all issues that could constrain an implementation of the inventory should be listed and considered. Issues include cost limits, availability of staff, presentation of the findings, the schedule, or the population for which estimates should be given. The listing of inventory objectives should not be confused with the list of attributes to be assessed or derived. Based on the objectives, the attributes for field assessments, remote sensing imagery, or other data sources have to be derived. Figure 1 gives an example for the indicator “growing stock.” The objective

754

M. Ko¨hl and M. Marchetti

Table 1 Criteria and indicators for sustainable management (After FAO 1998) Criterion 1. Extent of Forest Resources and Global Carbon Cycles

Indicator Area of forest cover Wood growing stock Successional stage Age structure Rate of conversion of forest to other use 2. Forest Ecosystem Health and Vitality External Influence Deposition of air pollutants Damage by wind erosion Forest Vitality Indicators Incidence of defoliators Reproductive health Forest Influence Indicators Insect/disease damage Fire and storm damage Wild animal damage Anthropogenic Influence Competition from introduction of plants Indicators Nutrient balance and acidity Trends in crop yields 3. Biological Diversity in Forest Ecosystems Distribution of forest ecosystems Ecosystem Indicators Extent of protected areas Forest fragmentation Area cleared annually of endemic species Area and percentage of forest lands with fundamental ecological changes Forest fire control and prevention measures Species Indicators Number of forest-dependent species Number of forest-dependent species at risk Reliance of natural regeneration Resources exploitation systems used Measures in situ conservation of species at risk Genetic Indicators Number of forest-dependent species with reduced range 4. Productive Functions of Percentage of forests/other wooded lands managed according Forests to management plans Growing stock Wood production Production of non-wood forest products Annual balance between growth and removals of wood products Level of diversification of sustainable forest production Degree of utilization of environmentally friendly technologies 5. Protective Functions of Soil conditions Forests Water conditions Management for soil protection Watershed management Areas managed for scenic and amenity purposes Areas and percentage of forest lands managed for environmental protection Infrastructure density by FMU category (continued)

Objectives and Planning of Forest Inventories

755

Table 1 (continued) Criterion Indicator 6. Socioeconomic Functions and Conditions Value of wood products Indicators for Economic Value of non-wood products Benefits Value from primary and secondary industries Value from biomass energy Economic profitability of SFM Efficiency and competitiveness of forest product production, processing, and diversification Degree of private and non-private involvement in SFM Local community information and reference mechanisms in SFM Indicators for the Distribution Employment generation/conditions of Benefits Forest-dependent communities Impact on the economic use of forests on the availability of forests for local people Quality of life of local populations Average per capita income in different forest sector activities Gender-focused participation rate in SFM 7. Political, Legal, and Legal framework that ensures participation by local Institutional Framework governments and private landowners Technical and regulatory standards of management plans Cadastral updating of the FMU Percentage of investment on forest management for forest research Rate of investment on the FMU level activities: regeneration, protection, etc. Technical, human, and financial resources

Fig. 1 Objectives, indicators, and attributes

756

M. Ko¨hl and M. Marchetti

“productive function of forests” is described by several indicators such as “percentage of forests managed according to management plans,” “growing stock,” “wood production,” or “annual balance between growth and removals of wood products.” The indicator “growing stock” is quantified by the attribute “volume over bark,” which is derived by volume functions. Attributes assessed in the field (e.g., diameter at breast height, tree height, upper stem diameters) are used as input variables for regression functions with the volume over bark as dependent variable.

A Typology of Forest Inventories Forest inventories can be differentiated according to combination and emphasis of different inventory objectives and the size of the area to be surveyed. Global forest resources assessments are conducted to determine forest resources at the global level. This usually means the compilation of results from individual national inventories. Thanks to advances in remote sensing techniques, satellite data can now be used to determine the distribution of forest vegetation throughout the world (Gerrand et al. 2009; Huete and Saleska 2010; Achard et al. 2010; Stibig et al. 2014). Global forest inventories were conducted by the FAO since 1946 (FAO 2010). The first FAO-led assessments were dominated by studies on timber supply. Today the Forest Resources Assessment (FRA) adopted the design of sustainable forest management as a reporting framework. The economic, ecological, and social elements of forests are covered as well as the legal, policy, and institutional framework. The FRA does not conduct an own survey. For the FRA 2010, a total of 233 countries and areas contributed data on the base of a harmonized reporting scheme. Results are presented for six regions and 12 subregions and for seven themes (FAO 2010): • • • • • • •

Extent of forest areas Forest biological diversity Forest health and vitality Protective functions of forest resources Productive functions of forest resources Socioeconomic functions of forests Legal, policy, and institutional framework

A national forest inventory (NFI) is carried out in a specific country and provides information about forest resources. The reporting framework is generally oriented toward sustainable forest management and the multiple ecosystem services and functions of forests. Often the data assessed by NFIs are used to develop timber supply studies that predict future increment and potential harvesting volumes for production forests. The results of NFIs support policy decisions, are utilized by the timber sector (e.g., timber supply, investments) and environmental administrations, and serve as input to scientific studies.

Objectives and Planning of Forest Inventories

757

National forest inventories can be realized as sample-based inventories covering the entire country or as a compilation of stand-level data assessed for forest management planning. The latter was common practice in Eastern European countries. The origins of sample-based NFIs date back to the 1920s, when surveys were conducted in the European Nordic countries and in the United States (Ilvessalo 1927; Tomppo et al. 2010). Today NFIs are generally tailor-made sample surveys designed for specific needs and forest characteristics and thus vary from country to country (Tomppo et al. 2010). This holds true for the systems of nomenclature applied as well, which aggravates comparisons of NFI results between countries (Ko¨hl et al. 2000). Generally, field surveys and remote sensing serve as key data sources. NFIs are designed as inventories at successive occasions, utilizing either periodic assessments or “rolling” surveys that assess a proportion of the country each year (Scott et al. 1999). Regional forest inventories register only a part of the national forested area and usually cover some hundreds of thousands up to 2 Mio. hectares. Similarly to national inventories, they are intended to provide a general picture of the situation regarding forestry. Land-cover and land-use inventories record not only forest resources but also the spatial distribution of other types of land cover. There is a distinct difference between land use and land cover: • Land cover is the observed (bio)physical cover of the Earth’s surface. In a strict sense, this concept applies to vegetation and man-made features. For practical applications, land cover includes bare rock and soil as well as water surfaces. • Land use relates to the actions of people in their environment. Table 2 presents examples for the differences of land use and land cover. As the major output of land-cover and land-use surveys are maps, the use of aerial photographs and satellite data is of special importance. At the community level, typical methods for assessing land use are walking or windshield (conducted from a car) surveys utilizing GPS tracking systems. Importing land-use data to a GIS system renders further spatial analyses possible (e.g., landscape fragmentation, buffer zone analysis). Table 2 Land use versus land cover

Land use Soccer field Rangeland Golf course Recreation area

Protected area

Land cover Grass

Forest Beach Built-up area/shopping mall Forest Bare sand Tideland

758

M. Ko¨hl and M. Marchetti

An essential step in planning land-cover and land-use inventories is the selection of a land classification system, which is an abstract representation of the real situation on the ground. The class boundaries need to be clearly defined, preferably by objective criteria. A classification system has to be scale and source independent and can be hierarchical (broad-level classes with subclasses) or nonhierarchical. FAO and UNEP created the Global Land Cover Network (GLCN) in order to develop a harmonized approach for the preparation of land-cover data on the local, national, and international level (www.glcn.org). Reconnaissance inventories aim at furnishing a rough outline of the forest conditions. As well as the location and extent of forested areas, they may aim to register access, species composition, tree dimensions, the distribution of various forest types, and a crude assessment of timber quality (Touber et al. 1989). Through the employment of remote sensing imagery and the restriction of field surveys to the minimum, reconnaissance inventories can be conducted at little cost. They frequently serve the preparation of a more intensive forest inventory. Data on the degree of variation and time-and-motion studies conducted during a reconnaissance inventory facilitate the planning of the definitive inventory. Exploitation surveys or logging plan surveys are conducted in forests to provide a basis for the planning of programs for timber harvesting. The main focus is on determining the standing crop, classified according to commercial species, dimensions, timber quality, and assortments and describing the accessibility of the area concerned. Little or no attention is paid to increment, ecological conditions, or sustainability. Where the economic potential of establishing a wood-processing industry is to be examined, a “forest industries feasibility study (FIFS)” is a standard practice. A FIFS comprises the collection of data not only on the forest resources as such but also on the situation regarding demand and marketing, potential sites for processing plants, the job market, sources of water and power, transport possibilities, and existing industries. As the establishment of a timber industry is only worthwhile where there is a steady supply of raw material, it is necessary to determine the forest resources in considerably more detail than is usual in exploitation surveys. In particular, the sustained yield of exploitable timber must be estimated. Working plan surveys are an intensive type of assessing managed forests. The preparation of working plans for intensively managed but restricted areas requires relatively detailed information. Usually the data are computed on a stand-by-stand basis for each species. Information on increment, detailed forest maps, and data on the quality of the various sites are just as necessary as details on topography, ownership, and access (Schmid 1969; Schmid-Haas et al. 1978). The various types of inventory are not clearly defined but overlap each other. Neither is their classification fixed, as the increasing demands on the forest will lead to new inventory forms. In addition to the types of inventories described above, special surveys are sometimes conducted, for instance, to determine regeneration, available biomass, or carbon sequestration. Forest inventories may also be classified in terms of time. Static inventories may be conducted simply to determine conditions at a given point in time and do not require consideration of possible

Objectives and Planning of Forest Inventories

759

subsequent inventories – a fact which considerably simplifies their planning. Nevertheless, the additional expense of permanently marking the sample plots is often a worthwhile investment.

Inventory Planning Phases The execution of a forest inventory can be divided into four main steps: 1. Definition of the inventory objectives and the information desired 2. Development of the inventory design, including the sampling design, assessment procedures, organization, work flow, and budget 3. Data assessment (field surveys, remote sensing, and gathering of information from other data sources) 4. Analysis of the collected data and publication of the results Each of these steps requires the knowledge and cooperation of many experts. This includes expertise in inventory statistics, information technology, field assessments, remote sensing, GIS, sustainable forest management, forest ecology, and communication. Due to the need for networking between different tasks, it is advisable to base the entire planning on a sound project management design. Good project management deals with three factors: time, cost, and performance. Projects are successful if they are completed on time, within budget, and to performance requirements. In order to bring the many components of a large project into control, there is a large toolkit of techniques, methodologies, and tools. These techniques provide the tools for managing different components involved in a project: planning and scheduling, developing a product, managing financial and capital resources, and monitoring progress (Meredith and Mantel 2012). However, the success of a project will always rest on the abilities of a project manager and the team members. A project life cycle includes the four phases: (1) study phase, (2) design phase, (3) development phase, and (4) operation phase. The following overview presents individual steps in the four project management phases and is intended to serve as a checklist for inventory planning (Lund 1998; Ko¨hl et al. 2006). A. Study phase User needs assessment

Screening and identifications of potential user groups Validate identified user groups by studying information flows between local and upper level as well as among different user groups Discuss with individuals their information demands Assess information required by laws, mandates, policies, international reporting Prepare concise statement of information needs (continued)

760

Initial investigation

Inventory objectives

Determination of the administrative and logistic situation

Study phase report

B. Design phase General system review Compilation of the data catalog and stipulations for measurements

M. Ko¨hl and M. Marchetti

Available information on forest resources Previous resource assessments in the inventory region Organizations with expertise in forest resources assessments, GIS, remote sensing, etc. Identify constraints Possibilities/needs of outsourcing of specific tasks Necessity of the inventory, information needed Potential users of the results Formulation of the inventory objectives Priorities of the objectives Bodies responsible for the execution Budget (available funds, bodies providing funds, financial administration, time available) Legal basis (right of access to privately owned forest, labor laws, protection of private forest owners from information leaks) Available information (maps, aerial photographs, data from previous forest inventories and other types of survey, scientific studies in the inventory area, general details on the forest, data on variation, description of the terrain, accessibility, and climatic conditions) Potential use of aerial photography and remote sensing imagery Possibilities for recruiting qualified staff Available equipment (vehicles, computers and software, measuring instruments, tents) Responsible bodies (staff management, financial administration, monitoring of data security, data release, dissemination of data, definition of inventory objectives and methods, execution of field surveys, data evaluation, formulation and release of the final results, publication, additional analyses) Prepare draft study phase report Review of draft by users, administrative bodies, and constituents Finalize study phase report

Listing of all variables to be analyzed (depending on inventory objectives) Definition of qualitative data Instructions for measurement of quantitative data Methods of volume determination (e.g., volume functions, points of measurement on the tree, volume inside or outside bark) (continued)

Objectives and Planning of Forest Inventories

Inventory design

Database design Procedures for quality assurance and control Software selection Equipment selection/acquisition Staff recruitment Field manual

761

Development of sampling design alternatives Selection of the optimal (cost-efficient) design Description of the design to be employed Description of the sampling units, especially their form, size, number, and distribution Computation of the necessary sample size for each inventory level, survey intensity Description of inventory levels (aerial photographic survey, interpretation of satellite data, field surveys, questionnaires) Map construction Area estimation Description of statistical methods for evaluation, estimation procedures, correlations to be applied Description of the assessment of road and transport networks See section “Quality Assurance and Quality Control”

Fieldwork organization Access to plot Establishment of permanent plots Data collection in the plot Data transmission

Plans for work progress Design phase report C. Development phase Implementation planning Computer program design User review Equipment acquisition and installation Field tests/pilot survey Computer program testing System testing Reference manual preparation Personnel training Changeover plan preparation Development phase report preparation User acceptance review (continued)

M. Ko¨hl and M. Marchetti

762

D. Operation phase Interpretation of aerial photographs and/or remote sensing data

Field surveys

Data evaluation

Final report

Instruments (interpretation instruments, computers, software) Organizations, staff, competence, duties Documentation and backup of the results Organization, central coordination Communication between field survey teams and central coordinators Recording and delivery of data Training of field staff (localization of sample plot centers, assessments on sample plots, use of instruments) Check cruises Digitalization of data Checking and correction of data Data analysis Operating, system management, data security Preparation (outlet format printed and/or internet) Approval for release Reproduction Dissemination

Performance evaluation

Selection of the Optimal Inventory Design A crucial step in planning a forest inventory is the selection of the optimal inventory design. Given the multitude of inventory objectives, the selection of the most suitable inventory design can be a controversial issue. For example, in a prelogging survey of a concession area, the optimal design for the assessment of tree species diversity is markedly different from the optimal design for the assessment of the market value of potential crop trees. According to FAO, forest inventories and assessments aim to “contribute to the sustainable management of forests . . . by providing decision makers and stakeholders with the best possible, most relevant and cost-effective information for their purpose at local, national and international levels” (Saket et al. 2002). Thus, the inventory design is a compromise between the information needed, the reliability of results, and the cost of the survey. Irrespective of the objective of a survey, alternative inventory concepts exist to choose from, including the utilized data sources (field assessments, remote sensing, maps, etc.) and the design of the sampling units, sampling rules, and sample sizes. The potential design alternatives are influenced by a variety of factors such as the variability of the target population, budget allowance, or availability of auxiliary data sources and information (e.g., maps, remote sensing imagery, volume

Objectives and Planning of Forest Inventories

763

equations). A rational decision about the optimal design can be made only by comparing the set of alternatives under objective selection criteria that combine information on survey cost and the achievable reliability of the results. This allows for selecting the most cost-efficient design that either provides the best reliability under a given budget or provides the desired reliability by least cost. A measure of cost-effectiveness is the relative efficiency (Cochran 1977). For fixed cost, relative efficiency of alternative A to alternative B is defined as the ratio of the variances of the two alternatives. Conversely, the cost can be compared for a given precision level. The latter situation is, however, rare in the initial phase of forest surveys. A fixed budget is normally specified. Thus the survey design has to be selected that offers the largest gain in precision for given cost. The design effect, DEFF, was presented by Kish (Kish 1965) as an alternative concept for comparing sampling designs. Given the same sample size, DEFF is the ratio of the variance of an estimate obtained from a complex sample to the variance obtained from a simple random sample. Both measures, DEFF and relative efficiency, are used for analytical comparisons of sampling design alternatives. For the practical decisions to be made in the survey planning process, the relation of cost and efficiency, especially the effect of a change of costs on the precision, is needed. In order to meet these requirements, cost-precision relations may be presented in graphical form in order to facilitate the understanding of cost-efficiency relation of different sampling scenarios and of the effect of increased costs on the precision of estimates. As the information needs of different stakeholders need to be taken seriously, the optimization process should not be targeted to a single attribute. A design tailored for the assessment of commercial timber volume might be insufficient for the assessment of biodiversity. A prioritization of information needs and their underlying attributes is essential.

Cost Survey costs are made up of fixed and variable cost components. Fixed costs are those that do not vary with sample sizes and design alternatives but are common to all alternatives. Examples for fixed cost are expenses for administration, office rent, or software purchasing. As fixed costs are design independent, they are not to be considered in the optimization process (Scott and Ko¨hl 1993; Groves 2004). Design dependent costs include additional fixed costs for specific design alternatives and variable costs. Costs for visiting and measuring field samples are a typical example for variable costs, which are proportional to the number of field samples assessed. The additional costs of, e.g., stratified sampling would include costs for acquisition, enhancement, and classification of remote sensing data as well as validation of the classification results. The purchase of aerial photographs or digital imagery would be a fixed cost component associated to a design utilizing stratification.

764

M. Ko¨hl and M. Marchetti

Hardcastle and Baird (2008) studied the readiness of 25 tropical countries for monitoring forests and reporting on REDD under the IPCC guidelines. For each country, cost estimates are provided for implementing REDD monitoring and reporting systems, the major divers of costs being forest extent, stratification, and the chosen reliability level, i.e., tier. They present the initial and recurrent cost separately for four alternatives: 1. Tier 2, approach A: an accurate land-cover map is available, 300 sample plots are assessed in situ, all carbon measurements are performed once at the beginning of the program, and future monitoring is focused on the assessment of AD by remote sensing data and requires only minimal field work. 2. Tier 2, approach B: no accurate land-cover map is available, in situ assessments are performed when activity monitoring by remote sensing identifies locations under change, and the in situ sampling intensity is considerably lower than under Tier 2, approach A. 3. Tier 3, ignoring degradation: AD und EF are assessed as under alternative 1 (Tier 2, approach A), but remeasurements are made in permanent in situ sample plots (about 1/3 of the original sample locations). 4. Tier 3, including degradation: alternative 3 is enhanced by further stratification of forests into the two classes: “intact forests” and “non-intact forests,” and the number of field plots is moderately increased. The inventory concepts applied by (Hardcastle and Baird 2008) are generic rather than case specific, as they do not result from a sound inventory design and optimization process on the national level. However, they are used for an approximate comparison of cost required to implement an operation REDD monitoring and reporting scheme on the national level. Figure 2 presents the respective costs for the four alternatives over forest area. The cost per unit area decreases with increasing forest area, as the share of fixed costs in total costs decreases.

Fig. 2 Cost estimates [US$/ha] in relation to forest area in the scope of REDD (Data from Hardcastle and Baird 2008)

Objectives and Planning of Forest Inventories

765

Table 3 Aboveground biomass stock in m3 per forest formation (Adapted from IPCC (2003))

Africa Asia and Oceania: continental Insular America

Tropical forests Moist with short dry Wet season 310 260 (131–513) (159–433) 275 182 (123–683) (10–562) 348 (280–520) 347 (118–860)

Moist with long dry season 123 (120–130) 127 (100–155)

Dry 72 (16–195) 60

290

160

70

217 (212–278)

212 (202–406)

78 (45–90)

Montane moist 191

Montane dry 40

222 (81–310)

50

362 (330–505) 234 (48–348)

50 60

Note: Data are given in mean value and as range of possible values (in parentheses)

Variability Sample sizes and thus survey costs are directly linked to the variability of the sample population. Variability data for a population can be obtained by prior knowledge or by a pilot survey. For each variance component that is enclosed in the estimation procedures, variability figures have to be specified. For stratified sampling, this means specifying the variance by stratum within the smallest unit of reference for each attribute of interest. Table 3 presents mean values and possible ranges of aboveground biomass stock per hectare, as given by IPCC (2003) for tropical forests.

Sample Design Alternatives Different sampling design alternatives can be used in the scope of forest inventories. These sampling designs can employ in situ (field plot) data, remote sensingbased data, or a combination of the two. Combined in situ/Earth observation sample designs use information obtained by remote sensing and field sampling systems simultaneously. The Earth observation data can consist of derived data, such as a classification of remote sensing data into land-cover classes, or reflectance data from optical, radar, or lidar sensors. Variables of interest such as biomass or carbon stock are assessed on a small sample of field plots, and these data are combined with the more densely sampled Earth observation data using statistical estimation procedures in order to generate estimates. Especially where airborne instead of spaceborne sensors are used, it can be prohibitive to cover large areas with remote sensing imagery. Similarly, field data collection campaigns can be costly, especially in areas that are hard to access.

M. Ko¨hl and M. Marchetti

766

Optimization For each sampling alternative, there exists an optimum combination of sample sizes. These optimum combinations should be used to compare the various design alternatives. In the optimization process, variance functions and cost functions have to be linked in order to derive the optimal (i.e., most cost-efficient) sampling alternative. The optimum sampling design can be defined in two ways: 1. Minimizing cost for a specified level of precision 2. Minimizing variance for a specified cost In either case, the optimization requires that the cost and precision be expressed in terms of the sampling design and sample sizes. It is good practice to present costefficiency graphs for the design alternatives compared, where the achievable percent standard error is plotted over the number of field plots (Fig. 3) or cost (Fig. 4). The selection of the optimal design by deriving design alternatives is subject to constraints. Therefore a sensitivity analysis needs to be performed where input variables such as cost and variance are inflated. The result will help to identify those designs that are most sensitive for changes in underlying constraints.

Selection Criteria The selection of the optimal design is not only driven by cost-efficiency. It is good practice to adopt additional selection criteria, such as the tolerance for violations in the selected variance and cost figures, flexibility for future methodological developments, the possibilities to integrate other data sources for sophisticated analyses, the intuitive comprehensibility of the design and the provided results, the

Percent Standard Error

60 50 40 30 20 10 0 0

50

100

150

200

Number of Field Samples Alternative A:

Alternative B:

Alternative C:

Fig. 3 Sampling design alternatives: percent standard error over number of field samples

Percent Standard Error

Objectives and Planning of Forest Inventories

767

60 50 40 30 20 10 0 0.0

0.5

10

1.5

2.0

2.5

Cost [US$/ha] Alternative A:

Alternative B:

Alternative C:

Fig. 4 Sampling design alternatives: percent standard error over cost

complexity of estimation algorithms, the acceptance by the users, or the involvement of local populations. Hence, the selection of the optimal design is a participatory process that involves not only survey statisticians or academia but appropriate and qualified set interest groups.

Defining the Sample Population It seems to be simple: forest inventories provide information on forests located in the area of interest. What at first glance seems to be simple is in fact a complex issue. There is neither a common definition of forests nor is the area for which the inventory should provide results self-explanatory. For example, one might get information on the forests of a country but omit mangrove forests. Or, does an inventory of logging concessions include natural forest reserves within the concession area? A key question is how to define a forest.

Forest Area Definitions There exists no unique concept about what qualifies as a forest. The hyperdictionary1 provides a definition of forest that represents two different aspects: • The trees and other plants in a large densely wooded area • Land that is covered with trees and shrubs When we talk about forest area, we relate to the area covered with trees and shrubs. National legislations often have a legal definition of forest, but they are 1

http://www.hyperdictionary.com/dictionary/forest

768

M. Ko¨hl and M. Marchetti

Fig. 5 Forest or non-forest?

generally not applicable to forest resources assessments. Forest is a qualitative attribute of an area, which cannot be measured directly. Figure 5 presents three different situations. While the picture on top gives clear evidence of a forest, the picture on the lower left shows a tree outside forests. On the lower right side, an example from the Swiss Alps is given, where trees grow close to the timberline. A predominant characteristic of forests close to natural timberlines is that tree density is gradually lowered toward the timberline. Thus, forest area definitions need to draw the borderline between trees inside a forest and trees outside forests. Forest area definitions utilize a set of quantifiable and measurable attributes to separate forest from non-forest land. In order to increase the reliability of forest area assessments, forest area definitions are based upon attributes that can easily be measured, such as crown cover, size or width of the forested patch, tree height, or productivity. For those attributes, threshold values are specified, and whenever a patch qualifies for the selected set of attributes, the patch is considered to be forested land. Forest area definitions may also contain specifications about the allowed or disallowed use of forests and forest types. The set of attributes selected as well as the specified threshold values vary in individual forest area definitions. Table 4 shows some forest area definitions used by international organizations.

Objectives and Planning of Forest Inventories

769

Table 4 Forest area definitions UNESCO

FAO Forest Resources Assessment 1990 (FAO 1995) FAO Forest Resources Assessment 2010 (FAO 2001)

Closed forest: tree 5 m or taller with crowns interlocking Woodland: trees 5 m or taller with crowns not usually touching but with more than 40 % canopy cover Forest (developing countries): 10 % crown cover for trees and/or bamboos Forest (developed countries): tree crown cover (stand density of more than 20 % of the area Land with tree crown cover (or equivalent stocking level) of more than 10 % and area of more than 0.5 ha. The trees should be able to reach a minimum height of 5 m at maturity in situ. May consist either of closed forest formations where trees of various stories and undergrowth cover a high proportion of the ground or of open forest formations with a continuous vegetation cover in which tree crown cover exceeds 10 %. Young natural stands and all plantations established for forestry purposes which have yet to reach a crown density of 10 % or tree height of 5 m are included under forest, as are areas normally forming part of the forest area which are temporarily unstocked as a result of human intervention or natural causes but which are expected to revert to forest Includes: forest nurseries and seed orchards that constitute an integral part of the forest; forest roads, cleared tracts, firebreaks, and other small open areas within the forest; forest in national parks, nature reserves, and other protected areas such as those of special environmental, scientific, historical, cultural, or spiritual interest; windbreaks and shelterbelts of trees with an area of more than 0.5 ha and a width of more than 20 m. Rubberwood plantations and cork oak stands are included Excludes: land predominantly used for agricultural practices

Ko¨hl et al. (2000) studied the effect of different national forest area definitions on the estimated size of forest area in a simulation study. Differences in the spatial distribution of trees and forested patches were simulated in computer-generated forest/non-forest maps. The computer-generated forests were used to simulate the impact of exchanging one national definition with another in a number of European countries. This approach allowed to estimate the effect of different national forest area definitions in absolute terms. For example, depending on the chosen definition, the range of the Spanish forest area could vary from about 240,000 km2 (reference United Kingdom) to 274,000 km2 (reference Luxembourg). A universal forest definition remains elusive. Sample-based estimates of forest area must therefore carefully consider the existing forest definition and take competing definitions into account to ensure that estimates of forest areas can be obtained for more than one definition. In the early phase of planning a forest inventory, it should be considered whether only areas currently supporting forest vegetation are to be surveyed or whether former forest areas and sites suitable for reforestation or afforestation should be included (Pancel 1984; Weaver and Birdsey 1986).

770

M. Ko¨hl and M. Marchetti

Target Population and Sample Frame The term population is used to define a set of elements from which a sample is selected and for which inference should be drawn. An element is defined as the basic unit that comprises the population. For example, a population can be defined as the entirety of all trees within a given forest area. Each element (here: an individual tree) can be characterized by one or several attributes (e.g., tree species, diameters, or tree height). In addition to the general term population, the more specific terms target population and sample population are used (Rossi et al. 1983). A sample population is a clearly defined population from which the sample will be selected; a target (or inferential) population is a clearly defined population to which the results will be applied. Ideally, sample and target populations coincide. However, many practical applications involve the risk that the target and the sample population differ. There are many examples of this: • Field assessments are carried out in the vicinity of roadways or rivers in order to reduce walking time; consequently the interior parts of forests are missed. • Locations in remote areas and thus subject to laborious and time-consuming access are less frequently visited. • Non-accessible areas are not excluded from the target population. • Permanent sample plots are omitted in clear-cut activities and lead to wrong information on fellings. A sample frame (synonyms: sampling frame or survey frame) is the actual set of units from which a sample is drawn. Ideally, it coincides with the target population. Sometimes the appropriate unit is obvious, e.g., the trees in a forest stand or the logs delivered to a sawmill. However, in forest inventories, the definition of the sample frame is generally more complex. Due to the extensive areas covered, it is impossible to generate a list including all trees growing in the inventory area. This problem is resolved by defining a sample frame that includes identifying information about characteristics of the individuals. In forest inventories, it is good practice to define a sampling frame not by reference to individual trees but by defining what is considered as a forest by a unique forest area definition (see section “Forest Area Definitions”).

Quality Assurance and Quality Control In addition to a sound design, quality assurance (QA) and quality control (QC) are essential for delivering quality information and services. According to the American Society for Quality (ASQ), QA is The planned and systematic activities implemented in a quality system so that quality requirements for a product or service will be fulfilled.

Objectives and Planning of Forest Inventories

771

QA is any systematic process of checking to see whether a product or service being developed is meeting specified requirements. A QA system is aiming at increasing end users’ confidence and the inventory’s credibility and improving work processes and efficiency. A major advantage of QA systems is to catch deficiencies of the entire inventory process before they get into the final reporting. The inventory management team decides quality assurance policies and objectives. Next, policies and requirements and how the staff can implement the quality assurance system are formally written down. Once this guideline is in place and the quality assurance procedures are implemented, the documentation of the quality of field data and data quality evaluation is ensured for operational applications (Ferretti et al. 1999; Kaufmann and Schwyzer 2001; Kitahara et al. 2009). Highquality standards and their verification lead to complete, unbiased, and factual representation of forest resources (USDA Forest Service 2012b). The terms quality assurance (QA) and quality control (QC) are sometimes confused. QA is an overall management plan or system, which includes the organization, planning, data collection, quality control, documentation, evaluation, and reporting activities. QA provides the information needed to ascertain that data meet defined quality standards. Quality control (QC) refers to routine technical activities, which are used for error control. Since errors occur in the field, in remote sensing analysis, in the laboratory, in data analysis, and in the office, QC measures must be part of any of these tasks. QA and QC should result in a quality assurance project plan (QAPP), which is a written record of the entire QA/QC program.

QA Components QA is planned and systematic activities implemented in a quality system so that quality requirements for a producing the final report will be fulfilled. It is a continual program that aims at continually improving the quality of data. The information obtained is necessary for interpreting and evaluating survey results, develop realistic objectives for measuring data quality, revising methodology to reduce efforts, improve the effectiveness of training sessions, and revising the remeasurement program for QC data (USDA Forest Service 2012). A QA plan identifies all operations and procedures that require control. In forest inventories, that comprises all planning, development, and implementation phases (see section “Inventory Planning Phases”) and thus relates to a wide range of activities, including staff recruitment and training, software development, remote sensing imagery analysis, and field assessments. For each operations and procedures, appropriate control protocols need to be defined, documented, and implemented. QA has three components: prevention, assessment and appraisal, and correction. Prevention aims at processes that ensure that a process results in an output with desired quality before the process begins. For field assessments, prevention activities include the development of standardized definitions and procedures, which are

M. Ko¨hl and M. Marchetti

772

documented in a field guide, and the agreement on data tolerance limits as well as training and calibration methods. Tolerance limits or minimum quality objectives (MQOs) define the precision of any attribute assessed in the field. For attributes on metric scales, they give the range in which measurement errors can be accepted (e.g., dbh  1 cm). They are more complex for attributes on nominal or ordinal scales. Calibration generally refers to measurement instruments but can be extended to any measure taken to ensure the reproducible assessment of data. A comprehensive field manual is the base for high data quality. It lays down all definitions and measurement rules as well as procedure for assessments. FAO (2004) and USDA Forest Service (2012a) provide good examples for field manuals. In addition, it is good practice to implement cross-checking routines and plausibility checks in mobile data loggers. For example, when for a tree with a diameter of 60 cm a tree height of 10 m is entered, a warning should be flagged. The check of data entries on the fly guarantees for the completeness of data, enables immediate corrections and verification on the plot, and thus improves data quality. A major component of prevention activities is training, which ensures that field crews have the required skills to meet the minimum quality objectives for data assessment. Training is carried out in several steps: field crews are introduced to the assessment processes and methodology, then they practice the methodology, and finally the crew performance is evaluated and documented. It is good practice to introduce a certificate at the end of a training course that approves the qualification of staff. Only certified staff members should be tasked with field assessments (Fig. 6). Assessment and appraisal relates to the measure of data quality, which is described by accuracy, precision, completeness, comparability, detection limit, and measurement range. Precision is the degree of agreement among repeated measurements of the same characteristic on the same object. It shows how close measurements are to each other and how consistent and reproducible methods are. It does not mean that the results actually reflect the “true” value but rather that the assessment method provides consistent results under similar conditions. When only two replicate measurements are available, precision can be determined by the relative percent difference RPD: RPD ¼

x1  x2

100 ðx1 þ x2 Þ=2

where x1 is the larger and x2 is the smaller of two measurements. Where many replicate measurements are available, the relative standard deviation RSD is a useable measure to quantify precision: s RSD ¼ 100 x where s is the standard deviation and x is the mean of the replicate measurements. The smaller RSD and RPD are, the more precise are the measurements.

Objectives and Planning of Forest Inventories

773

Accuracy refers to the size of deviations from the true value. The smaller the difference between a measured and a “true” value, the more accurate is the measurement. Measurement accuracy can be determined by comparing a standard reference material to measurements of this material. This process will help to identify if measurement techniques or measures of individual observers are biased. Where data on nominal or ordinal scales are assessed (e.g., tree species), RSD and RPD cannot be calculated. Here the percentage of correctly classified observations is a good measure for data quality. Completeness describes a measure of the number of samples or measurements that need to be taken. Field crews can either not collect as many data as planned on a field plot or exclude entire plots. The percent completeness %C is a measure on how much of the desired data is assessed in the field: %C ¼

v T

100

where v is the number of measurements taken and T is the total number of desired measurements. Comparability describes the extent to which data can be compared directly to either data from previous assessments or data from other studies. It is a useful measure for ensuring data quality in inventories at successive occasions. Detection limit and measurement range applies to the reliability of measuring instruments. Measurement range specifies the range of reliable measurements. Detection limit is defined as the lowest resolution the methods and equipment can detect and report to be greater than zero. Readings below the detection limit are too unreliable to use. As readings approach the detection limit, they become less and less reliable. Manufactures of instruments generally provide information on detection limits and measurement ranges. Correction aims at the improvement of measurement and assessment procedures. The information from assessment and appraisal is used to identify individual procedures that result in data that do not comply with the desired quality and need improvement. In pilot surveys that are conducted under given guidelines in order to verify the suitability of assessment methods, corrections are an important step to improve data quality. In exceptional cases, corrections can be applied during regular field assessments. However, in these situations, the comparability of data before and after corrections needs to be carefully studied.

QC Components QC is the observation techniques and activities used to fulfill requirements for quality. In forest inventories, they are routine procedures to control the data collection process. Inspections carried out during the entire data collection phase are a vital part of QC. It is good practice to identify an experienced field crew as QA inspection crew, who are responsible for the inspection and training of production

774

M. Ko¨hl and M. Marchetti

Fig. 6 Structure of training (After USDA Forest Service 2012)

crews, are involved in the routine data assessment. According to USDA Forest Service (2012), there are different types of inspections: Hot checks are inspections during training courses. The QC inspection crew visits a plot together with a production crew. Immediate feedback on data quality is provided, and data errors are corrected. Hot checks can either take place on training plots, which are established for training only, or on ordinary field plots. Exit checks are used for training and for the evaluation of the readiness of a production crew for certification (see Fig. 6). The QC inspection crew has the complete set of data available and visits the plot together with the production crew. Data errors are discussed and corrected. Cold checks are performed on production plots only and are part of the training or the regular QC program. Any production plot recently visited by field crews can be selected for cold checks. The production crews do not know, which plot will be selected, so that they cannot adjust their performance as a result of that knowledge. Cold checks are single-blind measurements, as the QC inspection crews has the completed data sets assessed by the production crew at hand. The inspection can relate to the whole or only to parts of the regular measurements on plots. Data errors are corrected. It is good practice to conduct cold checks within a couple of weeks after training with high intensity in order to verify if production crews meet the quality standards under field conditions. The QC inspection crew remeasures the plots and before leaving the plot compares the measurements with the data recorded by production crews. The results are documented and discussed with the field crews on short notice. Field crews who fail to meet the quality standards are identified and subject to additional training. Field crews failing to meet the quality standards repeatedly may be removed from the program. Where data quality is poor, plots may be rejected and remeasured. Blind checks are a reassessment of plots by QC field crews or by an alternative field crew without knowing the data of the previous assessment. Field crews do not know which plots will be selected for blind checks. Plots for blind checks are

Objectives and Planning of Forest Inventories

775

randomly selected after production plot data have been submitted. It is good practice to select up to 10 % of the production plots for blind checks (Kaufmann and Schwyzer 2001; Kitahara et al. 2009, USDA Forest Service 2012).

References Achard F, Stibig H-J, Eva H, Lindquist E, Bouvet A, Arino O et al (2010) Estimating tropical deforestation from Earth observation data. Carbon Manag 1(2):271–287 CBD (1995) Preliminary consideration of components of biological diversity particularly under threat and action which could be taken under the convention, COP 2 Decision II/8, http://www. cbd.int/decision/cop/default.shtml?id=7081 Cochran WG (1977) Sampling techniques, 3rd edn. Wiley, New York Corona P, Marchetti M (2007) Outlining multi-purpose forest inventories to assess the ecosystem approach in forestry. Plant Biol 144(2):243–251 Corona P, Ko¨hl M, Marchetti M (2003) Advances in assessments for sustainable forest management and biodiversity monitoring. Kluwer, Dordtrecht Cotta JH (1804) Systematische Anleitung zur Taxation der Waldungen. Sander, Berlin FAO (1995) Forest resources assessment 1990 - global synthesis. FAO forestry paper. Rome FAO (1998) Guidelines for the management of tropical forests 1. The production of wood. FAO Forestry Paper. FAO, Rome FAO (2001) Forest resources assessment 2000, main report. FAO forestry paper 140. FAO, Rome FAO (2004) National forest inventory field manual template. Working paper. FAO Forest Resources Assessment Programme, Rome FAO (2010) Global forest resources assessment 2010: main report. Food and Agriculture Organization of the United Nations, Rome Ferretti M, Bussotti F, Cenni E, Cozzi A (1999) Implementation of quality assurance procedures in the Italian program of forest condition monitoring. Water Air Soil Pollut 116:71–76 Gerrand A, Lindquist E, Wilkie M, Shimabukuro Y, Cumani R, Hansen MC et al (2009) The 2010 global forest resource assessment remote sensing survey. Joint Research Centre of the European Commission, Ispra, Italy, pp 1–4 Groves RM (2004) Survey errors and survey costs. Wiley-Interscience, Hoboken Hardcastle PD, Baird D (2008) Capability and cost assessment of the major forest nations to measure and monitor their forest carbon. In: Report prepared for the Office of Climate Change, Penicuick Hartig GL (1819) Neue Instructionen f€ ur die ko¨niglich-preussischen Forst-geometer und Forsttaxatoren, durch Beispiele erkla¨rt. Kummerische Buchhandlung, Berlin Huete AR, Saleska SR (2010) Remote sensing of tropical forest phenology: issues and controversies. Int Archive Photogramm, Remote Sens Spatial Inform Sci 38:539–541 Hush B, Beers TW, Kershaw JA (2003) Forest mensuration. Wiley, New York Ilvessalo Y (1927) The forests of Suomi (Finland). Results of the general survey of the forests of the country carried out during the years 1921–1924. Communicationes Ex Instituto Quaestionum Forestalium Finlandiae 11 IPCC (2003) Good practice guidance for land use, land-use change and forestry. Institute for Global Environmental Strategies (IGES), Hayama Kaufmann E, Schwyzer A (2001) Control survey of the terrestrial inventory. In: Swiss national forest inventory: methods and models of the second assessment. WSL Swiss Federal Research Institute, Birmensdorf, pp 114–124 Kish L (1965) Survey sampling. Wiley, New York Kitahara F, Mizone N, Yoshida S (2009) Evaluation of data quality in Japanese national forest inventory. Environ Monit Assess 159:331–340

776

M. Ko¨hl and M. Marchetti

Ko¨hl M, Traub B, Pa¨ivinen R (2000) Harmonisation and standardisation in multi-national environmental statistics – mission impossible? Environ Monit Assess 63(2):361–380 Ko¨hl M, Magnussen S, Marchetti M (2006) Sampling methods, remote sensing and GIS multiresource forest inventory. Springer, Berlin/Heidelberg Lund HG (1998) IUFRO guidelines for designing multiple forest resources inventories. IUFRO, Vienna Meredith JR, Mantel SJ (2012) Project management. Wiley, Hoboken, p 608 Pancel L (1984) Agroforstliche Landnutzungsmethode spezieller subhumider Vegetationsformen in Guinea Bissau. Forstarchiv 55(5):186–194 Rossi PH, Wright JD, Anderson AB (1983) Handbook of survey research. Academic, San Diego Saket M, Altrell D, Branthomme A, Vuorinen P (2002) FAO’s approach to support national forest assessments for country capacity building, Kotka IV: Expert consultation on global forest resources assessments – linking national and international efforts, Kotka, Finland, 1-5-July 2002, Background Paper 6.6 Schmid P (1969) Die Weiterentwicklung der Leistungskontrolle in der Schweiz. Wiss. Zeitschrift d. techn. Univ. Dresden 16(2):545–549 Schmid-Haas P, Werner J, Baumann E (1978) Kontrollstichproben: Aufnahmeinstruktion. Bericht 186. Eidg. Anst. Forstl. Versuchswes, Birmensdorf, p 57 Scott CT, Ko¨hl M (1993) A method for comparing sampling design alternatives for extensive inventories. Eidg. Forschungsanstalt f€ ur Wald, Schnee und Landschaft, Birmensdorf Scott CT, Ko¨hl M, Schnellba¨cher H-J (1999) A comparison of permanent versus periodic surveys. For Sci 45(3):433–451 Stibig H-J, Achard F, Carboni S, Rasi R, Miettinen JI (2014) Changes in tropical forest cover of Southeast Asia from 1990 to 2010. Biogeosciences 11(2):247–258 Tomppo E, Gschwantner T, Lawrence M, McRoberts RE (2010) National forest Inventories – pathways for common reporting. Springer, Heidelberg, p 612 Touber L, Smaling EMA, Andriesse W, Kakkeling RTA (1989) Inventory and evaluation of tropical forest lands: guidelines for a common methodology, Tropenbos technical series. Tropenbos Foundation, Wageningen United Nations (1992) Non-legally binding authoritative statement of principles for a global consensus on the management, conservation and sustainable development of all types of forests No. A/CONF.151/26 (Vol. III), Rio de Janeiro USDA Forest Service (2012a) Forest inventory and analysis national core field guide. USDA Forest Service, Washington, DC USDA Forest Service (2012b) Quality assurance plan for annual forest inventory in the South. USDA Forest Service, Southern Experiment Station, Forest Inventory and Analysis, Knoxville. p 16 Weaver PL, Birdsey A (1986) Tree succession and management opportunities in coffee shade stands. Turrialba 36(1):47–58 Wright JD, Marsden PV (2010) Survey research and social sciences: history, current practice, and future prospects. In: Marsden PV, Wright JD (eds) Handbook of survey research. Emerald Group Publishing Limited, Bingley, p 886

Sampling in Forest Inventories Michael Ko¨hl and Steen Magnussen

Contents Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Population, Samples, Attributes, and Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Probability Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Definitions and Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sampling Designs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simple Random Sampling (SRS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Systematic Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cluster Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two-Stage Cluster Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stratified Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two-Phase Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design-Based and Model-Dependent Inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Trees on Sampling Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fixed-Area Sampling Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Point Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Point Sampling Versus Fixed-Area Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sampling at the Forest Edge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sampling on Successive Occasions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Independent Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Continuous Forest Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sampling with Partial Replacement of Plots (SPR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Errors in Forest Surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Non-sampling Inventory Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

778 778 780 781 782 783 791 793 796 798 803 806 808 808 813 817 818 820 821 822 824 828 830 832

M. Ko¨hl (*) Center for Wood Sciences, Institute of World Forestry, University of Hamburg, Hamburg, Germany e-mail: [email protected] S. Magnussen Natural Resources Canada, Victoria, Canada e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_72

777

778

M. Ko¨hl and S. Magnussen

Abstract

In sampling, a part of a population is selected and used to obtain estimates of characteristics of that population. The current chapter gives an overview on sampling methods applied in the scope of forest inventories, describes their general approaches and estimation procedures, and discusses advantages and disadvantages of the individual designs. Fixed area plots and point sampling for the selection of trees on sampling units are presented. Alternative designs for the estimation of change by sampling on successive occasions are introduced. The final section gives an overview of sampling and non-sampling errors occurring in forests surveys. Keywords

Bias • Errors • Estimation procedure • Fixed-area plot • Forest edge • Forest inventory • Point sampling • Sampling • Sampling designs • Sampling on successive occasions • Selection

Forest inventories are conducted to obtain current information about a forest or a particular area of a forest. Where areas are large, time and cost constraints rule out a total tally (census). Therefore, extensive forest inventories utilize sampling. Sampling is an example of inductive logic by which conclusions are inferred on the basis of a limited number of instances (Fuller 2009). The first part of this chapter presents some basic terms and concepts, while the second part describes several sampling procedures utilized in forest inventories. Please consult (for examples) Cochran (1977), Schreuder et al. (1992, 1993), and Ko¨hl et al. (2006) for further reading.

Basic Concepts Population, Samples, Attributes, and Estimates A population comprises all units from which the sample is taken. It may be defined very simply, for instance, the trees in a forest stand or the staff in a forest enterprise. In forest inventories, the definition of the sample population is generally based on a description of a characteristic arrangement of trees that is considered as forest. An example is given by FAO, where forest is defined as “land with tree crown cover of more than 10 % and area of more than 0.5 ha. The trees should be able to reach a minimum height of 5 m at maturity in situ” (FAO 2010). Depending on the definition, a population can be either infinite or finite. A population defined as the forest in a given region may comprise an infinite number of spatial locations but only a finite number of trees. Where the development of a population is to be assessed over time, we also need a temporal definition of a population (Roesch and Van Deusen 2013). Few populations remain constant over time, and most undergo

Sampling in Forest Inventories

779

changes; due to regeneration, mortality, and timber harvest, the number of trees in a forest is changing with time. A sample consists of a number of sampling units1 (or simply units) selected from the population by some design. The population from which the sample is selected is termed the sampled population and must match the target population for which information is to be derived. Only then can representative conclusions for the target population be drawn (Cochran 1977). Each population unit possesses a series of attributes of interest. Attribute values often exhibit a considerable variation between units (elements). The attribute may be intrinsic or derived. The tree height would be an example of an intrinsic property, which can be measured and quantified directly. A market value, on the other hand, is an example of a derived attribute – an attribute that can only be obtained via other attributes (e.g., marketable tree stem volume, assortments, timber prices). Besides measurable attributes, we may define countable attributes. For example, the attribute of interest may simply be the trees in a forest with the unit attribute being “tree.” To characterize this attribute beyond a mere count of trees, we may choose to measure associated tree-level variables such as height and stem diameter at some reference height, identify the tree species, and assess the crown form. The number of attributes to include depends on what is needed to be known about the population units of interest. A parameter is a characteristic of a population. It is usually unknown and has to be estimated. For a given population at a given point in time, a parameter is, in most cases, a fixed value and does not vary. Parameters include aggregates (e.g., total volume, total area) and averages (e.g., mean tree height) of values associated with a population element or unit. Ratios of pairs of population parameters (e.g., volume per hectare as the ratio of total volume and total area), counts (e.g., number of trees), and proportions (e.g., proportion of forest area with a specific attribute) are also examples of population parameters. A statistic is a characteristic of the sample and calculated from the sample data. As it is possible to select more than one sample from the population, the statistic will vary from sample to sample. A statistic is used to provide information about unknown parameters of the population. For example, the mean calculated from the data in a sample is used to provide information on the overall mean in the sampled population. The aim of surveys is to estimate population parameters or functions of one or more of them. The process by which sample data are applied to indicate values of unknown quantities in the population is called estimation. An estimator is any quantity calculated from the sample data and used to provide information about the unknown population quantity. For example, the estimator of the population mean μ X is μ ¼ x ¼ xi =n, where the size of the sample is n and x1, x2,. . .., xn are the values of the n sampled units. The indicated value of a population parameter

1

A synonym for sampling units is the term sampling elements (or simply elements).

780

M. Ko¨hl and S. Magnussen

derived from a sample is called an estimate. For example, if the estimator x in a particular sample is found to be 10, then 10 is the estimate of the population mean μ. Estimators can be design based or model based (Sa¨rndal et al. 2003). The estimators treated here are in most cases design based. The underlying principle behind a design-based estimator is that the population from which samples are taken is considered as a fixed entity. The random selection of units/elements to include in the sample is the only source of stochastic variation (standard error). Model-based estimation, by contrast, is conditioned on the realized sample but requires more assumptions about the model and population parameters of interest.

Probability Sampling The general principle of sampling (Fig. 1) is to select a subset of units (i.e., a sample) from a population, to measure this subset intensively, and to draw inference from the sample to the entire population. There are a many different approaches by which a sample can be selected from a population. It would seem intuitively obvious that the sample should represent the entire population. The term “representative” – as used in everyday language – suggests that the sample should be a scaled-down replica of the population. However, unless each unit in the population has an equal chance of being selected, the sample may not be “representative” in the everyday meaning of the word. Many widely used sampling designs assign varying selection probabilities to the individual units in the quest of efficiency (cost, time, standard error). The chance of being selected can be assigned with respect to an attribute or a quantitative measure of the units known in advance of sampling. A selection method complies with the conditions of probability sampling when a procedure is followed that ensures that each unit in the population has exactly the predetermined probability of being selected to the sample. The selection probabilities are used in the estimators of

Fig. 1 The principle of sampling

Sampling in Forest Inventories

781

parameters of interest and in estimators of sampling variance (Thompson 1992). The choice of selection probabilities and estimators is called a sampling strategy (Sa¨rndal et al. 2003). Given a specific population of N units, the set of all possible distinct samples, s1, s2, . . ., sv, can be defined and the units making up each sample can be designated (two samples are distinct if their union minus their intersection is not empty). If n units out of N are to be selected without replacement (a unit can only be selected once), there are 

N n

 ¼

N! n!ðN  nÞ!

(1)

possible distinct samples (Levy and Lemeshow 2008).2 For example, if a population contains 10 units and we wish to take a sample of 2 units, 45 possible samples can be selected. For a population of 200 units and a sample size of 25, the total number of possible samples is 45, 217, 131, 606, 152, 448, 808, 778, 187, 283, 008, a truly astronomic figure. For each possible sample, say si, a selection probability π si can be specified. The sum of these selection probabilities over all possible samples is – by definition –1. In the inference process, the selection probabilities are used to expand individual sample attributes/variables to an unbiased estimate of population parameters. Probability sampling methods employ a selection process that ensures that each unit in the population has exactly its designated probability of being selected under a probabilistic sampling scheme. In practice that means that any unit selected as a member of the sample must be observed (recorded), irrespectively of any problems or difficulties in obtaining the desired unit-level information. When areas are excluded – in an ad hoc manner – from the sampling frame due to, for example, cost, time, perceived or real risks, or issues of access, some units will have little or no chance of being selected. Such practices must be avoided since there is no longer a control on the probability with which the units comprising the population are selected. Such samples are called non-probability samples; they do not qualify for design-based estimation procedures.

Definitions and Notations In probability sampling inventories, data on one or more attributes are collected for each selected unit of the population. In the case of selected trees, such variables may be the diameter at breast height (DBH), an upper stem diameter above the DBH, the tree height, or species. The value reflecting a variable of a unit forming the population is defined by yi (i = 1,. . .N), where N is the number of units in the population. A sample is composed of n units selected from the N population units;

n! = n(n  1)(n  2). . .(1) and 0! = 1.

2

M. Ko¨hl and S. Magnussen

782 Table 1 Population values and sample estimators of common parameters

Parameter Meanb

Population value

Total

Y

Ratio

R

Proportion

P

Y

Sample estimatora n X ^¼1 Y yi n i¼1

^ Y^ ¼ N Y

Xn y y Y^ i¼1 i ^ R ¼ ¼ ¼ Xn ^ x X x i¼1 i Xn δ i¼1 i P^ ¼ n

Simple random sampling; δ is an indicator variable specifying that a unit belongs to a given category, say A. δi is 1, if the ith unit attribute has attribute A and 0 otherwise b For finite populations a

the attribute values of a selected unit is denoted by yi (i = 1,. . .n). The sampling fraction is the proportion of units selected from the population, specified by the ratio n/N. Capital letters refer to population attributes (i.e., parameters), and lowercase letters refer to sample attributes (i.e., statistics). Sampled units are used to estimate parameters for the population. The four most important population parameters are: 1. The mean, Y (e.g., the mean tree height in a forest stand) 2. The total, Y (e.g., the total timber volume in the inventory area) 3. The ratio R between two means or totals (e.g., the proportion of fuelwood in the total growing stock) 4. The proportion, P, of units in a categorical class (e.g., proportion of a particular species) The sample provides us with estimates of population parameters. Estimates are distinguished from their true population values by adding a “^” (caret) placed above the associated symbol. The relationship between population values and sample estimators is given in Table 1 for the most common population parameters. The way in which means and totals are derived in different sampling procedures is described in section “Sampling Designs.”

Sampling Designs Sampling designs can be divided into two main groups (see Fig. 2): – Sampling designs without auxiliary information – Sampling designs with auxiliary information

Sampling in Forest Inventories

783

Fig. 2 Sampling designs (after Pelz and Cunia (1985)), numbers in brackets indicate the chapters dealing with the respective design

In sampling designs without auxiliary information, only the observations on the variables of interest are used to derive the parameters. In addition to the observations on the variables of interest, other information linked to the variables of interest is often available or can easily be obtained, e.g., from aerial photos. Designs that exploit auxiliary information in the estimation phase are plenty. As a rule, sampling designs with auxiliary information are more efficient than those without. The major types of the two groups are shown in Fig. 2.

Simple Random Sampling (SRS) Simple random sampling is not widely used in forest inventory applications. However, we begin the detailing of common inventory sampling designs with SRS because a presentation of SRS and its estimators will make it easier to appreciate more complex designs and their estimators.

General Description Given a finite population, SRS selects n units from the N units comprising the entire population. The probability of a single unit or element to be selected is n/N. The selection process is done in a way that all possible samples of size n have the same selection probability. As there are 

N n

 ¼

N! n!ðN  nÞ!

784

M. Ko¨hl and S. Magnussen



 N . The principle n of equal selection probabilities extends to infinite populations (e.g., points on a surface), but we have no means of calculating the selection probabilities as N is infinite (Mandallaz 2008). There are two types of SRS: SRS with replacement and SRS without replacement. In SRS with replacement, all units have the same selection probability n/N in each draw, so a single unit may be selected more than once. When a unit is selected more than once, no new information about the population is provided. Hence sampling with replacement is for small populations with a relatively high sample fraction less efficient than sampling without replacement. In sampling forest populations, it is rarely an issue of importance. In SRS without replacement, a selected unit is removed from the sampling frame before the next unit/element is selected. Thus, for a distinct unit, remaining in the sampling frame after completion of k draws the probability of selection at the (k + 1)th draw is (nk)/N and so on for k = 0,. . ., n1. When the sample sizes are small compared to the size of the population of interest, the differences will often be trivial. This situation holds true for most forest inventory applications. Therefore, only SRS without replacement is considered in the following. The simplest way of selecting sample units is to number all the units in the population; choose n numbers randomly from the complete list of enumerated units. Here, however, it must be ensured that all population units have been listed – but this is rarely if ever possible in forest inventories. In forest inventories using SRS, remote sensing imagery, aerial photographs, or a map is needed to establish a frame from which the sample is to be taken. Sample point coordinates are randomly chosen and the survey is then conducted at the corresponding points (Fig. 3). These coordinates may designate, for example, the centers of fixed-area plots, point samples, stem distance methods, or the location from which to include the nearest tree (see section “Selection of Trees on Sampling Units”). possible samples, the selection probability of each sample is 1=

Fig. 3 Allocation of samples in SRS

Sampling in Forest Inventories

785

Estimation Procedures Population Mean The population mean is given by Y¼

N 1X y N i¼1 i

(2)

^ is an unbiased estimator of the population mean: Under SRS, the sample mean Y ^ ¼1 Y n

n X

yi

(3)

i¼1

Variance, Standard Deviation, and Coefficient of Variation The units of a population or sample are not identical with respect to the attributes of interest. The degree of variability may differ from population to population, thus forming an essential characteristic. One measure of this variability is termed the variance. This defines the dispersion of individual unit values about their mean. The variance of individual unit values yi in a population, Var(yi), is defined as N  X

Var ðyi Þ ¼

yi  Y

2

i¼1

(4)

N1

where N is the population size and Y is the population mean. The standard deviation of the population is the square root of the variance Var(yi): Sdevðyi Þ ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Var ðyi Þ

(5)

Where SRS is employed, a sample-based design-unbiased estimator of the ^ ðyi Þ, is population variance, Var Xn  ^ ðyi Þ ¼ Var

i¼1

^ yi  Y

n1



2

(6)

^ is the sample mean. A sample-based estimator of where n is the sample size and Y the population standard deviation of yi, (Sˆdev(yi)), is obtained by taking the square ^ ðyi Þ. root of Var It is often convenient to remove the effect of measurement scale from estimators of variability. Variability expressed in relative terms – with respect to the mean of the variable to which they refer – is scale invariant. The coefficient of variation (CV) is an example of a popular scale-invariant measure of variation:

M. Ko¨hl and S. Magnussen

786

CV ðyi Þ ¼

Sdevðyi Þ Y

(7)

Sample-based estimators of the coefficient of variations are obtained by replacing the population quantities by their respective estimators. Standard Error and Confidence Intervals The combination of sample size and inherent variability in a sample of attribute values determines the uncertainty in an estimate. The standard error is a widely used statistic quantifying the uncertainty in an estimate. For a given sample size, a greater inherent variability among attribute values leads to a greater standard error3 in an estimate of interest and vice versa. As we shall see, it is not necessary to take several samples from the same population in order to determine the standard error. Instead we make use of the central limit theorem (CLT) which says that the mean of n randomly selected population values of a variable is asymptotically ðn, N ! 1Þ normally distributed with a variance that is the variance of the random variable divided by n (Casella and Berger 2008). If the sampling fraction n/N is small, an ^ ^ estimator of the standard error SE Y of an estimate is  rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ffi ^ ^ ^ ^ SE Y ¼ Var Y

(8)

where  Var ^ ¼ ^ ðyi Þ ^ Var Y n

(9)

The standard error depends on the variability of the units in the sample and the sample size. The concept of standard error is often not intuitively clear to many users of inventory data and they may find it difficult to assess the significance of a standard error and interpret it correctly. Estimates arising from a sample-based inventory ought to include a measure of uncertainty of the estimates. Confidence intervals for sample estimates provide an intuitive easily understood measure of the impact of a standard error. A confidence interval for an estimate gives the range within which one can expect the true population parameter to be located. The bounds of the confidence interval are termed confidence limits. The interval should have the property that the probability of the true value being located within the confidence limits is known, say 1α. The quantity 1α is called the confidence coefficient and the interval is called the 100(1α)% standard interval. Typically the 95 % confidence interval is presented (i.e., α = 0.05). It is expected that the 95 % confidence interval covers for 95 out of 100 replicate samples of size n the true value of the population. 3

The term “sampling error” is sometimes used as a synonym for standard error.

Sampling in Forest Inventories

787

Conversely, there is a 5 % chance that the true value is outside this interval. A specific sample-based estimate of the confidence interval either includes the true value or not. The distribution of sample estimates under repeat sampling is usually assumed to be normal (invoking CLT) with a mean equal to the obtained estimate and a variance equal to the estimated variance divided by the sample size. Under this assumption and continuing with the example with  above a population mean as the ^ ^ parameter of interest, the lower Y and upper Y limits of the 100(1α)% L

U

confidence interval for a sample-based estimate are  ^t ^ ¼Y ^ ^ Y L n1,1α2  SE Y

(10)

 ^þt ^ ¼Y ^ ^ Y U n1,1α2  SE Y

(11)

The symbol tn1,q is the qth quantile of Student’s t-distribution (Casella and Berger 2008). For large n and 95 % confidence probability, t is approximately 2, and the confidence interval is called the 95 % confidence interval. For a 68 % confidence probability, t is approximately 1. The question arises when to use the t-distribution and when to use the z-distribution for constructing intervals. Under the CLT, the sampling distribution of the sample mean ðyÞ is approximately normal regardless of the shape of the original distribution of the variable. When n  30, the sampling distribution of the sample mean becomes approximately normal. If n 30, skewness and heavy tails in the distribution of population values of an attribute will influence the shape of a sampling distribution of a parameter estimate and slow the rate at which it approximates a normal distribution with increasing sample size. Therefore, it is not advisable to assume that the sample distribution is normal. Instead, we will use a t-distribution, which is designed to give us a better interval estimate of the mean when we have a small sample size (Zar 2010).

Totals

^ by the number of units in the The total Y^ is calculated by multiplying the mean Y population: ^ Y^ ¼ N Y

(12)

The standard error of the estimated total volume is    ^ ^ Y^ ¼ N SE ^ Y SE

(13)

Confidence intervals for totals are calculated in analogy to Eqs. 10 and 11. It is assumed that N is known, when applying Eqs. 12 and 13. When N is not known, it has to be estimated and is thus associated with a standard error.

M. Ko¨hl and S. Magnussen

788

Sampling for Proportions and Percentages Forest inventory results for discrete and categorical attributes may be presented in terms of counts, proportions, or percentages. Examples would be the number of commercial species, proportion of dead trees, or percent of teak volume in a concession area. Counts, proportions, and percentages usually involve units belonging to a defined class or exhibiting a given characteristic. Where there are more than two mutually exclusive classes, the term “multinomial variable” is used. The results obtained on the basis of multinomial variables are presented as classifications of nominal scale (e.g., tree species, soil type) or ordinal scale (e.g., timber quality, stand layer). Multinomial variables are frequently analyzed and presented as a sequence (vector) of proportions summing to 1.0. The following discussion is limited to the relatively simple case of SRS where each sampled unit exhibits a binary class value. There is extensive literature on the more complex analysis of proportions and percentages in other sampling designs (e.g., Cochran 1977; Agresti 1992; Lloyd 1999). Binary variables assume one of two values, typically the value yi = 1 when they belong to a given class and yi = 0 otherwise. The number of population units in the class assigned a value of 1 is given by Y¼

N X

yi ¼ N P

(14)

i¼1

and their proportion P by XN P¼Y¼

y i¼1 i

N

(15)

The complementary proportion of units that do not have the class attribute is Q = 1  P. The population variance of yi is hereafter Var ðyi Þ ¼

N 1 X N Pð 1  PÞ ð y i  PÞ 2 ¼ N  1 i¼1 N1

(16)

Sample-based estimators of the population proportion and its variance are obtained from the above equations after substituting n for N and adding carets to distinguish them from the true population values: XN ^¼ P^ ¼ Y ^ ðyi Þ ¼ Var

y i¼1 i

n

N   2 1 X n ^ P 1  P^ yi  P^ ¼ n  1 i¼1 n1

(17)

(18)

Sampling in Forest Inventories

789

For large n, a good approximation to the standard error of P^ is given by sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ffi ^ ^   P 1P ^ P^ ’ SE n1

(19)

The confidence interval for P^ can be calculated from 2 P^  4tn1;α=2

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ffi P^ 1  P^ n1

3 15 þ 2n

(20)

where the term 1/2n is a correction for continuity, which is necessary as P is not a continuous variable (Cochran 1977; Zar 2010). If more than one observation is made on the binary trait in every sampled unit, the estimation of proportions has to be modified since the number of observations per sample unit can vary. Fixed-area plots are a typical example. The number of, say, trees per plot varies naturally among plots. The estimators of P^ and the variance of P^ are now given by mi n X X

P^ ¼

i

j¼1

n X i¼1

n X

yij

mi

¼

i¼1 n X

a (21)

mi

i¼1

where the subscripts i refer to sample unit and j to the population units in the ith sample unit. There are mi population units in sample unit i of which ai units belong to the binary class given a value of 1. The corresponding estimator of variance becomes   1 v^ar P^ ¼ 2 m

Xn i

Xn Xn a2i  2P^ i ai mi þ P^2 i m2i n1

(22)

where m is the mean number of population units per sample unit.

Ratio Estimates Ratio estimators are widely used in forest inventories. An attribute related to an area such as the number of trees per hectare or the volume per hectare is provided by a ratio estimator. The population ratio R is obtained by dividing the population total of the attribute (total volume, total number of stems) in the numerator of the ratio by the population total of the attribute in the denominator of the ratio. A sample-based estimator of R is the ratio of the two sample estimates of population totals:

M. Ko¨hl and S. Magnussen

790

Xn R^ ¼ Xi¼1 n

yi

x i¼1 i

¼

^ Y^ Y ¼ ^ X^ X

(23)

This is a ratio of means estimator, which has a bias of the order of 1/n. There is no unbiased sample-based estimator of R. For sample sizes n over 30, the bias is often negligible, but skewed population distributions of Y and especially X can introduce a serious bias in a sample estimate (Rao 1988; Hess and Bay 1997). The variance of a ratio of means estimate has a similar structure as the estimator in Eq. 22: ^ ^ ^ ^   ^ ^ R^ ¼ Var ðyi Þ þ R Var ðxi Þ  2R Covðyi , xi Þ Var ^2 nX

(24)

^ where Cov(y i, xi) is the covariance between the two attributes/variables y and x. The covariance is estimated as Xn ^ ðyi , xi Þ ¼ Cov

y x  n1 i¼1 i i

Xn

n1

y i¼1 i

Xn

x i¼1 i

(25)

It is often possible to estimate a ratio Ri for each plot in a sample, for example, the number of trees per ha in a plot. However, the estimation of the population ratio should not be based on these individual ratios because the mean of these individual ratios, as an estimate of R, has more bias than the ratio of means even if n is large (Cochran 1977). The exception is when yi ¼ R  xi for all units in the population. In this case, the mean of yi/xi is obviously R everywhere. Furthermore, Ri is often unstable and exhibits a large variance and a skewed sampling distribution. Determining Sample Size A sufficient SRS sample size is determined by the inherent (natural) variability of the attribute values in the population, the degree of precision required for the results, and the choice of confidence coefficient. In SRS, the sample size needed to satisfy a desired width (W ) of a 100ð1  αÞ% confidence interval is calculated, for say a mean, according to n¼

t2n1,1/=2 Var ðy^Þ W2

(26)

In practice we would normally not know the variance of the variable of interest. It must be replaced by an estimate derived from either historic information, related surveys, or from a qualified expert. Prudence dictates a conservative estimate. The variance Var(y) may be estimated either from previous data or a pilot study. Where neither is available, a rough estimate can be made using (Snedecor and Cochran 1989)

Sampling in Forest Inventories

791

var ðyÞ ffi

maxðyi Þ  minðyi Þ 12

2 (27)

Advantages and Disadvantages of SRS Strict adherence to the principles of simple random selection guarantees unbiased and consistent estimates of population parameters and their standard errors. Yet there are often other sampling designs for which the expected standard error for a given sample size is lower than the standard error expected under SRS. Relative to a more efficient design, an SRS requires a greater sample size for a given expected standard error, which typically translates into higher costs. Note, however, that the expected efficiency of a design relies on theoretical expectations. The survey planner has to obtain estimates of the expected standard error under different competing designs and their cost implication before a rational choice is possible. An SRS design often requires surprisingly large investments in organization, checking, and location of the samples, investments that can be more time consuming and more expensive than for other more efficient designs. Also, through random selection, an irregular spatial distribution of sample locations is a likely outcome with moderate to small sample sizes typical in forest inventories. Thus the population as a whole may not be uniformly represented by the sample. Although this outcome is expected under an SRS design, it is clearly unsatisfactory and perhaps even unacceptable to proceed to estimation and inference with a sample that one suspects will yield estimates far from the true population value(s). For these reasons, the SRS design is mostly limited to smaller homogeneous subpopulations as part of a more complex design.

Systematic Sampling A form of systematic sampling is the procedure most commonly used in forest inventories. As the term implies, the samples are not randomly distributed over the inventory area but arranged in a systematic pattern, usually either a squared grid or a triangular network. Thus the sample consists of one randomly selected unit, which determines the location of all other units in the sample (Fig. 4). The application of a systematic grid results in a maximal minimum spatial distance between sample units. Examples for a systematic layout of sampling units are given by Pandey (2008) or Veloso de Freitas (2009). In systematic sampling, the necessary sample size is usually determined on the basis of optimizing the SRS, which eventually leads to somewhat more accurate inventory results than originally anticipated (see section “Systematic Sampling,” “Estimation Procedures”). In a systematic sampling from a squared grid, the sample size determines the scale of the grid to be used for implementing the systematic selection process. For a population occupying a square with an area F and a desired

792

M. Ko¨hl and S. Magnussen

Fig. 4 Systematic sampling grids

sample size of n, the appropriate grid spacing between sample locations should be pffiffiffiffiffiffiffiffi F=n . For those who decide to use a triangular grid as the frame for a pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi systematic sampling process, the distance between points becomes 2=3 ⋆F=n ’ pffiffiffiffiffiffiffi 1:075 ⋆ F =n when the population occupies a square. In a triangular grid, the pffiffiffiffiffiffiffi distance between rows and columns is 0:557 ⋆ F =n . To facilitate the location of the sample units in the field, it is common practice to round off distances to the nearest 0.1 m. In establishing the grid, however, it should be borne in mind that some points may fall outside the target population. The density of the grid should be increased according to the proportions of sample units expected to fall outside the population of interest. The proportions can be nontrivial when the outline of the forest population is highly irregular or the forest is composed of many smaller, irregularly shaped spatial patches.

Estimation Procedures It is generally impossible to provide unbiased estimators of the variances when systematic sampling is used. Attempts have been undertaken to find estimators with little bias and low variance (Bellhouse 1985; Wolter 1985; Sherman 1996). The most commonly used approach is based on the assumption that a systematic sample is equivalent to a random sample. However, this assumption holds only when population attributes are randomly distributed over the population. With the assumption of SRS equivalency, means and totals are computed using the formulae applicable to SRS (see section “Simple Random Sampling,” “Estimation Procedures”). SRS estimators of standard errors applied to estimates from systematic sampling are usually conservative; they overestimate, on average, the actual error. An overestimation of about one third is not unusual, but more extreme results have been reported (Hartley 1966; Bellhouse 1988; Stehman 1992). An interesting alternative to the systematic placement of sample locations on a grid is a spatially balanced design with unequal probability sampling (Grafstro¨m 2010).

Sampling in Forest Inventories

793

Advantages and Disadvantages of Systematic Sampling The main advantages of systematic sampling are the uniform coverage of the population and the generally better efficiency than with SRS. As a rule, a “spatially balanced” sample design will have a lower root mean square error when sampling from a population with patterned variation (Matern 1980; Olsen et al. 1999; Stevens and Olsen 2004). As the joint selection probability of selecting two distinct population units in the sample is either positive or 0, depending on the systematic sampling protocol, the selected units are not independent of each other. This feature makes systematic sampling fundamentally different from SRS. In systematic sampling, it is assumed that the population exhibits a random spatial pattern with respect to the attributes of interest. This assumption can be violated where the orientation of a grid coincides with regularly arranged features with an effect on the survey attributes. Examples include rivers and topographic features like valleys and mountain ranges. In those circumstances, a systematic sampling may lead to substantial bias in the estimates of interest.

Cluster Sampling Every sampling design rests on the assumption that the population is composed of a finite countable number of clearly defined units. In cluster sampling, two or more elements or two or more units are included in the sample at each sample location. The inclusion of two or more units at each sample location intensifies the sampling effort at each sample location by simultaneously reducing travel costs. One example for such clustering is the establishment of several sample plots in a fixed geometric configuration at each sample location instead of just a single plot (Fig. 5). The grouping of units into clusters lent the procedure its name. Allocating a fixed number of plots in clusters generally reduces assessment cost, as the unproductive traveling times to plots can be reduced. However, if cluster sizes become too large, the efficiency of the design suffers. The trade-off between cost savings due to large cluster sizes and meeting a target precision on estimates of population parameters depends on the distribution of the variance of attribute values across spatial and temporal scales. If units/elements in a cluster are more alike than units/elements selected at random, then we do not learn as much about the population from one cluster with m units/elements as we would from m independent units/elements. Thus, for a cluster sampling approach to be attractive in terms of precision for a given overall number of sampled units/elements, the variation within a cluster must be large relative to the among-cluster variance (Saborowski and Smelko 1998; Barnett and Stohlgren 2003; Gray 2003b). For homogeneous forest areas, only a few units per cluster are needed to obtain a sufficient picture of the population. On the other hand, the surveying of large clusters in areas where conditions vary greatly within small distances is still justifiable since there is little redundancy in the collected information. In order to optimize cluster sampling, a balance between cluster size and costs must be found;

794

M. Ko¨hl and S. Magnussen

Fig. 5 Spatial arrangements of plots in clusters

this is decisively influenced by the variability and the spatial distribution of the units (Magnussen 2002). Cluster sampling is frequently applied in forest inventories. Examples are the national forest inventories of Brazil, India, Suriname, the USA, and many European countries (Tomppo et al. 2010). At each sample location, plots are allocated in a fixed geometric configuration such as square, rectangular, or more complex forms within each cluster. The term cluster is rarely used explicitly for the spatially grouped plots. Instead the word “tract” has become widely accepted as a quasisynonym for a cluster (of sample plots). An early example of cluster sampling is the camp-unit system, introduced in Thailand for inventorying teak stands (Loetsch 1957). In this system, a camp located in the center of a cluster (i.e., sample location) is surrounded by what are termed satellites, each satellite comprising several sample plots which can be surveyed by a field team within a single day.

Estimation Procedures The simplest form of cluster sampling is the surveying of clusters of constant size. To facilitate the understanding of cluster sampling, this version of cluster sampling is detailed. However, in practical applications, cluster sizes are rarely constant. Fixed-area sample plots are in effect clusters of trees and it is obvious that the cluster size, i.e., the number of trees per plot, changes from cluster to cluster. Even clusters designed with a fixed number of plots may exhibit unequal sizes, as some plots in a tract (cluster) may be located outside the population boundary or straddle the boundary. Equations 28 and 29 relate to clusters of equal size. Where clusters extend beyond the population limits, clusters with varying cluster sizes are obtained and

Sampling in Forest Inventories

795

require the application of more complex equations for clusters with unequal cluster sizes (Cochran 1977; Sukhatme et al. 1984; Ko¨hl and Scott 1994). The sample-based estimator of the population mean for n clusters of equal size m selected by simple random sampling is 1 ^ Y clust ¼ n

n m n X 1X 1X ^ yij ¼ Y i m n i¼1 j¼1 i¼1

(28)

where yij is observed value for the jth unit within the ith cluster (i = 1,. . .n, j = 1,. . .m). This is simply the average of cluster means computed over the units ^ . The estimator of the sampling variance of the with the ith cluster, i.e., Y i

estimated population mean in a finite population composed of N clusters is  n n   2 1 X ^ ^ ^ Y N ^ Var Y Y ¼ clust i clust n  1 i¼1

(29)

This is simply the among-cluster variance of cluster means. The corresponding estimator, for sampling with clusters of unequal size, is a weighted average of cluster means with weights proportional to cluster size (Cochran 1977; Ko¨hl and Scott 1994). Estimators for population totals are obtained by ^ Y^clust ¼ N  Y clust    ^ ^ Y^clust ¼ N 2  Var ^ Var Y clust

(30) (31)

Advantages and Disadvantages of Cluster Sampling Cluster sampling is generally less efficient than SRS for the same number of sample plots (n in SRS and n m in cluster sampling). The efficiency of cluster sampling depends on the similarity of observations within a cluster and can be quantified by the intraclass-correlation coefficient ρc:  Cov yij , yi0 j  ρc ¼ Var yij

(32)

where yij is the value of the ith unit in cluster j, with i = 1,. . .m and j = 1,. . .,n. 1 ρc 1 . For clusters of equal size, ρc can take values in the range of  m1 With increasing cluster size m, the lower (negative) limit approaches 0. The variance of the estimated population mean can be rewritten as (Sukhatme et al. 1984)

796

M. Ko¨hl and S. Magnussen

  N  n Var ðY Þ Var Y clust ¼ (33) f1 þ ðm  1Þρc g N  1 nm     When the cluster size m is 1, ρc is zero and v^ar Y clust is equal to v^ar Y , the variance obtained by simple random sampling. Increasing the cluster size reduces the efficiency of cluster sampling but might reduce the assessment cost disproportionately. The magnitude of ρc depends on the spatial pattern of the population and the distance between units of the cluster. As mentioned before, a major advantage of cluster sampling is the reduction of unproductive travel time. This holds especially true in remote areas. Besides cost, the optimal cluster design is depending on features such as the cluster size, the geometrical arrangement of clusters, and the distance between units of the cluster. Thus designing a cluster needs to take into account the spatial pattern and autocorrelation structures of the population which may prove difficult when critical information is lacking. Large clusters can also be used to derive information about the variation in the population at scales covered by a cluster (Diggle and Ribeiro 2007; Chilés and Delfiner 2012).

Two-Stage Cluster Sampling Two-stage sampling offers an option for reducing inventory costs through aggregating units. The method assumes that the units of a population can be further subdivided into smaller units. In two-stage cluster sampling or simply two-stage sampling, the entire population is divided into N clusters. A sample of n clusters is selected. The ith cluster is assumed subdivided into Mi equally sized smaller units called second-stage units. A sample of mi second-stage units is taken from the ith cluster. This is termed subsampling since the attributes of interest are observed (recorded) in fewer than Mi subunits within each cluster. The staged sampling design is not limited to two stages but can be extended to more stages. Figure 6 Fig. 6 Two-stage cluster sampling

Sampling in Forest Inventories

797

gives an example of two-stage sampling, where a sample of six clusters is selected. Each cluster is subdivided into 9 s stage units, from which two are randomly selected. Two-stage procedures are frequently used in forest inventories. In the simplest incarnation, every cluster contains the same number of secondary units, and both clusters and secondary units within clusters are randomly selected at each stage. Two-stage sampling is particularly attractive when access to individual sample locations is time consuming and costly, a topical situation in many tropical forest regions. Two-stage sampling becomes more complicated when the number of secondary units per primary unit varies and sampling techniques other than SRS are employed. These may include proportional sampling or stratification. Each of the resultant possible combinations requires its own specific algorithms for computing population parameters and standard errors. Bowden et al. (1979) give an excellent description of multistage sampling procedures.

Estimation Procedures With SRS sampling of n clusters out of N population clusters of size M and SRS of m (m < M) secondary units within the ith cluster, an unbiased two-stage estimator of the population mean based on all the m  n units in the sample is (Cochran 1977; Sukhatme et al. 1984) n X m n 1 X 1X ^ ^ Y yij ¼ Y clust2 ¼ i nm i¼1 j¼1 n i¼1

(34)

where n is number of selected clusters, m is the number of selected subunits per cluster, and yij is the value obtained for the jth subunit in the ith cluster. The samplebased estimator of the variance of the above-estimated mean is (Cochran 1977; Sukhatme et al. 1984) 2 n   n Xn  ^ m ^ ^ 2 n X m  1 1 Y  Y y  Y X i clust2 ij i i¼1 ^ N M ^ Var þN Y clust2 ¼ mn n n1 n ð m  1 Þ i¼1 j¼1 (35) The first term on the right side of Eq. 35 is the variance of the cluster means (i.e., the between-cluster variance), while the second term is the variance of the means of the clusters (i.e., the within-cluster variance). When n/N is negligible, a good approximation of the standard error can be computed from the simpler equation: 



^ ^ Var Y clust2 ¼

Xn  i¼1

^ ^ Y Y i clust2

nð n  1Þ

2 (36)

798

M. Ko¨hl and S. Magnussen

When all secondary units in all clusters are sampled (mi = Mi for all i), we revert to estimators appropriate for single-stage cluster sampling. For n = N, that is, all clusters are sampled, but mi < Mi for at least some i, the two-stage estimator is identical to the estimator for stratified random sampling with clusters serving as strata (see section “Stratified Sampling”). The above two-stage estimators assumed that the population was divided into a unique set of N clusters that, in turn, were subdivided into a fixed number of secondary units. N would be known in this situation. When first-stage clusters are merely a fixed-area-sampling device located at random in the population, N is unknown and must be estimated by dividing the area of the population by the area of a first-stage unit. With a small number of inventory plots placed at random or in some geometric configuration inside a first-stage unit, we do not a priori know M but we can estimate M by dividing the area of first-stage unit (cluster) by the area of a second-stage unit (plot).

Advantages and Disadvantages of Two-Stage Cluster Sampling Two-stage sampling is particularly suitable for forest inventories in which access to the sample plots is costly and thus especially useful for tropical forests. Two-stage sampling is often proposed as a technique for combined inventories, i.e., inventories in which data are collected from both terrestrial assessments and remote sensing imagery. The use of remote sensing imagery allows for a subdivision of the inventory area into primary units. If a full coverage of the inventory area by remote sensing imagery is not possible, modified two-stage techniques have to be applied (Saborowski 1990; Gregoire et al. 2011). Two-stage sampling competes, to some degree, with two-phase sampling (see section “Two-Phase Sampling”). Often, a two-phase sampling design is more efficient when the first phase is not used exclusively in the selection process but also in the estimation phase by providing auxiliary information on the units collected in the first phase.

Stratified Sampling In stratified sampling, we use auxiliary information to partition the population into subpopulations or strata. Strata are formed based on more or less similar attribute values or administrative, jurisdictional, geophysical, or other characteristics. The strata should be mutually exclusive in a way that a population unit can be assigned to only one stratum. Stratification aims at dividing a population into a number of parts, which are as homogeneous as possible. This allows for partitioning the total sums of squares (SS) of the population values into two components: (1) the SS within the strata and (2) the SS among strata. Stratification aims at making the contribution of the among-strata SS to the total SS as large as possible. Once stratification is done, samples are selected independently in each stratum. Estimates are calculated for individual strata, and afterwards strata estimates are combined to population estimates. The fundamental attraction of stratified sampling is a potential gain in efficiency: for a given sample size, the achievable precision improves

Sampling in Forest Inventories

799

with decreasing population variability. By applying this idea to several more homogeneous strata, the overall precision of a population estimate improves compared to an estimate obtained from a simple random sampling design with the same number of sampled units. Besides improving the variance efficiency of estimators, other reasons to choose a stratified sampling design are as follows: i) estimates for defined subpopulations (strata) may be required, (ii) the desired precision is not the same for all subpopulations, (iii) assessment cost and/or attributes of interest are not the same for all subpopulations, and (iv) different sampling protocols apply to different subpopulations. A diverse spectrum of criteria can be used to stratify a population. Some examples are major timber type, vegetation type, stand structure, species mixtures, site quality, protective status, habitat, ecological sensitivity, wetland status, recreational use, non-timber resource values, and political and administrative units. Where satellite imagery provides the auxiliary information, the stratification is often done on the basis of the value of various indicators of vegetation types, such as the normalized difference vegetation index (Wulder and Lavigne 1996; Baker and Wilson 2000; Sims and Gamon 2003; Zhu et al. 2003; McRoberts et al. 2012; Laurans et al. 2014; McRoberts et al. 2014; Schroeder et al. 2014; Schlund et al. 2015). Cochran (1977) and Dalenius and Gurney (1951) give generally valid rules for the selection of an optimum stratification. The Jenks natural breaks classification method assigns values to a given number of strata with the objective of minimizing variances within classes while maximizing between class means (Jenks 1967). In multi-resource inventories, the consideration of different attributes may result in substantial differences in what constitutes an optimal stratification. For many practical applications, 5–10 strata appear to give substantial reductions in variance. Cochran (1977) gives a very good example of this issue and argues for why the gain by adding yet another stratum decreases very quickly. Units within a stratum are selected independently from the selections of units in other strata. This accommodates strata-specific sample sizes, selection criteria, and survey methods. When SRS is applied in all strata, the procedure is termed stratified random sampling. In assessment, the strata are first evaluated separately and the results are then compiled to give population-level estimates of, for example, means (totals) and standard errors. The fact that stratified sampling provides strata specific estimates of interest is a distinct and important advantage of this design.

Sample Allocation Deciding the number of samples to take in each stratum is perhaps the most important decision the inventory designer has to make after deciding on a stratified sampling design. The allocation of samples to strata may be done in various ways. The decision is often one of allocating a fixed total sample of size n to individual strata. The expected precision and cost of the resulting design can then be approximated from subject knowledge, experience, or qualified guesses. Design

M. Ko¨hl and S. Magnussen

800

alternatives for different n and for different attributes of interest can then be compared, and the one judged most attractive against a set of global objectives is typically favored. There are generally three common approaches to allocate the total sample size to strata: Equal allocation: an equal number of samples are taken in each stratum, nh = n/L. Equal allocation, however, is rarely effective, as smaller strata are sampled with disproportionately higher intensities than larger strata. Proportional allocation: this popular allocation scheme generates strata sample sizes in proportion to the size of the strata. The size of stratum h (Nh) is measured in the number of units. With this approach, the sample size in stratum h becomes nh ¼

Nh n: N

(37)

Optimal allocation or “Neyman allocation”: At times, the inventory designer will have some ideas or estimates of the expected strata-specific variance of the attribute of interest, s2h. When both the within-stratum variance and stratum size are considered together in the allocation problem and the objective is to minimize the expected variance of an estimate of a population total (mean), the solution is termed optimal allocation. With this approach, the sample size in stratum h becomes N h sh nh ¼ X n N h sh

(38)

Cost constraints may necessitate a shift away from these allocation schemes toward an affordable design. Further criteria may be applied in selecting the allocation, for instance, differences in costs for various survey methods or differences in the importance of the strata.

Estimation Procedures For the hth stratum, the estimators of the mean and the variance for the stratum are as follows: nh X ^ ¼ 1 Y y h nh i¼1 hi

 ^ ¼ v^ar Y h

nh  1 X ^ 2 yhi  Y h nh  1 i¼1

(39)

(40)

where yhi is the value obtained for the ith unit in stratum h. The estimators for the population mean and its variance under stratified random sampling with L strata are given in Eqs. 41 and 42, respectively:

Sampling in Forest Inventories

801

^ Y STR ¼ 



^ v^ar Y STR ¼

L L X X Nh ^ ^ Yh ¼ Wh Y h N h¼1 h¼1

 ^  W 2h v^ar Y h

XL

nh

h¼1

1

(41)

nh Nh

 (42)

The variance estimator is simplified if the sample fraction in each stratum is negligible and if the sample allocation is proportional to the stratum size (e.g., area). It frequently happens that we do not know the total population size N nor the size of individual strata (Nh, h = 1, . . ., L). In these circumstances, it would be incorrect to replace the actual strata weights Wh h = 1,. . ., L by sample-based weights wh ¼ nnh as it leads to biased estimators. A bias will remain constant even as the sample size increases. When the attribute of interest is expressed in units per unit area, the area of the population A and the area of individual strata (Ah) are used instead of N, viz., Nh. Area-based strata weights then replace the weights based X on A, the size of a unit, viz., an element. For a stratum area Ah and total area A ¼ h h the area weight for stratum h becomes W h ¼ Ah  A1 . When strata areas are known to within a negligible error – a situation that is common when the strata information comes from a classified remotely sensed image – the bias arising from using estimated weights w^h in place of the true area weights Wh can safely be ignored. If proportional allocation is used and Wh is replaced by the area proportion of stratum h, most of the potential gain of a stratification compared to simple random sampling will, in most cases, be retained.

Post-Stratification The term post-stratification applies to a procedure for which samples are stratified to a set of known strata after the completion of the sampling. In other words, auxiliary strata information is used upon the completion of the sampling process. Post-stratification may be an option when a field survey is completed before the auxiliary information (e.g., remotely sensed data) became available. Poststratification facilitates forest surveys, as field sampling and analysis or interpretation of remote sensing data can be done with fewer constraints on timing. However, it should be emphasized that the ensuing allocation of the samples to the post-strata may not safeguard any gains in efficiency over and above what is provided by the original sampling design. In its simplest form, post-stratification applies to data from an SRS. Using the previous notation for stratified random sampling – but for an addition of a suffix “.ps” to distinguish a post-stratum from an a priori stratum – the post-stratification estimator of the population mean becomes ^ Y STR:ps ¼

L:ps X h:ps

^ W h:ps Y h:ps

(43)

M. Ko¨hl and S. Magnussen

802

and, assuming we can ignore the finite population correction factor, the estimator of the sampling variance of the post-stratified estimate of the population mean is 

^ var Y STR:ps



" # L:ps L:ps  X     1 X 1 ^ W h:ps v^ar Y W h:ps 1  W h:ps v^ar yh:ps ¼ h:ps þ n h:ps nh:ps h:ps (44)

The first term in the variance estimator is identical to the variance under stratified sample with proportional allocation of sample sizes and within-strata SRS. The second term reflects an increase in the variance due to the random nature of the strata weights. The strata weights in post-stratification are the expected value of a random binary variabletaking thevalue of 1 if a sample is in stratum h and zero otherwise. The term W h:ps 1  W h:ps is the well-known variance of a binomial random variable (Snedecor and Cochran 1989). As before, we can replace the above strata weights based on strata sizes in terms of population units with area-based weights. The above post-stratification estimators are not changed if the initial sample is not obtained under SRS but from a systematic sample. The implicit proportional allocation is more likely to be satisfied in this case than when the initial sample is obtained under an SRS scheme. According to Cochran (1977), post-stratified sampling is almost as precise as stratified sampling with proportional allocation of sample units to strata, provided that the post-stratified sample is reasonably large in each stratum (nh.pc > 20), and the effects of possible errors in the weights Wh.pc can be ignored. As the increase in variance above that of stratified sampling will be small if the average post-strata sample size is sufficiently large, the application of Eqs. 41 and 42 without further adjustments for post-stratification can be justified.

Advantages and Disadvantages of Stratified Sampling Among the advantages of stratified sampling is the fact that estimates for subpopulation means or totals and their sampling variances are readily available. As the sample selection procedures for separate strata must be independent (otherwise there may be a covariance between results from different strata), sampling designs and sample sizes can be chosen freely to fit separate strata. In that regard, the stratified sampling design is quite flexible. In almost all cases, a gain in precision of population estimates of means, totals, ratios, and proportions is possible from either a prior or a post-stratification. While stratification for the estimation of a single population attribute is generally advantageous, the situation is less clear for a multipurpose inventory with its many attributes of interest. For each variable of interest, a different optimum stratifications are more likely to emerge than not. Consequently, the final stratification is a compromise dictated by thresholds and limits of precision not to be sacrificed in any attribute or subpopulation. The utility of stratified sampling should be particularly critically examined when the inventory is to be repeated in whole or part in later years. Changes in the design

Sampling in Forest Inventories

803

can be a result of, for example, the change of sampling units from one stratum to another (e.g., if the stratum is related to tree age or stand volume), new stratification rules are to be applied (e.g., due to the progress of remote sensing techniques), new strata must be defined to accommodate a possible strata  design effect in the estimates, or the inventory design is modified. A resultant nontrivial increase in the number of strata to deal with should be a concern as it complicates estimation and increases the likelihood of very small sample sizes in more than a few strata. Furthermore, if the population attributes of interest have changed between two inventories, the stratification used in the older inventory may no longer be efficient. An example of such a shift in focus is the current emphasis toward nonproductive functions of forests at the expense of a more narrow focus on timber values. In these situations, a post-stratification of the older inventory to the new classes may be an option. Fortunately, most problems that arise from shifts in strata and attributes of interest can be mitigated effectively by a proportional allocation of sample sizes to strata and a systematic sampling within strata. Inventory designs with these characteristics may be suboptimal in terms of efficiency, but they offer advantages – not to be underestimated – of flexibility and permanence.

Two-Phase Sampling Two-phase sampling or double sampling is a sampling procedure where two samples are taken from the population. The idea is to exploit an association between the attribute values in the two samples. In a first sample, a large number of easy-to-assess or low-cost sampling units are taken in order to measure one or more auxiliary variables. From this sample, a second smaller sample is taken for the purpose of assessing the attribute/variable of interest. The statistical link between the auxiliary variable(s) and the variable(s) of interest can be established either by a linear regression (two-phase sampling with regression estimators) or by using the auxiliary variable to estimate the size of strata (two-phase sampling for stratification). The two-phase design extends naturally to three and more phases (Ko¨hl and Kushwaha 1994; Magnussen 2003). Two-phase sampling can also be applied in the time domain, when from the n units measured at time 1, a subset is selected and measured again at time 2 in order to estimate change.

Two-Phase Sampling with Regression Estimators Estimation Procedures In two-phase sampling with regression estimators, a large sample of size n1 is selected in the first phase. On the selected phase 1 sampling units, the auxiliary variable x is assessed. In the second phase, a random subsample of size n2 n1 is selected where the auxiliary variable, x, and the variable of interest, y, are

M. Ko¨hl and S. Magnussen

804

measured. The two-phase sample with  regression estimator of the population mean, ^ ^ Y , and its variance, var Y , are 2prgr

2prgr

 ^ ^ ^ ^ ^ Y 2prgr ¼ Y 2 þ b X 1  X 2   v^ar ðyÞ 1  r^2xy v^ar ðyÞ r^2xy v^ar ðyÞ ^ þ  ffi v^ar Y 2prgr n2 N n1

(45)

(46)

^ and X ^ are the sample-based estimates of the mean of x in the first-phase where X 1 2 ^ is the estimate of the population and in the second-phase sample, respectively, Y 2

mean of y obtained from the second-phase sample, and b^ is the least-squares regression coefficient of y on x computed from the second-phase sample. The second term in the two-phase sampling with the regression estimator of the population mean (Eq. 45) is a term that corrects the SRS estimate by an amount that is proportional to the difference between the first- and second-phase estimates of the population mean of the auxiliary variable and the average effect of a one unit change in the auxiliary variable on the expected value of the variable of interest. ρ^xy in Eq. 46 is the phase-two sample-based estimate of the product moment correlation coefficient between phase-two sample values of yi and xi. Note, if the regression model is selected based on the data mining of the second-phase sample, the estimates of error may underestimate the actual (unknown) errors. Advantages and Disadvantages of Two-Phase Sampling with Regression Estimators Two-phase sampling with regression is typically used in forest inventories where remote sensing data and field assessments are to be combined. Although two-phase sampling with regression estimators is often more efficient than two-phase sampling for stratification (link?), it bears specific problems when used in practical settings. The cost relationship between the assessment in the first and second phase is one factor to consider carefully. The other is the strength of the relationship between the variable of interest ( y) and the auxiliary variable (x). This strength is often measured in the fraction of the variance in y that is explained by x. The coefficient of determination (ρ2xy) quantifies the fraction of explained variance; ρ2xy is a real number between 0 and 1. A higher ρ2xy means a stronger relationship and conversely a lower variance of the two-phase sampling with the regression estimate of the mean. In large areas or in forests with a large spatial variability, ρ2xy values around 0.4 are not uncommon. Thus only 40 % of the variation in y can be explained through variation in x. In homogeneous or small-scale forest areas, higher ρ2xy values are commonplace. To expect ρ2xy values larger than 0.9, however, is unrealistic given the natural variation of the variables and the lack of perfect relationships among common inventory attributes. Furthermore, such high values are probably questionable and could be the result of transformations of x, y, or both or the result

Sampling in Forest Inventories

805

of forcing the regression through the origin when an intercept term is significant. The interpretation of an estimate of ρ2xy should always be prudent. A few nontypical observations in the secondary sample or simply a nonrepresentative sample can grossly inflate the sample-based estimates of the population value of the correlation (Royall 2001). A cursory glance at the approximation to the variance in Eq. 46 may give the misguided (yet common) perception that any correlation can be exploited advantageously in two-phase sampling with regression estimators. An implicit requirement for the application of regression analysis is that the assessment of the variable of interest and the auxiliary variable is done on the same unit. This can only be safeguarded if the sample plots in the two phases coincide exactly. Location errors of terrestrial plots can be up to 20 m when, for example, maps on a scale of 1:10.000 are used for navigation to a location. Smaller errors in the order of a few meters can be achieved by the use of GPS sensors (Halme and Tomppo 2001; Keller 2001; Valbuena et al. 2010; Tomppo et al. 2011). The fact that a relationship between two attributes/variables often change across locations (Gertner 1984; Walters et al. 1991) further strains the notion of a single linear relationship across the entire sample. Finally, measurement errors in an auxiliary variable attenuates (bias) its regression coefficient (Fiuller 2006). Recently two-phase sampling with regression estimators has gained interest in the context of LIDAR remote sensing, where tree height measurements from laser scanning are used as auxiliary variables for growing stock, biomass, and carbon assessments (Palace et al. 2015; Saarela et al. 2015).

Two-Phase Sampling for Stratification Two-phase sampling for stratification is similar to stratified sampling except for the fact that strata sizes are not known before sampling, but estimated from the first-phase sample. Strata sizes are estimated by counting the number of first-phase samples located in each stratum and scaling the counts according to the sampling design. The variables/attributes of interest are assessed on the basis of the phase-two sample. Phase two can be a subsample of the sample taken in phase one (i.e., the sample outcome is dependent on the units selected in the first phase) or an independent sample. We shall limit ourselves to the option with a dependent subsampling in phase two and simple random sampling in each phase. The stratum of each ground plot is known at the time of (phase-two) sampling. Population parameters are then estimated through a combination of estimates obtained for each of the strata. The uncertainty surrounding estimates of strata size must be included in estimators of sample variance (link). Estimation Procedures In two-phase sampling for stratification with L strata and sample sizes n1 in phase one and n2 in phase two, the estimator for the population mean, for example, is as follows (Cochran 1977): Y^ 2pstr ¼

L X n1l l¼1

n1

^ ¼ Y l

L X l¼1

^ ^l Y W l

(47)

M. Ko¨hl and S. Magnussen

806

^ is the secondwhere nl is the phase-one sample size in stratum l (l = 1, . . ., L ) and Y l phase estimate of the lth strata mean. We see that the population mean is estimated as a weighted sum of strata means with weights (Wˆl) equal to the phase-one estimates of the relative frequencies of units (elements) in each stratum. A sample-based estimator of the variance of the estimated mean is L L  N 1X  2 W l v^ar ðy2l Þ ðN  n1 Þ X ^ ^ v^ar Y þ W l Y 2l  Y 2pstr ¼ 2pstr N l¼1 n2l N ðn1  1Þ l¼1

(48)

where v^ar ðy2l Þ is a phase-two sample-based estimate of the variance of variable y in the lth stratum. As expected, the variance is the weighted sum of within- and among-strata variances corrected for sample fractions and a finite population size. The variance estimator for stratified simple random sampling can often be used with impunity in lieu of the estimator in Eq. 48 whenever the phase-one sample has a very large number of units in each stratum (say over 1,000). A qualifying example is with a large remotely sensed image with, say, over 100,000 units (N ) and a phaseone sample of, say, 10,000 distributed across, say, five strata with a minimum of, say, 1,200 sample units per stratum. Intuitively, with a large sample size in the first phase, the variances of estimates of relative strata sizes become  small and ignorable  ^ ^ . leading to vanishing (unimportant) differences between v^ar Y and v^ar Y 2pstr

str

Advantages and Disadvantages of Double Sampling for Stratification The advantage of double sampling for stratification over stratified sampling is that a potentially laborious assessment of strata sizes can be replaced by a quicker and less costly sampling procedure. Strata may be defined exclusively for the purpose of estimation and they may not otherwise form any meaningful subdivision of the population. The within-strata variance, however, must be smaller than the variance in a non-stratified population before there can be a payoff from the phase-one stratification in the form of a lower sampling variance. In comparison to double sampling with regression estimators, the derivation of regression functions is no longer relevant. In many practical situations, there is no suitable auxiliary variable that is uniformly and strongly correlated with the variable of interest across the entire population. Two-phase sampling for stratification could be an attractive option. This argument extends to the case where field samples cannot be located with any great precision or linked with a desired accuracy to observations made in phase one.

Design-Based and Model-Dependent Inference The estimators for means (totals) and sampling error (standard error) given in previous sections have been design based (DB). That is, the sampling design and sample inclusion probabilities are reflected in the estimators. Most agencies involved with surveys and inventories of natural resources subscribe to a

Sampling in Forest Inventories

807

design-based framework for estimation and inference. The most important reasons for this adherence is that probability sampling eliminates personal biases in selecting the sample and that the estimation method used is design unbiased (or nearly so) for any population, irrespective of its structure. You are not required to develop a model for the population you are surveying. If you choose the design-based approach illustrated in this book, there are two rules that must be obeyed: (i) any change to the original sampling design, for reasons of economy or time, must be incorporated into a set of revised designbased estimators; (ii) the consequences of a compromise of the original sample selection procedure must be resolved and an appropriate remedial action must be worked out. Changes to a sampling design and events compromising the sample selection are not uncommon in large area surveys and forest inventories. The most important action to take when this happens is a detailed reporting with a complete picture of all changes and compromises to the sample selection procedure. Only this way is there a chance that the consequences can be made clear. Even so, a design and selection procedure can be so compromised that a designbased approach to inference is no longer defensible. A typical example surrounds the problems encountered when inference is attempted for domains and small areas that were not identified and incorporated in the original sample design. When a design-based approach to inference is deemed infeasible, an analyst may instead rely on a postulated model for the attribute values of interest, including, possibly, auxiliary variables deemed to be good predictors of the target attribute. The ensuing inference will be conditional on the sample at hand and it will also be model dependent because the properties of the estimates depend not only on the taken sample but also on a correct model specification. Model-dependent estimators can also be model unbiased if given the model and the sample, the expectations of the target parameters taken over all possible designs giving rise to the obtained sample are equal to the true parameter. This implies that the onus is now on the analyst to defend the model (“all models are wrong, a few are useful”) and the properties of the sample vis-à-vis the population (is the distribution of sample values the same as in the population?). Reassuring answers to these questions are, in effect, much more difficult to establish than generally believed. You can frequently come upon applications with the model-dependent approach to inference in forest resource assessments supported by remotely sensed data. Often little or no attention is given to the fact that the reliability of estimates of uncertainty hinges on a correctly specified model. We realize that in practice a model-dependent approach to inference can be the only option available to an analyst. In that case, we strongly recommend An Introduction to Model-Based Survey Sampling with Applications by Chambers and Clark (2012) where one finds a wide collection of the most common and popular model-dependent estimators, many with a parallel to the design-based estimators presented in this book. For further information on model-dependent inference, see Thompson (1992), Gregoire (1998), Rao (2003), Little (2004), and Magnussen (2015).

808

M. Ko¨hl and S. Magnussen

Selection of Trees on Sampling Units The sampling units of forest inventories are usually not individual trees but groups of individuals satisfying some criterion. Sample trees may be selected by, for example, satisfying the criterion of location inside a sample plot (see section “Scaling of Individual Tree Data into Sample Plot Values”) exceeding a distance weighted size threshold in point sampling (see section “Selection of the Optimal Plot Design”) proximity to a survey line or satisfying a rank proximity criterion at a sample location as in, for example, fixed-count sampling (Prodan 1968; Pommerening and Schmidt 1998; Thomas et al. 2010). The rank proximity criterion includes a fixed number of trees closest to a sample location. For example, the six trees closest to a random sample location may be selected (Prodan 1965). Estimators based on this type of tree selection will often be biased, especially if the spatial distribution of trees is aggregated or in some other way displays distinct spatial patterns. Further, in dense forests, it can be difficult to determine the ranking of tree distances from a random point. In sampling on successive occasion, logged trees or ingrowth can alter the original set of trees at a sampling location and make change estimates difficult. Those who like to know more about it are referred to the literature (Prodan 1965; Pollard 1971; Payandeh and Ek 1986).

Fixed-Area Sampling Units Fixed-area sampling units are the simplest intuitive basis for selecting trees to be assessed in forest inventories. The term “plot” is applied to small circular, rectangular, square, or triangular areas. A “strip” is a rectangular sample area, whose length is a multiple of its width. Unbiased estimates can be computed for all sample areas, no matter what their shape. In planning an inventory, survey costs must be weighed against desired precision to determine the optimum size and shape of the sample plots. The shape of the sample plot is mainly determined by the costs and other practical considerations. In temperate latitudes, circular plots are usually employed as they have the smallest periphery in relation to area and consequently the lowest number of borderline trees for a given area of the plot. In tropical forests, where the undergrowth hinders both access and visibility and large areas must be surveyed, it is usual to take rectangles or squares because such plots are the easier to establish. Very often, strips of up to 30 m wide and several hundred meters long are recommended. Sample plots in areas with a high stem count may contain a large number of trees. As small plots are often more efficient in areas with a high tree density and larger plots being more efficient in areas with a low tree density, plot designs with multiple concentric plots have been introduced. Two or more plots of differing size are demarcated around a given sample point. In the smallest area, all trees with diameter greater than a given minimum fixed by design (e.g., 12 cm) are surveyed.

Sampling in Forest Inventories

809

Fig. 7 Field plot of the Swiss NFI (Zingg 1988)

16

36

30

18 14 9 24

28

R5 = 12,62

PFZ 14

10

R=

7,9

8

16 40

46

18

10

32

9 14

14

In the larger plots, the minimum diameter threshold is higher. This design often allows a considerable reduction of survey time without a noticeable decrease in efficiency compared to a single plot size applicable to all trees in the plot. Figure 7 shows the concentric plot design employed in the Swiss National Forest Inventory. On the smaller, 200 m2, plot, all trees with a DBH over 12 cm are measured, while on the larger, 500 m2, plot, only those with a DBH of 35 cm or above are measured. For sampling on successive occasions, permanent fixed-area plots are to be particularly recommended, as they allow easy determination of growth components, such as survivor growth, ingrowth, mortality, and cuts (Scott 1998; Pretzsch 2009; Weiskittel et al. 2011). Also, fixed-area plots are usually simple to survey, maintain, and analyze. For these reasons, they are preferred over variable radius plots (Rice et al. 2014).

Scaling of Individual Tree Data into Sample Plot Values The statistical approach to sampling designs generally assumes that sample plots represent the smallest (natural) sample unit. It is common to take an area of one hectare as the natural unit. However, actual sampling may not be done with this unit, as generally applied plot sizes are considerably smaller. Indeed plots of different size may be used or trees may be selected based on a criterion of inclusion. Thus, individual tree values sampled during a survey have to be scaled to the natural unit. The scaling is accomplished by area weighting of the attribute, say Y, of the jth tree on the ith sample location. Let aij denote the area of the sample plot used to sample the ith tree at the jth sample location. The area weight given to the attribute value Yij is wij ¼ a1 ij which scales the attribute value to one hectare. A synonym for

M. Ko¨hl and S. Magnussen

810 Table 2 Plot expansion factor Plot size [m2] 200

Weight wij

500

1 ha 1ha wij ¼ 500m 2 ¼ 0, 05ha ¼ 20 1 ha wij ¼ 1000m2 ¼ 0,1ha 1ha ¼ 10

1,000

Attribute value per hectare (Yij = 3 m3) Yijha = 3 m3  50 = 150 m3

1 ha 1ha wij ¼ 200m 2 ¼ 0, 02ha ¼ 50

Yijha = 3 m3  20 = 60 m3 Yijha = 3 m3  10 = 30 m3

the weight wij is the term “plot expansion factor.” Table 2 gives examples for plot expansion factors. This area weighting is flexible as it extends naturally to sampling with unequal probability of inclusion for individual trees. A simple example ought to clarify the concept. Let us assume that we are sampling with a set of fixed-area concentric sample plots. The effect of different plot sizes on selection probabilities has to be acknowledged in the estimation phase through a scaling to a common (natural) unit. This is the purpose of area weights. For each concentric plot, a separate scaling factor applies. If, for example, two concentric plots of size 0.02 ha and 0.05 ha are used, the scaling factors are calculated as 1 w ¼ 0:05 ¼ 20 for trees measured on the 0.05 ha sample plot and 1 w ¼ 0:02 ¼ 50 for trees measured on the 0.02 ha sample plot. In the above example, it was the size of the trees that determined whether they were measured on one plot or the other. Often the diameter at breast height (DBH) is used as the size criterion due to ease of measurement. In that case, the area weights (plot expansion factors) become functions of DBH. If all trees irrespective of their DBH are tallied in the smaller 0.02 ha plot and in the 0.05 ha plot, only trees with DBH  35 cm are tallied, we can express the weights as

wij ¼

50, if DBH ij < 35cm 20, if DBH ≧ 35 cm

(49)

Note, the scaling of inventory estimates to a unit area (here 1 ha) allows us to assess the effect of plot size, plot shape, and selection criterion on statistical estimates of interest. Recall, we do not consider a scaled attribute value as a ratio of two random variables since we assume throughout that the scale factor is known without error. Measurement errors in aij are not considered. After scaling individual attribute values Yij to a unit area (1 ha), we usually sum them to a single value Y iþ for the ith sample location Y iþ ¼

X j

wij Y ij

(50)

Selection of the Optimal Plot Design The selection of the optimum size is basically determined by the spatial distribution and the variability of the forest to be surveyed and by the cost associated with

Sampling in Forest Inventories

811

reaching a plot location. A fixed total sample area can be divided into a smaller number of plots with a larger plot size or a larger number of plots with a smaller plot size. With the constraint of fixed sample area, choosing a larger plot size means that the sample size goes down relative to a design with smaller plots. A large plot is likely to produce a lower among-plot variance than a smaller plot since large plots in general display more within-plot variation than do smaller plots. Yet the lower number of larger plots afforded under a fixed total sample area may actually produce a higher standard error than sampling with a smaller plot (Correll and Cellier 1987; Magnussen 1988; Gray 2003a). The optimum plot size in terms of minimum sampling variance for a fixed total sample area is determined by the spatial distribution and the variability of the forest to be surveyed. Small plots in homogeneous forests may furnish results with higher precision, as the number of independent observations for a given sampling intensity is higher. On the other hand, in heterogeneous forests, the coefficient of variation between small plots may increase so greatly that it would be better to use a larger plot. Consequently, not only the costs but also the variability of the inventory area must be taken into consideration. A key statistic to gauge the efficiency of different plot sizes is the intra-plot correlation coefficient (Cochran 1977) (see Chap. 2.3). The coefficient measures the similarity of observations within the sample plot. Basically, the plot size should be a decreasing function of the similarity of observations within a plot (Correll and Cellier 1987; Saborowski and Smelko 1998). Zeide (1980), as well as Mesavage and Grosenbaugh (1956), Tardif (1965), and O’Regan and Arvanitis (1966), examined various methods for optimizing plot design. Zeide (1980) weighed the time needed to locate a plot against the specified precision and found that the optimum plot design comes with the lowest expenditure for a specified precision. The optimum plot size a can be computed with a ¼ a1

 t 2 m

(51)

where a1 = size of plot used in a pilot survey, t = average travel time between two neighboring plots, and m = average measuring time on a plot of size a1. Zeide (1980) concluded that the greater the distance between plots, the larger the plots should be. Lombardi et al. (2015) provide examples for deriving the optimal plot size for the assessment of distinctive structural features. To compare two plot designs (say type 1 and 2), their relative efficiency for the estimation of, say, a population total is Eff type1:type2 ¼

  ^ Yjplot ^ Var type 1 ⋆cost of inventory with plot type 1   ^ Yjplot ^ Var type 2 ⋆cost of inventory with plot type 2

(52)

An efficiency ratio less than one means that plot type 1 is more efficient and vice versa. Note that the efficiency depends on the population parameter of interest. It is possible that one plot type is more efficient for one parameter but less efficient for another.

812

M. Ko¨hl and S. Magnussen

As a rule of thumb, small plots are preferable for dense stands with small trees; larger plots are better suited for open stands and large trees. A plot should be large enough to contain enough trees per plot for the survey results to be representative of the population at large while at the same time keeping the time to complete a plot to a minimum. Very often, a distinction is made between unproductive relocation time between plots and productive survey time. When travel time is significant, as often the case in a tropical forest inventory, the size of inventory plots tend to be relatively large; one can often encounter plots in the 0.4–0.5 ha range. The life span of an inventory plot is also important to consider when deciding on a plot size. Permanent fixed-area sample plots intended for multiple surveys are difficult to optimize. The number of trees and their size will naturally change over time. To ensure that the plot size is sufficient throughout the life of a plot, a permanent fixed-area plot tends to be relatively large.

Assessments in Sloping Terrain As the horizontal plane is the reference base for all inventory data, sample plots in sloping terrain must be adapted accordingly. A plot is demarcated on the sloping ground, but a horizontal projection is required. The horizontal projection of a circular plot on a sloping plane is elliptical. A rectangle or a square on a sloping ground becomes a parallelogram when projected to the horizontal plane. Note that horizontal projections of your plot on a sloping ground will have a smaller area than the plot demarcated on the inclined plane. Where plot sizes are not calibrated to account for the slope, a negative bias will manifest itself in all attributes expression on a per unit area. Three distinct calibration procedures given next; the first is general, while the second and third are for circular plots only. All plots: Demarcate the plot on the incline and then expand the plot proportionally to the degree of inclination in such a way that an orthogonal projection of the expanded plot onto the horizontal plane matches exactly the nominal plot in area. Circular plots: Demarcate an ellipse on a slope with a short and long axis determined in such a way that when the ellipse is projected orthogonal onto the horizontal plane, it coincided with the circular outline of the nominal inventory plot. Circular plots: Measure or compute the horizontal distance of candidate plot trees to the plot center. Only trees within a distance of the nominal plot radius are included and measured. In order to simplify the field survey and to facilitate an audit and a quality control of the inventory, a correction of the plot area according to the degree of inclination appears most suitable. Ko¨hl and Brassel (2001) showed for the Swiss NFI that there is no significant difference between the second and third expansion approach. In mountainous regions, an error in the corrected plot can induce a nontrivial error in the survey results. For this reason, corrections for slope inclination must be performed with great care, and all pertinent data about the expansion/ correction process should be captured to allow a control and possibly a correction of the errors.

Sampling in Forest Inventories

813

Point Sampling Compared with permanent sample plots, point sampling is a relatively new procedure. First presented by Bitterlich in 1947 (Bitterlich 1947), it was developed and theoretically substantiated by Keen (1950) and Grosenbaugh (1952). Alternative names for this method are “angle count,” “variable plot cruising,” and “plotless cruising.” All these names reflect one of the most important features of the method: the area surveyed varies (Hush et al. 2003; Ducey 2009). The first step in point sampling is the same as that in surveying with fixed-area plots: the selection of sample locations (points). Then attributes of interest are measured on all trees meeting a certain criterion of selection. Typically the criterion is DBH and the decision on whether to include or exclude the tree from selection is based on a measurement with an angle gauge instrument. The simplest form of an angle gauge has a crossarm attached horizontally to a vertical stick held at a known distance from the eye (Fig. 8). With this instrument, an angle in the horizontal plane and 1.3 m above ground is aligned with the trunk of each tree visible from the sample location. All trees with a DBH forming an angle greater than the angle subtended by the crossbar are selected. Trees with a smaller DBH are ignored (Fig. 9). Assessments and measurements are then carried out on the selected trees. The basal area is determined through multiplication of the number of trees by a constant factor derived from the given angle; no extra measurements are needed. Thus each tree assessed, independent of its diameter, represents the same basal area per hectare. The basal area per hectare at the sample location is determined through multiplication of the number of “in” trees by a constant factor derived from the given angle subtended by the horizontal cross bar; no extra measurements are needed. Thus each tree assessed, independent of its diameter, represents the same basal area per hectare; a proof is given below. The geometrical principle of point sampling is illustrated in Fig. 10 with a crossarm 2 cm wide (b = 2 cm) attached to a stick 1 m long (l = 1 m). For a tree at distance Ri with a diameter at breast height equal to DBHi and subtending an angle equal to the angle subtended by the angle gauge crossarm, we get b 2 cm DBH i ¼ ¼ , 50  DBH i l 100 cm Ri

Fig. 8 Stick-type angle gauge

(53)

M. Ko¨hl and S. Magnussen

814 Fig. 9 Point sampling

Fig. 10 Geometrical principle of point sampling

b DBH

Ri

Any sample location within a distance of 50*DBHi from this tree would result in the tree being included in the sample. In other words, the sample area for this tree is π*(50*DBHi)2. The attribute value of the tree is therefore given an area weight equal to the inverse of this area. If the attribute value is the basal area, i.e., π4 DBH 2i , the area weighted attribute value is simply π π DBH 2i DBH 2i πDBH 2i 1m2 m2 4 4 ¼ ¼ ¼ ¼ 1 10000m2 ha πR2i π 10000 DBH 2i π ð50di Þ2

(54)

This means that, in this example, every selected tree represents a basal area of 1 m2/ha. The basal area per hectare is estimated by simply counting the number of tallied trees. The simple derivation shown above is only valid for b/l = 2/100. If a gauge with a different subtended angle (α) is used, a more general equation must be employed (see Fig. 10 for details and definitions): Ri ¼

di 2sin

α 2

(55)

Sampling in Forest Inventories

815

For any given gauge angle α and count Ncount of “in” trees, the basal area (G) per hectare is estimated as ^ ¼ N count 104 sin2 α ¼ N count BAF G 2

(56)

where BAF stands for “basal area factor.” The BAF value is indicated on commercially available angle gauges. After a 360 sweep and deciding on “in”/“out” for every visible tree, one obtains the basal area per hectare by multiplying the number of “in” trees (Ncount) by the BAF. The chosen angle and thus the factor determine the number of selected trees. A wider angle translates to a higher factor and consequently to a lower number of selected trees. In tropical forests, factors between 8 and 10 are popular; they ensure reasonable counts (between 10 and 40). With angle point sampling, the measured attribute (say population mean) of the trees counted as “in” should be expanded to a common reference area of one hectare in order to remove the effect of differential inclusion probabilities. The expansion takes the form zj zj n X n X ^ yij BAF X yij Y BAF X ¼ ¼ π ha n i¼1 j¼1 gij n i¼1 j¼1 DBH ij 4

(57)

where gij is the basal area of the jth “in” tree at the ith sample location (i = 1,. . .n). Consequently, the basal area of all “in” trees must be determined or, conversely, estimated from a measurement of DBH and the assumption of a circular outline of the stem cross section. Per hectare estimates of stems and basal area deserve special attention. For stem count, the attribute value of each “in” tree is 1, so the estimator for the stem count per hectare becomes countðjÞ n NX N^ BAF X 1 ¼ ha n i¼1 j¼1 gij

(58)

and the estimator for the basal area per hectare is n NX n count ðiÞ ^ BAF X gij BAF X G ¼ ¼ N count ðiÞ n i¼1 j¼1 gij n j¼1 ha

(59)

As illustrated above, point sampling with an angle gauge is essentially sampling with probability proportional to size (basal area) (Fig. 11). In fixed-area sampling, all trees have the same probability of selection, a probability that only depends on plot size. For any attribute related to size, a selection with the probability of selection proportional to size will result in a more efficient sampling (Brewer and Hanif 1983; Sa¨rndal et al. 2003). The estimated standard error for a given number of selected trees will be lower than for sampling with equal selection probability.

M. Ko¨hl and S. Magnussen

816 Fig. 11 Selection of trees in point sampling

It can happen that the angle subtended by a tree’s DBH appears to be exactly equal to the gauge angle. Such trees are termed “borderline trees”; their diameter and distance from the point center must be measured and the decision as to whether to include them or not based on the equation DBH b α ≧ ¼ 2 sin Ri l 2

(60)

There are many variations on horizontal point sampling as described above, e.g., vertical point sampling, vertical line sampling by Strand (Strand 1958), the critical height method for volume assessment, or angle counting by Wenk (see Loetsch et al. 1973; Hush et al. 2003). Uebelho¨r (1988) describes the use of point sampling with the wide-scale relascope in the Philippines National Forest Inventory and recommends point sampling for measurements in tropical rain forests to reduce the cost of field surveys. Other applications of point sampling in tropical forests are presented by Boon (1970), Puffenberger (1976), Da Silva (1982), and Banyard (1987).

Instruments for Point Sampling Commercially available instruments for point sampling are based on one of two different principles. One uses the above-outlined principle of two divergent lines starting at the viewpoint and extending to a fixed reference distance and beyond until intercepted by an obstacle (Fig. 8). The practical problem arising with this type of instrument is that a close object (reference distance) and a distant object (the tree) have to be focused and two lines (right and left side of the tree) observed

Sampling in Forest Inventories

817

Fig. 12 Wedge prisms for point sampling

simultaneously, often by a human observer. This renders decisions about whether or not to include borderline trees difficult. Angle gauges sold under the name of relascope include a feature for automatic correction for inclinations from the horizontal. A wide-scale relascope was developed for application in tropical forests (Loetsch et al. 1973). The second principle is based on the diffraction of light rays, in our case diffraction of light rays from trees as they go through a wedge prism in front of the observer (Fig. 12). The observer will see two superimposed images of the tree stem. An actual non-diffracted image is superimposed on a diffracted image. The tree is selected when its actual image coincides with the diffracted image. Trees with a diffracted image displaced laterally relative to the actual image are not selected. It is much easier to decide about borderline trees with this type of instrument than with a relascope or stick-type angle gauge. Ease of use made them popular, especially in Canada and the USA.

Point Sampling Versus Fixed-Area Plots Forest resource sampling is a challenge due to the frequent encountered problem of uniquely defining population elements/units and consequently the problem of defining a sampling frame. The point paradigm, by which a population is defined as all possible spatial locations of a sample unit, is adopted out of necessity. When there is no natural sampling unit, the survey designer has to decide on how observations are gathered at each sample location. The decision as to whether to employ point sampling or sampling with fixed-area plots depends on the individual aims and needs of the inventory. In a study on an area of 60 ha in Suriname, Schreuder et al. (1987) compared the efficiency of fixed-area plots, point sampling, and horizontal line sampling. Fixed-area plots gave the best results in terms of root mean square error for tree number, horizontal line sampling for basal area, the sum

818

M. Ko¨hl and S. Magnussen

of tree diameters, and the average tree diameter. Point sampling was superior for estimating the number of trees in the upper diameter classes while fixed-area plots fared better for the smaller diameter classes. These findings apply in general. With point sampling in stands with a high stem count, with clusters of big and small trees, or with dense undergrowth, there is a nontrivial risk that trees may be hidden and consequently that a negative bias may be incurred. The time to implement repeated checks for hidden trees and their inclusion in the local sample quickly erodes any practical advantage of point sampling. In such cases, it is preferable to use fixed-area plots. In the consideration of fixed-area plots versus point sampling, the survey analyst must also take into consideration the life span of a sample location. Will the sample locations be used in future inventories or will they be part of an ongoing monitoring? If plots are to be used again and again over time for the purpose of estimating change and trends in population parameters, the fixed-area plot has some distinct advantages. In point sampling, the inclusion probability of a tree depends on the attribute value of the tree (commonly BA) at the time of sampling. Thus, the inclusion probability does not remain constant over time for attributes/variables that change over time. In the context of estimating the population parameter “tree growth,” the change in inclusion probabilities generates some unique estimation problems. At the time of remeasurement, you can have trees included in the point sample that were not included at the previous time of sampling. The estimation of growth of individual trees becomes cumbersome when their inclusion probability has changed between times of measurement (Flewelling and Thomas 1984; Scott 1998). There are two distinct events that would allow a tree to enter the later measurement but not the earlier. First, the DBH of the tree did exceed the minimum threshold diameter on the first occasion but it was located beyond the critical inclusion radius. The growth estimate for this tree would equal its size. The common terminology for this type of growth is “ongrowth.” The second event that allows a tree to enter the sample between sampling times is when its size exceeds the inclusion threshold at the second but not the first measurement occasion. The growth calculated for this tree is called “ingrowth.” Kuusela (1979) describes various estimators of growth based on point sampling with repeated measurements at a fixed set of sample locations. The complex nature of these estimators suggests that they should only apply in exceptional cases. Procedures for estimation of increment and growth components in fixed-area plot sampling are simple in comparison.

Sampling at the Forest Edge Sample locations at the forest edge present a special estimation problem in forest inventories. Since the population of interest is restricted to areas classified as forest, it can happen that the effective sample area for these locations is less than the nominal area associated with sample locations in the interior of the forest. It would

Sampling in Forest Inventories

819

be wrong to simply disregard such sample locations as this would mean that trees growing in areas along the forest edge would have a different probability of being selected compared to trees growing further away from this edge ((Williams 1996; Gregoire and Scott 2003). Since growing conditions and tree species along the forest edge are often distinctly different from the interior forest, the disregard of edge plots could lead to a serious bias in the inventory estimates. The surveyor has several options for correctly dealing with the boundary effects. The problem and solutions are best understood if we adopt a tree-centric view of sample areas. For the fixed-area plot, the sample area of a tree is simply the area covered by the sample plot when the centers of the plot and the tree location coincide. For point sampling, the sample area is the area covered by the circle centered at the tree location and with a radius equal to the critical distance of selection. Trees located along the edge of the forest will have part of their sample area outside the population of interest. They are therefore less likely to be selected than a tree further away from the boundary. The solutions presented next are for sampling with fixed-area plots, but they apply equally to point sampling by simply replacing the word “sample plot” by “sample area.” One recommended option involves finding the exact intersection of the forest edge with the inventory plot and then compute the area of the plot that is inside the population. The attribute values observed on this partial plot are scaled according to the area of the plot inside the population. The weighting scheme can also be applied to individual plot trees. For trees with a distance to the forest edge less than the radius of the appropriate sample area (e.g., 12.62 m for a circular plot of 0.05 ha), the part of the sample area inside the forest is determined for each tree and converted to an area weight wij (Fig. 13).

Forest

Non-Forest

Area inside forest

Expansion factor

a i = 0.05ha w i = 1ha/0.05h =20

Fig. 13 Sampling at the forest edge

a i =0.05ha/2 =0.025ha w i = 1ha/0.025ha = 40

a i = 0 ha a i = 1ha/0ha =0

820

M. Ko¨hl and S. Magnussen

Another rather ingenious solution is the “fold-back” or the mirror reflection method (Schmid 1969). In a mirror reflection of a plot, the part of the plot that is outside the forest is projected orthogonally back into the forest with the forest edge serving as the axis of projection. The surveyor records all attributes for the part of the plot that is fully inside the forest and then all attributes on the mirror reflection of the part that was outside. In other words, a part of the plot is measured twice and occasionally three and four times if a boundary is on a corner and reflected portions of the plot overlap. Correctly applied, this method produces unbiased results, but it assumes that the forest boundary can be located accurately. In practice it is often quite difficult to decide on the exact location of a boundary. Any nonlinear edge will also generate practical problems with the mirroring. When the forest edge cannot be defined in precise and generally valid terms, the method becomes problematic. Two easy-to-implement remedies, but resulting in biased estimates, consist of (i) relocating the straddling plot further away from the boundary to avoid any overlap with the outside non-forest area and (ii) expanding, during the estimation phase, the part of the plot inside the forest to compensate for the area outside the forest. Gregoire and Scott (1990) compared four unbiased and three biased methods for dealing with sample plots at the forest edge in a mixed hardwood and mixed softwood stand in Maine. They concluded that no single method was uniformly best. Their performances depend on the nature and magnitude of the “edge effect.” Some biased methods of plot relocation performed, at times, better than the unbiased methods. In practice, it is often the case that only plots with a plot center inside the forest are accepted and tallied. The existence of sample locations outside the forest boundary, however, raises questions about the integrity of the sample frame and the multipliers to use to when you scale per hectare estimates to population totals. If a sample location is judged outside the forest, but the outside location is actually a part of the forest estate (in an administrative or legal context), it can be argued that the “outside” location should have been included in the sample. In areas with illegal forest clearing, for example, the discard of “outside” sample locations could lead to a serious overestimation of per hectare attribute values.

Sampling on Successive Occasions Sampling on successive occasions has the following main objectives: – To determine the status of the forest resource at the time of the first inventory – To determine the status of the forest resource at the time of the second inventory – To determine changes in the forest resource between two successive inventories The basic principles behind a quantification of forest growth and yield by repeated measurements were born in the last century in Europe. In Germany, the first permanent plots were established in 1860 (Graves 1906). Foresters in France (Gurnaud 1878) developed a set of rules for how to estimate increment from

Sampling in Forest Inventories

821

successive measurements. Biolley (1921) was the first to apply the rules in the forest of Couvet, Switzerland, which was measured ten times between 1890 and 1946 in intervals of six to seven years. In the USA, the idea of obtaining quantitative estimates of the change in standing wood volumes through repeated measurements in a set of plots gained support and acceptance over a period stretching from 1929 to 1950 (Stott and Semmes 1962). Alder and Synnott (1992) give a detailed presentation of methods for permanent sampling in mixed tropical forests. Mixed tropical forests can only be managed sustainably if it is possible to regulate felling in a way that safeguards their maintenance as a productive ecosystem (Alder and Synnott 1992). Sampling on successive occasions provides information on tree increment, mortality, regeneration, and ingrowth. It is therefore – in many situations – a prerequisite for sustainable forest management and determination of allowable cut. Sampling on successive occasions can utilize both temporary and permanent plots. Temporary plots are visited only once and do not include the time dimension directly. Permanent plots are, in contrast, remeasured periodically. Hence, they also provide data on change over time for the growth components: tree increment, mortality, cut, and ingrowth (i.e., young trees reaching a minimum DBH threshold). On permanent plots, individual trees are identified either by visible marks or by recording their coordinates. There are several designs tailored to sampling on successive occasions. They mainly differ in how they deploy and utilize permanent and temporary plots. The designs fall into three broadly defined categories: (1) independent samples, (2) permanent plots, and (3) a mixture of both plot types.

Independent Samples In rare circumstances, plots may be selected independently at each occasion (Fig. 14). Because plots are visited only once, plot locations and tree locations within a plot do not have to be recorded. Change estimates are obtained by ^ ¼Y ^X ^ C 2, 1

(61)

^ is the estimated mean of plots assessed at the first occasion (time 1) and Y ^ is where X the estimated mean  of plots assessed at the second occasion (time 2). The variance ^ ^ of C2,1 , viz: v^ar C 2, 1 , is obtained by    ^ ^ ^ þ v^ar X ^ar Y v^ar C 2,1 ¼ v

(62)

  ^ and v^ar X ^ are the estimated variances of observations Y at time where v^ar Y i 1 and Xi at time 2.

822

M. Ko¨hl and S. Magnussen

Fig. 14 Independent samples at successive occasions

If only the current status of the resource is to be considered, temporary sample plots are often more cost effective than permanent plots, since – as mentioned earlier – there are no required expenditures for monumenting plot centers and registration of sample tree locations. If change has to be estimated, however, a standard error computed from a set permanent sample plots will be lower, since the average difference between two sets of independent observations is not due to change alone but also due to the contributions of two independent sampling errors.

Continuous Forest Inventory In the 1930s, a sampling method, known as continuous forest inventory (CFI), was developed in the USA. CFI is based on repeated measurements of a set of sample plots (Stott and Ryan 1939). Despite the fact that the term “continuous” in CFI may seem parochial, as the method is neither continuous in the spatial or in the time domain, we shall, however, continue to use the term CFI as it is widely used in texts on forest inventory. Stott and Semmes (1962) give a historic overview of early CFI application. With the CFI method, all sample plots, which were measured on the first occasion, are measured again in successive inventories (Fig. 15). The estimators of population parameters under CFI are time specific. We indicate the time dependency of CFI estimators by a suffix t for time. The suffix takes on values 1, 2, 3, . . . for the estimates of population parameters at the first, second, third, and so on inventory. We are usually interested in the estimation of a change between two successive inventories. To simplify notation when the context is clear, we shall simply refer to estimators and estimates in two successive inventories as “first” and “second.”

Sampling in Forest Inventories

823

Fig. 15 Continuous forest inventory (CFI)

In continuation of our example for the estimation of a population mean under ^ (Eq. 63) and 2nd, Y, ^ (Eq. 65) occasion, the CFI estimators are SRS at the first, X, Xn1 ^¼ 1 X X i¼1 i n1

(63)

Xn2 ^¼ 1 Y Y i¼1 i n2

(64)

Changes in a population parameter between two inventories can be derived as the difference between the second (Eq. 64) and first (Eq. 63) estimate of the population parameter. Specifically, ^¼Y ^X ^ C

(65)

The estimator of the variance of the change in the mean becomes rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi     ffirffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  ffi ^ ¼V Y ^ ^ þV X ^  2^ ^ v ar X r YX v^ar Y v^ar C

(66)

as the same set of plots is remeasured on both occasions. r^YX is an estimate of the correlation coefficient between the observations of the second and the first occasion. The correlation r^YX is restricted to values between 1 to +1. A higher positive correlation between paired observations from the first and second inventory leads to a smaller variance of their difference. Equation 66 explains the reasons for this important result. The advantage of using the CFI method rests with a reduction of the variance in an estimate of change. In extreme cases, the correlation (viz., covariance) can be negative, e.g., in heavily exploited tropical forests when highvolume stands are cut between the two occasions and young stands are growing fast. Should this occur, the variance of an estimate of change will be larger than a variance estimated from independent samples since the last term in Eq. 66 is now positive.

824

M. Ko¨hl and S. Magnussen

If the primary objective is to estimate net change with an error below a stated maximum tolerated error, a CFI with permanent sample plots can, despite the extra cost involved in monumenting plots, marking, and locating trees, still be more cost efficient than two independent surveys. This is possible because CFI, for a given sample size, tends to produce smaller estimates of standard error than surveys with independent samples. Hence CFI may reach the set target accuracy with fewer (permanent) plots with a lower cost than possible with larger independent samples. If, in contrast, an estimation of the current state is to be considered as the primary inventory objective, a design with temporary sample plots will most often be more cost effective than a design with permanent plots, since expenditures for marking the sample plot centers and the registration of sample tree locations can be dispensed and avoided. Ranneby and Rovainen (1995) present methods for the determination of time intervals between measurements of permanent plots. The CFI method, despite obvious advantages, encounters practical and inferential problems. Over time the locations of sample plots may become known beyond the surveyors with the risk that the plots may deliberately be treated differently than their surroundings. This is a nontrivial risk that is especially relevant when sample plots are visibly marked. With permanent plots, an omnipresent inferential problem is unavoidable. As paraphrased by Schmid-Haas (1983), “there is no guarantee that sample plots, visible or not, will remain representative of the target population.” Inventories that potentially do not represent reality will lose credibility. This risk is especially acute in managed forests and in places with frequent landuse changes. Ko¨hl et al. (2015) discuss the effect of treatment bias for CFI and SPR (see d). Also, changes in primary inventory objectives can be difficult to accommodate in a CFI with a legacy of plots established and tailored to meet past objectives.

Sampling with Partial Replacement of Plots (SPR) Practical survey objectives are often a blend of target precision on estimates of state and change. In this case, a design with a mixture of permanent and temporary plots appears attractive. The ideas behind survey designs with both plot types arose from such mixed objectives. Sampling with partial replacement (SPR) utilizes a mixture of both permanent and temporary plots. New, temporary plots are established at the second and every following occasion (Fig. 16). Temporary plots allow an analyst to assess the above-discussed potential of a possible “treatment bias” in the remeasured plots (Ko¨hl et al. 2015). In addition, plots lost due to land-use changes can be “replaced” by new plots so that the number of forested plots does not diminish over time. SPR was introduced to forest inventory around 1960 (Bickford et al. 1963; Ware and Cunia 1965). Scott (1984) presented a consistent set of estimators for SPR, which will be presented in the following.

Sampling in Forest Inventories

825

Fig. 16 Sampling with partial replacement (SPR) of plots

For two occasions, three types of sample plots can be considered: – Sample plots that are measured on the first occasion as well as on the second occasion (permanent, matched sample plots) are referred to as the n12 sample. – Sample plots that are only measured on the first occasion (unmatched plots) are referred to as the n1 sample. – Sample plots that are only measured on the second occasion (new, unmatched plots) are referred to as the n2 sample. The SPR estimation procedures involve four steps (Scott 1984): 1. The current state is obtained by (a) a mean based exclusively on the most current ^ (Eq. 67), and (b) a regression(i.e., second) measurements in the n plots, Y 2

A

^ , comprised of the sum of the time 2 mean based estimate of the current state, Y B of the n12 plots plus an adjustment for the difference between the time 1 mean of all plots assessed at time 1 and the time 1 mean of all permanent plots (Eq. 68). ^ is formed by updating the time 1 mean using the simple linear regression Y B

between time 1 and 2 on the remeasured plots. This regression, in effect, updates the values of the sample plots that are not remeasured (Y1). A second mean is derived from the new (temporary) sample plots (Eq. 68): Xn2 Y 2j j ^ ¼ ^ ¼Y Y (67) A 2 n2  ^ þ β^ X ^ ^ X ^ ¼Y Y (68) B 12 1 2 YX

M. Ko¨hl and S. Magnussen

826

where ^ ¼ time 1 mean of all plots assessed at time 1 X 1 ^ X12 ¼ time 1 mean of all permanent ðremeasuredÞ plots ^ ¼ time 2 mean of all permanent ðremeasuredÞ plots Y 12

^ ¼ mean of all time 2 temporary plots Y 2 Y2j = measurement of time 2 temporary plot j, j = 1,. . ., n2 β^YX ¼ slope coefficient in the simple linear regression of Y12 on X12 ¼ ss2XY

X12

sXY ¼

 Xn12  ^ ^ Y  Y  X X 12j 12j 12 12 j ðn12  1Þ s2X12 ¼

Xn12  ^ 2 X  X 12j 12 j ðn12  1Þ

(69)

(70)

X12j = measurement of permanent plot j at time 1, j = 1,. . ., n12 Y12j = measurement of permanent plot j at time 2, j = 1,. . ., n12 2. For both means, the variance is calculated: 2 Xn2  ^  s2 Y  Y 2j 2 j ^ ¼ Y 2 ¼ v^ar Y A n2 n2 ðn2  1Þ 2 3  2 ^ ^ 2 2  X1  X12 1 7 sY 12  sY:X ^ ¼ s2 6 þ þ v^ar Y 4 5  B Y:X n12 Xn12 n1 ^ 2 X12j  X 12 j

(71)

(72)

where Xn12  s2Y 12 ¼

s2Y:X ¼

ð1  r 2 Þ

Xn12  j

^ Y 12j  Y 12

ðn12  2Þ

j

Y 12j  Y^ 12

2

ðn12  1Þ

2 ¼ mean squared error of the regression

r ¼ sX sXYsY ¼ estimated correlation between observations at time 1 and 2 for the 12

12

n12 plots 3. Through weighting both means with their inverse variance, a combined estimator is derived. If the regression estimator has a larger variance, it therefore receives a lower weight and vice versa. These weights minimize the variance:

Sampling in Forest Inventories

827

^ þ w^ Y ^ ^A Y B B A ^C ¼w Y ^ w   ^ , w^ ¼ 1=^ ^ , and w ^A ¼ 1=^ ^¼w ^A þ w^B . where w var Y var Y A B B

(73)

4. As the last step, the variance of the combined estimator is calculated: 2  6 ^ ¼ 61 þ v^ar Y C 4

 ^B 4w^A w

3 1 1 þ n12  1 n2  1 7 7=w^ 5 ^2 w

(74)

 ^ and its variance v^ar Y ^ Once the current mean Y C C have been computed, an  ^ are easily obtained. The most ^ and its variance v^ar C estimate of change C straightforward estimation of change between two occasions is the combined ^: ^ , minus the mean calculated at previous occasion, X current estimator, Y C

I

^ ¼Y ^ X ^ C C I

(75)

^ is An approximation of the variance of C Xn1  ^B w ^ 2  1 s2 2 β^YX X  X 1j 1 j ^ ¼ þ X1  w^ v^ar SPR C ^ n1 ð n1  1Þ w

(76)

While SPR estimators are relatively simple for measurements on two occasions, the estimators quickly become complex and cumbersome to work with when dealing with measurements on three or more occasions. For example, where SPR is applied for three occasions, we have seven different plot types. To wit: n123, n12, n13, n23, n3, n1, n2 (Scott and Ko¨hl 1994). It is common practice to present inventory results in the form of tables. When using SPR estimators for different subpopulation (e.g., volume per species or forest type), the sum of subpopulation totals may not add up to the total for the entire population. In other words, SPR does not guarantee additive tables of subpopulation totals. This problem has been discussed in detail by Scott (1986). It is possible to adjust the tables as described by Li and Schreuder (1985), Green et al. (1992), and Ranneby and Rovainen (1995), among others. The problems in analyzing inventory data with SPR methods should not be underestimated (Scott 1986). The problems encountered in practical implementation with SPR are clear detractors. In some survey regions of the USA, SPR has recently been replaced by more flexible and less complex designs (Hahn and Scott 2003, personal communication) such as a semi-systematic sampling design where plot location

828

M. Ko¨hl and S. Magnussen

is random within a regular tessellation of the population into equal-sized hexagons (Bechtold and Patterson 2005; Barabesi and Franceschi 2011). The problems with unwieldy estimators and nonadditive tables can be largely avoided by refraining from the updating step in Eq. 68. Doing so, current values for time 2 are derived from new and unmatched plots, while the estimation of change between time 1 and time 2 is computed solely from the matched plots. This procedure may be less efficient than SPR but certainly a great deal easier (Scott 1986; Ko¨hl 1994).

Errors in Forest Surveys In sampling surveys, two types of errors are distinguished: sampling errors and non-sampling errors. Sampling errors result from the facts that only part of the population is surveyed and population parameters are estimated from the sample. The estimated parameters may deviate from the true population values. One way in which sampling error may be expressed is through the standard error of, say, a mean. An estimate of standard error ought to be given for all estimators, as it is essential for the correct interpretation of inventory data. Non-sampling errors, on the other hand, are not connected with the problem of dealing with only part of the population but arise from inaccurate measurements, less than perfect measuring devices, mistakes in the execution of the sampling plan or protocol, sampling the wrong units/elements, and errors in the sampling frame. These kinds of non-sampling errors are likely to either inflate the apparent sampling variance, introduce a bias in the estimates, or both. When the sample observations are derived from model-based predictions and not a direct measurement per se, the collected data are also subject to model error. Volume and biomass are typical attributes that are derived from models. Since models only predict the expected value, the apparent sampling variance of model-based predictions will be too low, unless the residual model variance is taken into account (Domke et al. 2012). From a statistical point of view, the reliability of results can be quantified by reporting their precision, accuracy, mean square error, and bias. Note, estimates of accuracy, mean square errors, and bias are only possible if we either know the true value of a parameter of interest or we have historic data and studies to help us arrive at reasonable estimates thereof. We provide a definition for each these terms here as a considerable confusion still lingers about their correct use. Precision Precision refers to the size of deviations in the estimate of a population parameter in repeat application of a sampling procedure. The standard error or confidence interval quantifies precision. Increasing the number of (independent) observations increases the precision of a statistical estimate. A dependency among observations slows the rate at which the precision would otherwise improve as the number of independent observations increases.

Sampling in Forest Inventories

829

Fig. 17 Example for the concept of accuracy and precision of an estimator (after Vanclay (1994)). If we assume that the target center is the true but unknown population value and that the position of each shot represents the estimate obtained by a random sample, it follows that in (1) the estimator is precise and accurate, in (2) the estimator is precise but biased, in (3) the estimator is imprecise but unbiased, and in (4) the estimator is imprecise and biased

Accuracy Accuracy refers to the size of the deviation between an observed value and the true value. Thus if we knew the true value of a population parameter, we could estimate the accuracy of a survey, but sampling would obviously not have been necessary. Bias Bias is directly related to the accuracy of an estimate. Bias is the difference between the estimated and true value of a parameter. We cannot estimate bias unless we know the true value of a parameter. In practice, we do not have this knowledge, but the sampling design and the methods used to obtain sample data paired with historic information and studies may give us reasonable expectations as to the potential of a bias and its possible magnitude. First-order approximations to the bias can sometimes be obtained via a Taylor series expansion (Wolter 2007). The effect of precision and bias can be seen in Fig. 17. Mean Square Error A useful measure of reliability is the mean square error (MSE). It combines the precision of an estimate with the square of the bias. The MSE is a useful criterion for the comparison of two or more competing estimators, possibly with different amount of bias. A direct comparison of estimators, in terms of precision only, may skew the choice toward estimators that generate highly precise but biased estimates. According to Cochran (1977, p. 15), the MSE of an estimate of, say, a population total is

M. Ko¨hl and S. Magnussen

830

      ^ Y^ þ Y^  Y 2 MSE Y^ ¼ Var

(77)

For unbiased estimators, MSE and precision are asymptotically (n ! N) identical. The estimators presented in section “Sampling Designs” are with few exceptions (e.g., the ratio estimator) unbiased for the estimation of population parameters. The standard procedure for calculating sampling error does not allow for the effect of bias. Nevertheless, bias may inflate the sampling error by several magnitudes (Gertner 1984; Gertner and Ko¨hl 1995). Increasing the sample size may certainly decrease the sampling error but it will, unfortunately, at the same time increase the relative influence of bias (if any). Consequently, the possibility of a bias should be vigorously investigated and zealously reduced from the earliest stage of planning a survey. It is often possible to assess various sources of inventory errors (Gertner and Ko¨hl 1995) and gauge if they have the potential to introduce a bias. Bias can also arise with an unintentional choice of a biased estimator. Retrospective simulation studies may be needed to quantify the potential magnitude and direction of a bias before attempting or proceeding with efforts to reduce an anticipated bias in an estimate.

Non-sampling Inventory Errors Estimators of population parameters and their sampling variances have so far been presented as if the observations were not only complete but also free of errors. Practical surveys can rarely live up to this ideal; forest inventories are no exception (Goelz and Burk 1996; Chen 1998; Lesser and Kalsbeek 1999; Chen and Cowling 2001). Best possible observations mean that the most accurate technique for obtaining data was applied everywhere and is the widely used standard for “best practices.”

Nonobservation The sample can be incomplete due to nonobservation of sample units or due to errors in the population frame from which sample units were selected (Sa¨rndal et al. 1992). Nonobservations of sample units can occur when some units were not visited because (1) access was denied, (2) access posed a danger to the survey crew, (3) sample units could not be located, and (4) sampling was terminated due to time or budgetary constraints. Measurement Errors Directly observed or compiled attribute values of a sampled population element (unit) are rarely, if ever, completely free of errors or bias. An observation deviating from the best possible observation is said to be in error and possibly biased.

Sampling in Forest Inventories

831

Conceptually, we can write an observed or a compiled attribute value yi0 for the ith population element as a linear sum of the best possible value yi plus a series of error and bias terms arising from various sources (k). For a univariate attribute, we can write our observations (compilations) as y0i ¼ yi þ

X

e k ik

þ

X

b k ik

¼ yi þ ei þ bi , Eðeik Þ ¼ 0, bik 6¼ 0

(78)

where ei and bi denote the sum of errors and bias in the observation (compilation) yi0 . The errors eik will depend on the attribute and on the entire process including the design that generates the observed or compiled values, yi0 (Domke et al. 2012).

Frame Errors Errors in the population frame usually result in an undercoverage; certain population units have a zero probability of inclusion in the sample or conversely an inflated inclusion probability if some units appear more than once in the frame. Regardless of cause, an incomplete frame will, as a rule, lead to biased estimates of both the population parameter and its sampling variance. Sa¨rndal et al. (2003) list problems related to errors in the population frame, among which are: • Elements outside the population of interest may be sampled (e.g., a non-forest unit has been classified as forest). • Some units of the population cannot be sampled because they do not appear in the frame (e.g., inaccessible forest areas or forest mistakenly assigned to a different class). • The list of population elements is dated and no longer adequate: the sample design may become inefficient and introduce bias (e.g., in ratio estimators when the ratio includes an estimate of population size or the size of subpopulations (domains, strata)).

Estimating Non-sampling Errors and Bias Bias and measurement errors can seriously compromise the quality and precision of a survey. Diligence and high standards are required in all aspects and all phases of a forest inventory to keep measurement errors and bias within narrow (tolerated) bounds. Quality standards and quality checks are integral parts of a forest inventory. Still, measurement errors and bias are virtually impossible to eliminate in a forest inventory. It is therefore good practice to investigate the impact of measurement errors and bias on survey results (Ko¨hl and Marchetti 2014). An error budget discloses all possible sources and an expected impact of error in the entire process that begins with a visit to a sample unit and ends with a set of survey estimates of population attributes is ideally suited for this purpose (Gertner and Ko¨hl 1995; Kangas 1998). The error budget will ideally not only disclose a possible bias in estimators but also suggest how better standards can mitigate the impact of bias and measurement errors. In a well-designed

832

M. Ko¨hl and S. Magnussen

inventory with high measurement and compilation standards, the contribution of natural intrinsic variation in attribute values to the overall sampling variance is usually orders of magnitude larger than the contribution of measurement errors and bias (McRoberts and Westfall 2013; Breidenbach et al. 2014).

References Agresti A (1992) A survey of exact inference for contingency tables. Stat Sci 7:131–177 Alder D, Synnott TJ (1992) In: Institute OF (ed) Permanent sample plot techniques for mixed tropical forests, Tropical forestry papers. Oxford Forestry Institute, Oxford, p 124 Baker PJ, Wilson JS (2000) A quantitative technique for the identification of canopy stratification in tropical and temperate forests. For Ecol Manage 127:77–86 Banyard SG (1987) Point sampling using constant tallies is biased: a tropical rainforest case study. Commonw For Rev 66:161–163 Barabesi L, Franceschi S (2011) Sampling properties of spatial total estimators under tessellation stratified designs. Environmetrics 22:271–276 Barnett DT, Stohlgren TJ (2003) A nested-intensity design for surveying plant diversity. Biodivers Conserv 12:255–278 Bechtold WA, Patterson PL (2005) The enhanced forest inventory and analysis program-national sampling design and estimation procedures. USDA Forest Service, Southern Research Station, Asheville Bellhouse DR (1985) Computing methods for variance estimation in complex surveys. J Off Stat 1:323–329 Bellhouse DR (1988) Systematic sampling. In: Krishnaiah PR, Rao CR (eds) Sampling. Elsevier, Amsterdam, pp 125–145 Bickford CA, Mayer CF, Ware KD (1963) An efficient sampling design for forest inventory: the northeastern forest survey. J For 61:826–833 Biolley HE (1921) L’aménagement des foreˆts par la méthode expérimentale et spécialment la méthode du controˆle. Neuchatel, Paris Bitterlich W (1947) Die Winkelza¨hlmessung. Allgemeine Forst- und Holzwirtschaftliche Zeitung 58:94–96 Boon DR (1970) A critical review of timber survey methods in tropical rainforest. International Institute for Aerial Survey and Earth Sciences, Delft, p 35 Bowden DC, Dixon GE, Frayer WE, Graybill FA, Jeyaratnam S, Johnston DC, Kent BM, LaBau VJ, Roberts E (1979) Multi-level sampling designs for resource inventories. Colorado State University, Ft. Collins Breidenbach J, Anto´n-Fernández C, Petersson H, McRoberts RE, Astrup R (2014) Quantifying the model-related variability of biomass stock and change estimates in the Norwegian national forest inventory. For Sci 60:25–33 Brewer KRW, Hanif M (1983) Sampling with unequal probability. In: Brillinger D, Fienberg SE, Gani J, Hartigan JA, Krickeberg K (eds) Lecture notes in statistics. Springer, Berlin/ Heidelberg/New York, pp 1–164 Casella G, Berger RL (2008) Statistical inference. Cengage Lerning Emea, Andover Chambers RL, Clark RG (2012) An introduction to model-based survey sampling with applications. Oxford University Press, New York Chen D (1998) Measurement errors in line transect surveys. Biometrics 54:899–908 Chen D, Cowling A (2001) Measurement errors in line transect surveys where detectability varies with distance and size. Biometrics 57:732–742 Chilés J-P, Delfiner P (2012) Geostatistics: modeling spatial uncertainty. Wiley, Hoboken Cochran WG (1977) Sampling techniques. Wiley, New York

Sampling in Forest Inventories

833

Correll RL, Cellier KM (1987) Effects of plot size, block size and buffer rows on the precision of forestry trials. Aust For Res 17:11–18 Da Silva JA (1982) Aplicacao do reloscopio de banda larga em inventarios de florestas tropicais. Silvic Sao Paulo 16A:601–603 Dalenius T, Gurney M (1951) The problem of optimal stratification Skand Akt 34:133–148 Diggle PJ, Ribeiro PJ (2007) Model-based geostatistics. Springer, New York Domke GM, Woodall CW, Smith JE, Westfall JA, McRoberts RE (2012) Consequences of alternative tree-level biomass estimation procedures on U.S. forest carbon stock estimates. For Ecol Manage 270:108–116 Ducey MJ (2009) Sampling trees with probability nearly proportional to biomass. For Ecol Manage 258:2110–2116 FAO (2010) Global forest resources assessment 2010: main report. Food and Agriculture Organization of the United Nations. FAO, Rome Fiuller WA (2006) Measurement error models. Wiley, New York Flewelling JW, Thomas CE (1984) An improved estimator for merchantable basal area growth based on point samples. For Sci 30:813–821 Fuller WA (2009) Sampling statistics. Wiley, New York Gertner GZ (1984) The sensitivity of measurement error in stand volume estimations. Can J For Res 16:1120–1123 Gertner GZ, Ko¨hl M (1995) An assessment of some nonsampling errors in a national survey using an error budget. For Sci 38:525–538 Goelz SJ, Burk TE (1996) Measurement error causes bias in site index equations. Can J For Res 26:1585–1593 Grafstro¨m A (2010) On unequal probability sampling designs. Ph.D. thesis, SLU, Umea˚ Graves HS (1906) Forest mensuration. Wiley, New York Gray A (2003a) Monitoring stand structure in mature coastal Douglas-fir forests: effect of plot size. For Ecol Manage 175:1–16 Gray A (2003b) Monitoring stand structure in mature costal Douglas-fir forests: effect of plot size. For Ecol Manag 175:1–16 Green EJ, Ko¨hl M, Strawderman WE (1992) Improved estimates for cell values in a two way table Biometrie. u. Informatik. in Medizin. u. Biologie 23:24–30 Gregoire TG (1998) Design-based and model-based inference in survey sampling: appreciating the difference. Can J For Res 28:1429–1447 Gregoire TG, Scott CT (1990) Sampling at stand boundaries: a comparison of the statistical performance among eight methods. In: Burkhart HE, Bonnor GM, Lowe JJ (eds) Research in forest inventory, monitoring, growth and yield. School of Forestry and Wildlife Resources, Virginia Polytechnic Institute and State University, Blacksburg, pp 78–85 Gregoire TG, Scott CT (2003) Altered selection probabilities caused by avoiding the edge in field surveys. JABES 8:36–47 Gregoire TG, Sta˚hl G, Naesset E, Gobakken T, Nelson R, Holm S (2011) Model-assisted estimation of biomass in a LiDAR sample survey in Hedmark County, Norway. Can J Forest Res 41:83–95 Grosenbaugh LR (1952) Plotless timber estimates – new, fast, easy. J For 50:32–37 Gurnaud A (1878) Cahier l’aménagement pour l’application de la méthode par contenance exposée sur la foreˆt des Eperous, 160 Halme M, Tomppo E (2001) Improving the accuracy of multisource forest inventory estimates by reducing plot location error – a multicriteria approach. Remote Sens Environ 78:321–327 Hartley HO (1966) Systematic sampling with unequal probability and without replacement. J Am Stat Assoc 71:739–748 Hess GR, Bay JM (1997) Generating confidence intervals for composition-based landscape indexes. Landsc Ecol 12:309–320 Hush B, Beers TW, Kershaw JA (2003) Forest mensuration. Wiley, New York

834

M. Ko¨hl and S. Magnussen

Jenks GF (1967) The data model concept in statistical mapping. In: Frenzel K (ed) International yearbook of cartography. Rand McNally, Skokie Kangas A (1998) On the bias and variance in tree volume predictions due to model and measurement errors. Scand J For Res 11:281–290 Keen EA (1950) The relaskop. Emp For Rev 29:253–264 Keller M (2001) Aerial photography. In: Brassel P, Lischke H (eds) Swiss National Forest Inventory; methods and models of the second assessment. WSL, Birmensdorf, pp 45–64 Ko¨hl M (1994) Statistisches Design f€ ur das zweite Schweizerische Landesforstinventar: Ein Folgeinventurkonzept unter Verwendung von Luftbildern und terrestrischen Aufnahmen. In: Eidgeno¨ssischen Forschungsanstalt f€ ur Wald, S.u.L. (ed) Mitteilungen der WSL, Birmensdorf, p 141 Ko¨hl M, Brassel P (2001) Zur Auswirkung der Hangneigungskorrektur auf Scha¨tzwerte im Schweizerischen Landesforstinventar (LFI). Schweizerische Zeitschrift f€ ur Forstwesen 152:215–225 Ko¨hl M, Kushwaha SPS (1994) A four-phase sampling method for assessing standing volume using landsat-TM-data, aerial photography and field assessments. Commonw For Rev 73:35–42 Ko¨hl M, Marchetti M (2014) In: Pancel L, Ko¨hl M (eds) Objectives and planning of forest inventories, Tropical forestry handbook. Springer, Heidelberg Ko¨hl M, Scott CT (1994) Zur Auswertung von Gruppenstichproben bei extensiven Forstinventuren. Allgemeine Forst- und Jagdzeitung 165:101–106 Ko¨hl M, Magnussen S, Marchetti M (2006) Sampling methods, remote sensing and GIS multiresource forest inventory. Springer, Berlin/Heidelberg Ko¨hl M, Scott CT, Lister AJ, Demon I, Plugge D (2015) Avoiding treatment bias of REDD+ monitoring by sampling with partial replacement. Carbon Balance and Management Kuusela K (1979) Sampling of tree stock by angle gauge in proportion to tree characteristics. Commun Inst For Fenn 95(7):14 Laurans M, Hérault B, Vieilledent G, Vincent G (2014) Vertical stratification reduces competition for light in dense tropical forests. For Ecol Manage 329:79–88 Lesser VM, Kalsbeek WD (1999) Nonsampling errors in environmental surveys. JABES 4:473–488 Levy PS, Lemeshow S (2008) Sampling of populations: methods and applications. Wiley, New York Li HG, Schreuder HT (1985) Adjusting estimates in large two-way tables in surveys. For Sci 31:366–372 Little RJA (2004) To model or not to model? Competing modes of inference for finite population sampling. J Am Stat Assoc 99:546–556 Lloyd CJ (1999) Analysis of categorical variables. Wiley, New York Loetsch F (1957) Report to the government of Thailand on inventory methods for tropical forests. FAO report. FAO, Rome Loetsch F, Zo¨hrer F, Haller KE (1973) Forest inventory. BLV Verlagsanstalt, M€ unchen/Bern/ Wien Lombardi F, Marchetti M, Corona P, Merlini P, Chirici G, Tognetti R, Burrascano S, Alivernini A, Puletti N (2015) Quantifying the effect of sampling plot size on the estimation of structural indicators in old-growth forest stands. For Ecol Manage 346:89–97 Magnussen S (1988) Tree height tarifs and volume estimation errors in New Brunswick. North J Appl For 15:7–13 Magnussen S (2002) Fast pre-survey computation of the mean spatial autocorrelation in large plots composed of a regular array of secondary sampling units (SSUs) depend on the average withinplot autocorrelation. Math Modell Sci Comput 13:204–217 Magnussen S (2003) Stepwise estimators for three-phase sampling of categorical variables. J Appl Stat 30:461–475 Magnussen S (2015) Arguments for a model based inference? For Int J For Sci 88(3):317–316

Sampling in Forest Inventories

835

Mandallaz D (2008) Sampling techniques for forest inventories. Chapman & Hall/CRC, Boca Raton Matern B (1980) Spatial variation. Springer, Berlin/Heidelberg/New York McRoberts RE, Westfall JA (2013) Effects of uncertainty in model predictions of individual tree volume on large area volume estimates. For Sci 60:34–43 McRoberts RE, Gobakken T, Næsset E (2012) Post-stratified estimation of forest area and growing stock volume using lidar-based stratifications. Remote Sens Environ 125:157–166 McRoberts RE, Liknes GC, Domke GM (2014) Using a remote sensing-based, percent tree cover map to enhance forest inventory estimation. For Ecol Manage 331:12–18 Mesavage C, Grosenbaugh LR (1956) Efficiency of several cruising designs on small tracts in North Arkansas. J For 54:569–576 O’Regan WG, Arvanitis LG (1966) Cost effectiveness in forest sampling. For Sci 12:406–414 Olsen A, Sedransk J, Edwards D, Gotway C, Ligget W, Rathburn S, Reckhow K, Young L (1999) Statistical issues for monitoring ecological and natural resources in the United States. Environ Monit Assess 54:1–45 Palace MW, Sullivan FB, Ducey MJ, Treuhaft RN, Herrick C, Shimbo JZ, Mota-E-Silva J (2015) Estimating forest structure in a tropical forest using field measurements, a synthetic model and discrete return lidar data. Remote Sens Environ 161:1–11 Pandey D (2008) National forest inventory in India. Workshop on “Monitoring of reduction of emissions from forest degradation” Paris, COMIFAC. http://pfbc-cbfp.org/docs/reddparis032008/13%20-%20Pandey%20National%20lForest%20Inventory-India.pdf Payandeh B, Ek AR (1986) Distance methods and density estimators. Can J For Res 16:918–924 Pelz D, Cunia T (1985) Forest inventories in Europe. Mitt Abt Forstl Biom, Freiburg, vol 3, p 85 Pollard JH (1971) On distance estimators of density in randomly distributed forests. Biometrics 27:991–1002 Pommerening A, Schmidt M (1998) Modifizierung des Stammabstandsverfahrens zur Verbesserung der Stammzahl- und Grundfla¨chenscha¨tzung. Forstarchiv 69:47–53 Pretzsch H (2009) Forest dynamics, growth and yield: from measurement to model. Springer, Berlin/Heidelberg Prodan M (1965) Holzmesslehre. Sauerla¨nders, Frankfurt Prodan M (1968) Punktstichprobe f€ ur die Forsteinrichtung. Forst- u Ho/zw 23:225–226 Puffenberger HE (1976) Forest inventories in the tropics: a consideration. J For 74:28–38 Ranneby B, Rovainen E (1995) On the determination of time intervals between remeasurements of permanent plots. For Ecol Manage 71:195–202 Rao CR (1988) Variance estimation in sample surveys. Elsevier, Amsterdam Rao JNK (2003) Small area estimation. Wiley, Hoboken Rice B, Weiskittel A, Wagner R (2014) Efficiency of alternative forest inventory methods in partially harvested stands. Eur J For Res 133:261–272 Roesch FA, Van Deusen PC (2013) Time as a dimension of the sample design in national-scale forest inventories. For Sci 59:610–622 Royall RM (2001) Model robust confidence intervals using maximum likelihood estimators. Int Stat Rev 54:221–226 Saarela S, Grafstro¨m A, Sta˚hl G, Kangas A, Holopainen M, Tuominen S, Nordkvist K, Hyyppa¨ J (2015) Model-assisted estimation of growing stock volume using different combinations of LiDAR and Landsat data as auxiliary information. Remote Sens Environ 158:431–440 Saborowski J (1990) Scha¨tzung von Varianzen und Konfidenzintervallen aus mehrstufigen Stichproben am Beispiel von Wladschadensinventuren. In, Schriften aus der FOrstlichen Fakulta¨t der Universita¨t Go¨ttingen und der niedersa¨chsischen forstlichen Versuchsanstalt Go¨ttingen Saborowski J, Smelko S (1998) Evaluation of inventories based on sample plots of variable size. Allg Forst- und Jagdz 169:71–75 Sa¨rndal C-E, Swensson B, Wretman J (2003) Model assisted survey sampling. Springer, Heidelberg

836

M. Ko¨hl and S. Magnussen

Schlund M, von Poncet F, Kuntz S, Schmullius C, Hoekman DH (2015) TanDEM-X data for aboveground biomass retrieval in a tropical peat swamp forest. Remote Sens Environ 158:255–266 Schmid P (1969) Stichproben am Waldrand. In: Eidgen. Anstalt Forstl Versuchswes (ed) Bericht, pp 234–303 Schreuder HT, Banyard SG, Brink GE (1987) Comparison of three sampling methods in estimating stand parameters for a tropical forest. For Ecol Manage 21:119–127 Schreuder H, Gregoire TG, Wood GB (1992) Sampling methods for multiresource forest inventory. Wiley, New York Schreuder HT, Li J, Scott CT (1993) Estimation with different stratification at two occasions. For Sci 39:368–382 Schroeder TA, Healey SP, Moisen GG, Frescino TS, Cohen WB, Huang C, Kennedy RE, Yang Z (2014) Improving estimates of forest disturbance by combining observations from Landsat time series with U.S. Forest Service Forest Inventory and Analysis data. Remote Sens Environ 154:61–73 Scott CT (1984) A new look at sampling with partial replacement. For Sci 30:157–166 Scott CT (1986) An evaluation of sampling with partial replacement. in: use of auxiliary information in natural resource inventories. Society of American Foresters (SAF), Blacksburg, pp 74–79 Scott CT (1998) Sampling methods for estimating change in forest resources. Ecol Appl 8:228–233 Scott CT, Ko¨hl M (1994) Sampling with partial replacement and stratification. For Sci 40:30–46 Sherman M (1996) Variance estimation for statistics computed from spatial lattice data. J R Stat Soc Ser B 58:509–523 Sims DA, Gamon JA (2003) Estimation of vegetation water content and photosynthetic tissue area from spectral reflectance: a comparison of indices based on liquid water and chlorophyll absorption features. Remote Sens Environ 84:526–537 Snedecor GW, Cochran WG (1989) Statistical methods. 8th edn. Iowa State University Press, Iowa City Stehman SV (1992) Comparison of systematic and random sampling for estimating the accuracy of maps generated from remotely sensed data. Photogramm Eng Remote Sens Environ 58:1343–1350 Stevens D, Olsen A (2004) Spatially balanced sampling of natural resources. J Am Stat Assoc 99:262–278 Stott CB, Ryan EJ (1939) A permanent sample technique adapted to commercial timber stands. J For 37:347–349 Stott CB, Semmes G (1962) Our changing inventory methods and the CFI system in North America. In: Proceedings of the 5th world forest congress, Seattle ˚s Strand L (1958) Sampling for volume along a line. Norske Skogsforsøksv, A Sukhatme PV, Sukhatme BV, Sukhatme S, Asok C (1984) Sampling theory of surveys with applications. Iowa State University Press, Ames Tardif G (1965) Some considerations concerning the establishment of optimum plot size in forest survey. For Chron 41:93–102 Thomas L, Buckland ST, Rexstad EA, Laake JL, Strindberg S, Hedley SL, Bishop JRB, Marques TA, Burnham KP (2010) Distance software: design and analysis of distance sampling surveys for estimating population size. J Appl Ecol 47:5–14 Thompson SK (1992) Sampling. Wiley, New York, p 343 Tomppo E, Gschwantner T, Lawrence M, McRoberts RE (2010) National forest inventories – pathways for common reporting. Springer, Heidelberg Tomppo EO, Heikkinen J, Henttonen HM, Ihalainen A, Katila M, Ma¨kela¨ H, Tuomainen T, Vainikainen N (2011) Designing and conducting a forest inventory – case: 9th National Forest Inventory of Finland. Springer, Dordrecht Uebelho¨r K (1988) Praktische Erfahrungen mit Winkelza¨hlprobe und Breitbandrelaskop im tropischen Regenwald. Forstarchiv 59:47–52

Sampling in Forest Inventories

837

Valbuena R, Mauro F, Rodriguez-Solano R, Manzanera JA (2010) Accuracy and precision of GPS receivers under forest canopies in a mountainous environment. Span J Agric Res 8:1047–1057 Vanclay JK (1994) Modelling forest growth and yield: applications to mixed tropical forests. CAB, Wallingford Veloso de Freitas J (2009) Forest monitoring: the national forest inventory-Brazil. In: Brazilian Norwegian workshop on forest and marine monitoring (ed). http://www.dsr.inpe.br/Brazil_ Norway_Workshop/JOBERTO_FREITAS_The_National_Forest_Service_Inventory_and_ Monitoring_Program.pdf Walters DK, Burkhart HE, Reynolds MRJ, Gregoire TG (1991) A Kalman filter approach to localizing height-age equations. For Sci 37:1526–1537 Ware KD, Cunia T (1965) Continuous forest inventory, partial replacement of samples and multiple regression. For Sci 11:480–502 Weiskittel AR, Hann DW, Kershaw JA, Vanclay JK (2011) Forest growth and yield modeling. Wiley, Chichester Williams DL (1996) Adaptive optimization of renewable natural resources: solution algorithms and a computer program. Ecol Modell 93:101–111 Wolter MK (1985) Introduction to variance estimation. Springer, Berlin/Heidelberg/New York Wolter KM (2007) Introduction to variance estimation. Springer, New York Wulder FSE, Lavigne MB (1996) High spatial resolution optical image texture for improved estimation of forest stand leaf area index. Can J Remote Sens 24:441–449 Zar JH (2010) Biostatistical analysis. Prentice Hall, Upper Saddle River Zeide B (1980) Plot size optimization. For Sci 26:251–257 Zhu JJ, Matsuzaki T, Gonda Y (2003) Optical stratification porosity as a measure of vertical canopy structure in a Japanese coastal forest. For Ecol Manage 173:89–104 Zingg A (1988) Schweizerisches Landesforstinventar, Anleitung f€ ur die Erstaufnahme 1982–1986. In: Versuchswesen EAF (ed), Birmensdorf

Measurement, Reporting, and Verifications Systems in Forest Assessment €bler, Prem Raj Neupane, Daniel Plugge, Daniel Ku Konstantin Olschofsky, and Laura Prill

Contents Rationale: Why Information on Biomass and Carbon Is Needed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biomass Distribution World Wide: The Importance of Tropical Forests . . . . . . . . . . . . . . . . . . . . . Different Pools for Biomass Assessments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Estimating Biomass Changes Over Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Uncertainties and Different Tier Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Assessment of the Aboveground Biomass Pool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Destructive Assessment of Aboveground Biomass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nondestructive Assessment of Aboveground Biomass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biomass Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Belowground Biomass and Soil Organic Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Root Biomass Assessment Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Soil Organic Carbon (SOC) Assessment Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dead Organic Matter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Deadwood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Litter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Upscaling Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality Assurance and Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Standard Operating Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Field Crews . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Remeasurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

840 843 844 846 846 847 848 853 858 859 859 863 866 866 872 874 875 876 876 876

D. Plugge (*) SPC/GIZ Regional Program, REDD+ - Forest Conservation in Pacific Island Countries II, Deutsche Gesellschaft f€ ur Internationale Zusammenarbeit (GIZ), Suva, Fiji Institute for Worldforestry, University of Hamburg, Hamburg, Germany e-mail: [email protected] D. K€ubler • K. Olschofsky • L. Prill Institute for Worldforestry, University of Hamburg, Hamburg, Germany P.R. Neupane University of Hamburg, World Forestry, EFI (European Forest Institute) Project Centre SURF (Supporting the Global Implementation of REDD+ and FLEGT), Hamburg, Germany # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_73

839

840

D. Plugge et al.

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 877

Abstract

The assessment of biomass and carbon in tropical forests is gaining more and more attention. This is due to their role of providing livelihoods to more than a billion people, their vast timber resources, their provision of shelter to more than half of the known plant and animal species, as well as to the role of tropical forests in the global climate system and climate change. They represent the biggest terrestrial carbon pool and cycle more water than any other biome. The high uncertainty about the amount of biomass and carbon stored in tropical forests and their decrease due to anthropogenic deforestation and forest degradation renders more detailed knowledge necessary to better understand the real importance of this biome. In this chapter we are going to introduce different techniques for the assessment of biomass and its translation into carbon that is stored in tropical forests to describe the status quo at a given point in time. The calculation of changes over time requires assessments on at least two occasions and subsequent application of statistical methods which are described in other chapters of this book. We focus on the five biomass pools as stipulated by the Intergovernmental Panel on Climate Change (IPCC), which are aboveground and belowground biomass and deadwood, litter, and soil organic matter. For each pool the specific requirements for assessing biomass are explained and the single-assessment steps are described. Besides this, we give an overview on commonly used reference values from the literature for the case that a physical assessment of a certain pool is not feasible.

Rationale: Why Information on Biomass and Carbon Is Needed Knowledge about biomass in forests is getting more important for an increasing number of stakeholders. It can serve as information for the amount of energy that a specific forest area holds either for energy purposes or as the fuel load available for forest fires. It can satisfy the need for information on the amount of carbon that is stored inside forest ecosystems, e.g., for Clean Development Mechanism (CDM) projects or the REDD+ (reducing emissions from deforestation and forest degradation, the role of conservation, the sustainable management and enhancement of forest carbon stocks in developing countries) mechanism (see Box 1). It further can be important for the identification of the ratio of harvested wood for construction or timber and the amount of organic material that is left in the forest for habitat and final decomposition. Knowledge about biomass therefore requires more information than can be obtained from traditional forest inventories focusing on the amount of harvestable volume only. Young (1978) lists various additional aspects that need to be addressed during a biomass inventory compared to traditional inventories:

Measurement, Reporting, and Verifications Systems in Forest Assessment

841

• • • • • •

More parts of the tree have to be considered Smaller tree sizes are to be included More tree species, not only commercial ones, have to be assessed Additional parameters are required, e.g., information of the shape of the crown Different regression equations Different preferences with regard to the sampling procedures and inventory design • More complex calculations • The results will have a wider demand Box 1: Reducing Emissions from Deforestation and Forest Degradation – The Role of Conservation, Sustainable Management, and Enhancement of Forest Carbon Stocks in Developing Countries (REDD+)

REDD+ is a mechanism that is discussed under the United Nations Framework Convention on Climate Change (UNFCCC) (see also chapter “▶ International Processes: Framework Conditions for Tropical Forestry”). Its intention is to give a monetary value to the carbon that is stored in the (tropical) forests. The idea for this mechanism was first introduced by the Coalition for Rainforest Nations (CfRN), led by the governments of Papua New Guinea and Costa Rica to the Conference of the Parties (COP) at the 11th session of UNFCCC in Montreal in 2005. In their submission CfRN considered that so far, developing countries were not actively contributing to mitigation and adaptation activities and that the reduction of deforestation rates was not an eligible action for emission reductions under the Kyoto Protocol (UNFCCC 2005). CfRN stated that tropical forests are equivalent to a wealth of global importance, considered as a common good by the global society, despite the fact that it had to be administrated on a national level. Preserving forests against pressures driven by the needs of the rural poor and the financial and political power of large agricultural or livestock companies and the demand for commodities from developed countries is a futile task as long as the forest itself has low economic value. From 2005 onwards the idea of REDD and its implementation have been discussed broadly in the political and scientific arenas. Especially, the results of the Stern Review on the Economics of Climate Change (Stern 2006), stipulating that the reduction of tropical deforestation would be the most cost-effective way of tackling climate change, together with the estimate of IPCC (2007) that about one fifth of the anthropogenic release of GHGs is due to tropical deforestation, fuelled the action toward this topic. Since the Climate Change Conference in Bali 2007, REDD has been officially negotiated as a climate change mitigation mechanism for tropical forests under the Ad Hoc Working Group on Longterm Cooperative Action under the Convention (AWG-LCA) and Subsidiary Body for Scientific and Technological Advice (SBSTA). Regarding (continued)

842

D. Plugge et al.

international negotiations, the following five key decisions are of importance in describing the progress of REDD+ under the UNFCCC: • • • • •

Decision 1/CP.13 (Bali Action Plan) Decision 2/CP.13, both from COP13 in Bali (UNFCCC 2008) Decision 4/CP.15 from COP15 in Copenhagen (UNFCCC 2010) Decision 1/CP.16 from COP16 in Cancun (UNFCCC 2011) Decisions 9–15/CP.19 from COP19 in Warsaw (the Warsaw Framework for REDD+) (UNFCCC 2014)

While at the beginning a process to “avoid deforestation” was discussed, nowadays a global mechanism is negotiated that embraces the five national activities, namely, (i) reducing emissions from deforestation, (ii) reducing emissions from forest degradation, (iii) sustainable management of forests, (iv) conservation of forest carbon stocks, and (v) enhancement of forest carbon stocks and holds a multitude of social, ecological, and governmental safeguards that have to be respected. All this is subsumed in the term REDD+. Mostly, the implementation of REDD+ is outlined in a phased approach, moving in a stepwise fashion from pilot activities to full, result-based REDD+ implementation (Meridian Institute 2009; UNFCCC 2011). Fuller (2006) relates REDD+ to a new era of transparency in forest governance. Given the present state of the negotiations on REDD+, countries aspiring to generate benefits herein need to consider several components. They must implement, among other things, sound systems for measuring, reporting, and verification (MRV) of carbon stocks and carbon stock changes (UNFCCC 2012). These systems must allow for the identification of all processes leading to deforestation and forest degradation and for the sensible quantification of emissions hereof. In addition to the MRV system, an efficient implementation of the potential REDD+ mechanism requires a system for identifying and quantifying local and regional drivers of deforestation and forest degradation and adapting incentive schemes to manage with these drivers. Both preceding components allow for the definition of a forest reference (emission) level, against which the reduced emissions can be measured and benefits can be considered. Finally, an incentive scheme has to be set up offering different regionally or locally adapted options that en masse sustainably contribute on a national scale to the set of five REDD+ activities. It becomes obvious that while REDD+ originates from a simple idea, i.e., giving a financial value to the carbon stored in forests, the implementation of this idea requires a highly complex and multi-targeted approach (Adapted from Ko¨hl et al. 2014).

Measurement, Reporting, and Verifications Systems in Forest Assessment

843

This results in more areas that have to be inventoried, i.e., areas assigned to the class “other wooded land” (shrubland, wooded grassland, etc.) and thus increases the time and cost of such inventories. It is therefore very important to specify the information needs of stakeholders interested in the inventory results to avoid unnecessary efforts. This also includes the selection of biomass parts that have to be included in the assessment, e.g., whether they have a commercial value or add significant amounts to the fuel load of a forest. In the next subchapters the options for the assessment of the five different biomass pools as defined by the IPCC (2003) are presented. To control the effort needed for an assessment, certain parts within one specific pool may be excluded. For example, for the aboveground biomass (AGB) pool, decisions need to be taken whether understory trees and shrubs and ground vegetation should be included and if palms or bamboos are of interest for the survey. While the latter may constitute significant amounts of biomass in certain forest types, the former often accounts for only a small fraction of the overall aboveground biomass. Assessment of this vegetation can therefore be restricted to only a subsample of the complete assessment. Of high importance is also to clearly define threshold values for the inclusion of, e.g., small trees (via a minimum diameter at breast height), roots (via minimum diameter), or deadwood (via decay classes). The definitions chosen have to meet the demands of stakeholders while not comprising the inventory efforts in terms of time and costs.

Biomass Distribution World Wide: The Importance of Tropical Forests Carbon is sequestered through photosynthesis from the atmosphere and stored into the living but also dead biomass of plants. Forests display the biggest terrestrial biomass stocks and thus the biggest carbon pool on the land surface. Roughly 80 % of the total global land-based biomass is stored in forests, which contribute 75 % to the total biomass production (Beer et al. 2010; Kindermann et al. 2008). The overall above- and belowground forest biomass storage amounts to 600 Gt (equal to 600 Pg), which represents 149 t/ha on average over the whole area of forests (FAO 2010). Table 1 displays the biomass stock categorized by regions and subregions following the Forest Resource Assessment (FRA) of the Food and Agricultural Organization of the United Nations (FAO) in 2010 (FAO 2010). The highest stocks can be found in the tropical regions of South America and Western and Central Africa with nearly 250 t/ha. Pan et al. (2011) focus on the estimation of carbon that is stored in the three big global forest biomes. According to their findings, boreal forests store 272  23 Pg of carbon of which most is allocated in the soil. In temperate zones, 119  9 Pg of carbon is stored in the forest biomass. Compared to these two biomes, the tropical forests show the highest carbon storage with an estimated 471  93 Pg of which, in contrast to boreal forests, most is stored in the aboveground biomass.

844

D. Plugge et al.

Table 1 Biomass by region and subregion in 2010 (Source: FAO FRA 2010) Region/subregion Eastern and Southern Africa Northern Africa Western and Central Africa Total Africa East Asia South and Southeast Asia Western and Central Asia Total Asia Europe excl. Russian Federation Total Europe Caribbean Central America North America Total North and Central America Total Oceania Total South America World

Biomass Million tonnes 33,385 3,711 81,603 118,700 18,429 51,933 3,502 73,864 25,602 90,602 1,092 3,715 76,929 81,736 21,302 213,863 600,066

t/ha 124.8 47.1 248.7 176.0 72.4 176.4 80.5 124.7 130.7 90.2 157.5 190.5 113.3 115.9 111.3 247.4 148.8

Different Pools for Biomass Assessments The differentiation of Pan et al. between aboveground and soil biomass/carbon illustrates that there are different pools that need to be considered in estimating biomass and/or carbon in tropical forests. To completely assess the biomass and carbon that are stored in forests, other pools need to be considered as well: the belowground biomass (BGB), litter, and the coarse woody debris also often referred to as deadwood (see Fig. 1). These five pools are also suggested by the IPCC Good Practice Guidance for Land Use, Land-Use Change and Forestry (GPG-LULUCF) (IPCC 2003) to be considered for carbon stock assessments in forests (see Table 2). The IPCC organizes these pools into living biomass, dead organic matter, and soils. Table 2 displays the pools and their definitions according to the GPG-LULUCF. We are going to introduce the assessment of the five pools described above in the following. This chapter focusses on the different methods to estimate biomass from terrestrial measurements carried out directly in tropical forests. The capabilities of remote sensing techniques in estimating forest biomass are described in brief in subchapter (133.2.2) and extensively in a specific chapter of this book (Baldauf and Garcia, “▶ Image Processing of Radar and Lidar in Tropical Forestry”). Also, different statistical sampling techniques and the amount of samples needed to derive reliable estimates on biomass are described in a separate chapter of this book (“▶ Measurements and Assessments on Field Plots” and “▶ Objectives and Planning of Forest Inventories”).

Measurement, Reporting, and Verifications Systems in Forest Assessment

845

Fig. 1 The five forest biomass/carbon pools (Source: DiRocco et al. 2014)

Table 2 Definitions of the five different biomass pools according to the GPG-LULUCF (Adapted from IPCC 2003) Pool Living biomass

Dead organic matter

Aboveground biomass Belowground biomass Dead wood

Litter

Soils

Soil organic matter

Description All living biomass above the soil including stem, stump, branches bark, seed, and foliage All living biomass of live roots. Fine roots of less than 2 mm diameter are often excluded because these cannot be distinguished from soil organic matter or litter Includes all nonliving woody biomass not contained in the litter, either standing, lying on the ground, or in the soil. Includes dead wood, roots, and stumps larger than or equal to 10 cm Includes all nonliving biomass below a threshold diameter chosen by a country (e.g., 10 cm), lying dead, in various states of decomposition above the soil Includes organic carbon in mineral and organic soils (including peat) to a specified depth chosen by the country. Live fine roots (e.g., below 2 mm) are included with soil organic matter

846

D. Plugge et al.

Estimating Biomass Changes Over Time To assess changes in biomass over time, sampling on successive occasions (Cunia 1979; Ko¨hl et al. 1995) or the development of models to extrapolate data from one point in time to another (Hush et al. 2003) is needed. Cost- and time-efficient biomass monitoring mostly concentrates on the AGB pool (Murdiyarso et al. 2008). The monitoring of biomass changes over large areas can be facilitated with remote sensing techniques (Gibbs et al. 2007; link to chapter Baldauf and Jimenez), but, despite recent developments in technologies (see section “Estimating Biomass by Remote Sensing and Terrestrial Laser Scanning”), are often still not sensitive enough to provide reliable information of in situ changes (Baldauf et al. 2009). Therefore, field assessments, as introduced here, are needed to reduce the uncertainties of pure remote sensing-based estimations. In vast tropical areas, these assessments have to be sample based. Methods for the design and implementation of such field inventories are broadly available (Brown 1997; Ko¨hl et al. 2006) and described in this book (“▶ Measurements and Assessments on Field Plots” and “▶ Objectives and Planning of Forest Inventories”). Following IPCC methodologies (IPCC 2003, 2006), biomass changes per unit area can be estimated via the “Gain-Loss Method,” based on estimates of the actual annual change from gain and loss in biomass and the “Stock-Difference Method” which concentrates on the estimated difference in overall biomass carbon stocks between two points in time. The estimation of biomass over time is not in the scope of this chapter. However, the methods described here can be used to assess biomass at different points in time and thereby provide the input into change estimate methods.

Uncertainties and Different Tier Methods To identify the reliability of an estimate, information on the precision, accuracy, bias, and mean square error (MSE) are needed. Cochran (1977) gives an extensive introduction into this field of sampling statistics. Chave et al. (2005) state that tropical forest allometric models used for AGB estimation suffer from three important shortcomings: (i) they are constructed from limited samples, (ii) they are sometimes applied beyond their valid diameter range, and (iii) they rarely take into account available information on wood density. When using the default values and regressions, the sources of the values and equations should be checked, and their correspondence with specific conditions of the study area should be examined. Even when constructing own biomass functions for a specific area of interest with advanced models and techniques according to a Tier 3 approach, uncertainties will arise in all of the five biomass pools. It is important to differentiate between different types of errors to understand how the error sources can be tackled. A major difference exists between sampling and non-sampling errors. Sampling errors occur in any type of environmental assessment (Gertner and Ko¨hl 1992) and can be controlled by increasing the number of observations. Different parameters influence the reliability of sampling, e.g., when assessing root biomass as the by far

Measurement, Reporting, and Verifications Systems in Forest Assessment

847

least measured carbon pool due to the difficulty in measuring or modeling stocks or growth rates and high assessment costs (Ravindranath and Ostwald 2008). Several biotic (e.g., stand, tree age, tree species) and abiotic factors (e.g., soil moisture, soil fertility, soil texture, precipitation) might influence the root biomass. To reduce the uncertainty, the IPCC (2006) encourages to develop country- or region-specific root to shoot ratios and regression equations to estimate and predict the growth of root biomass. Besides sampling errors, Lessler and Kalsbeek (1992) describe different types of non-sampling errors, like frame errors, function errors, or classification errors. These errors need to be prevented in the best case before a survey takes place or can only be tackled with analytical measures after the survey. To allow countries to adapt estimation methods to their national circumstances and capacities, three methodological tiers with increasing complexity are offered by the IPCC. Tier 1 mainly provides equations and default values given in the IPCC guidelines and is therefore subject to high uncertainties up to a range of 50 %. Tier 2 utilizes country- or region-specific data with higher temporal and spatial resolution to determine the changes in forest areas and carbon stocks for the most important land-use categories. Tier 3 uses methods that include tailor-made country-specific models and inventory measurement systems. The methods consist of high-resolution remote sensing data and comprehensive field assessments at successive occasions amended by land-use data analysis with GIS-based systems. Higher-tier methods reduce the uncertainty of estimates but add to the complexity, time, and cost efforts for the detailed and high-resolution monitoring systems. As not all biomass carbon pools may be of the same importance for a specific assessment, the GPG-LULUCF allows for using different tiers for different pools.

Assessment of the Aboveground Biomass Pool In the following, the possibilities for assessing the aboveground biomass pool are described including destructive (direct) and nondestructive (indirect) measurement techniques. The direct assessment of biomass of any of the five pools is a laborintensive and tedious work. Therefore, in most cases where information about biomass or carbon of a specific forest area is needed, faster and less cost-intensive methods are required due to imposed time and budget restrictions. That is why most inventories focus on the indirect assessment of the aboveground biomass pool only, as this is the largest pool in tropical forests and it can be assessed with the least effort and well-defined assessment techniques and protocols (Murdiyarso et al. 2008). Nevertheless, direct assessments are indispensable to construct regression functions that allow for estimating biomass through the application of parameters derived from indirect assessments. It is advised to combine these two techniques in a two-phase approach. In the first phase a small subset of trees is destructively sampled to derive biomass functions for single species or groups of species. In the second phase, a samplebased broad-scale inventory is carried out, assessing tree parameters that provide a functional relationship to the biomass of a tree and serve as input into the previously

848

D. Plugge et al.

derived biomass functions. Picard et al. (2012) give an exhaustive introduction into the single steps of the first phase. Here, we want to summarize the main steps and introduce the necessary prerequisites for the derivation of AGB functions. To avoid misinterpretation of the term “aboveground biomass,” we use the definition as stipulated by Brown (1997) in the following: The total amount of above ground living organic matter in trees expressed as oven-dry tonnes per unit area.

Destructive Assessment of Aboveground Biomass The destructive assessment of aboveground biomass is the only way to derive functions that relate easily assessable tree attributes like diameter at breast height (DBH) and tree height to the total AGB of a single tree. As the workload of this task is enormous, it is oftentimes avoided and readily available biomass functions using a set of independent variables or biomass expansion factors (BEFs, see section “Biomass Functions”) which relate the volume of a tree to the overall AGB are used. Furthermore, under some circumstances, e.g., protected areas or dry forest areas with low regeneration and regrowth, destructive sampling may be prohibited. However, given the high number of different forest types in the tropics and the large amount of tree species, destructive measurement should be preferred whenever possible to achieve functions that are appropriate for a specific area. Any survey concerning the assessment of biomass therefore has to check carefully whether sufficient biomass functions or BEFs exist for their region or whether they need to be constructed first. In the latter case the following steps need to be considered.

Grouping of Tree Species In an ideal case for each species, a specific function would have to be derived via destructive sampling. This, however, appears to be unrealistic in tropical forests due to the high tree species diversity. Therefore, tree species should be grouped based on existing knowledge and data regarding their phenotypes, growth habits, and spatial distribution across the region of interest. If this information is not available before the survey, the grouping needs to be carried out after the destructive sampling took place. In both cases the groups need to be tested statistically to derive regression functions that are significantly different from each other to avoid overlapping between groups (Ko¨hl et al. 2006). Sample Tree Selection The selection procedure for sample trees is driven by three requirements (Cunia 1979): 1. The sample trees should be representative for the population of interest. 2. The selection procedure should allow for a valid regression analysis. 3. The selection procedure should be cost-efficient.

Measurement, Reporting, and Verifications Systems in Forest Assessment

849

A random selection of trees for destructive sampling is not advised as the trees should cover the whole range of sizes for the studied species. Stewart et al. (1992) introduced a method to allow for the objective selection of trees by forming classes of trees according to their diameter and then randomly selecting trees out of these classes. Trees should be selected independently for each species or group and across the whole range of site conditions and possible management types in the area of interest (see Ko¨hl et al. 1995). To derive the number of trees to be selected for each species or species group for the construction of regression functions, information on several variables has to be assessed before the actual felling of the trees takes place. The number of trees to be felled can then be calculated according to the variance and the allowable error level. It has to be noted that the amount of trees to be felled is higher if a certain amount (as a rule of thumb, 40 % of all sampled trees) is used for the verification of the derived regression functions.

Measurement of Dendrometric Variables To allow for the determination of required sample sizes to reliably estimate regression functions for a species or species group, many variables can be assessed without destroying the tree. Here, we provide a list of the parameters commonly used for such an assessment. Not all of the assessed attributes will serve as independent parameters in the regression functions. However, it is good practice to gather many attributes to have a large set of possible independent parameters. The actual measurement techniques and devices used are introduced in section “Nondestructive Assessment of Aboveground Biomass.” The most commonly assessed parameters are: • • • • • • • •

Diameter at 30 cm height (d0.3) (stump height) Diameter at breast height (d1.3, DBH) Upper-stem diameters at different heights (dx) Bark thickness Tree height (h) including bole length and stem height Crown parameters (length, diameter, shape, density) Tree species Layer/social position of the tree

Felling of the Trees After having assessed the dendrometric parameters, the selected trees will be cut down. It is important to proceed carefully in this step to avoid the mixture of wooden parts of the felled tree with woody debris already on the ground or resulting from damages to other trees or shrubs. To be accurate in this step, the usage of a heavy-duty plastic sheet in the area where the crown of the tree is foreseen to hit the ground and of ropes attached to the tree to avoid too much damage and guide the felling direction is advised. This will enable a sound conduction of the following steps. The tree should be felled as close to the ground as possible to allow for a complete aboveground biomass measurement.

850

D. Plugge et al.

Separation of Stem, Branches, and Twigs into Compartments The biomass of a single tree can either be assessed by measuring all parts of the compartments of the tree (see Fig. 2) and calculating their specific volume (full volume tally) or by weighing. If a volume tally is chosen, the different compartments should be marked to be identified easily in the following measurements. If weighing is chosen, the compartments have to be cut and separated from each other to facilitate the next steps. As shown in Fig. 2, the assignment of different parts of the tree into compartments facilitates the calculation of biomass. It is advised to divide the compartment stem, branches, and twigs into subclasses using a defined diameter as a threshold. Subclasses should be formed in a way that aligns with information needs, e.g., information on classes for commercial purposes. The amount of classes should be kept as low as possible to narrow the necessary measuring effort and to facilitate the combination of the different classes in the construction of biomass functions. The measurement of the threshold diameters should take place with a fixed caliper which is moved alongside the stem or branch until the threshold is reached. Deriving Biomass by Volume Tally via Randomized Branch Sampling As measuring the volume of each part of the tree is a rather labor-intensive work, it is advised to use a sampling approach to achieve a sound estimate of the total volume of the different classes. As long as dealing with monopodial trees for the

Fig. 2 Scheme of different biomass compartments of a tree (After: Young et al. 1964)

Measurement, Reporting, and Verifications Systems in Forest Assessment

851

different assortments of the stem, the volume can be calculated using Smalian’s or Newton’s formula: Smalian’s formula V ¼ Newton’s formula

V¼

gu þ gb L 2

gu þ 4gm þ gb L 6

(1) (2)

where gu, gm, and gb are the upper, middle, and bottom cross section of the assortment and L is the length of the assortment. For the class of the biggest branches, the above described formulas can also be applied given a regular, stem-like shape. In the case of bi- or multipodial trees and/or irregular shapes of the stems as well as for the smaller and irregularly formed branches and twigs, a volume assessment by measuring the length and the diameters with a tape is needed. This laborious work can be facilitated through randomized branch sampling. Randomized branch sampling (RBS) was first introduced into forestry by Valentine et al. (1984). In RBS there is no operational distinction made between the branch, stem, or twig. For RBS the tree or the part of the tree that is studied is divided into segments. This segmentation may lead to further subclasses in the classes already defined during the assignment. A segment is the part of a branch between two nodes. A path is the sequence of connected branch segments. At each node of branching, a decision has to be made which branch to follow in order to select the path through the entire tree. The continuation of the branch is selected with probability proportional to size (PPS), the “size” being equivalent to the estimated volume of the possible path continuation. The size can be estimated by the product of the length of the branch, l, and its diameter squared, d2, as this quantity is highly related to biomass (Fig. 3). By following this procedure, each selected branch has a defined probability of being selected from among the other branches at the node. By multiplying the subsequent probabilities, the conditional selection probability is calculated for each branch. The last segment of such a path has the lowest probability of selection. For more in-depth information on RBS, please refer to Valentine et al. (1984), Gregoire et al. (1995), Gregoire and Valentine (2004), and Ko¨hl et al. (2006). Volume figures have to be converted to weight figures by means of the corresponding wood density (see section “Treatment of Samples in the Laboratory”).

Deriving Biomass by Weighing The weighing of the different classes of a tree (stem, branches, twigs) can be very resource demanding. Especially when very large trees with huge trunks are cut, it would involve the usage of heavy-duty cranes, which in most of the cases in tropical forests is an unrealistic undertaking. Therefore, for the larger parts of the stem and for big branches, the abovementioned Smalian and Newton’s formulas should be applied to derive an estimate of the volume which will then be converted to weight using wood-specific density information.

852

D. Plugge et al.

Fig. 3 Segmentation for randomized branch sampling

For the other parts of the tree, the different classes need to be separated from each other and weighted individually. Weighing should take place as soon as possible after the felling of the tree to receive the fresh weight of the different classes. After weighing or the above described volume estimation, wood samples from each class have to be selected.

Selecting Wood Samples From each of the defined classes, samples have to be collected to be analyzed in the laboratory later on. In this step it is important to note that from the classes with smaller diameters, i.e., small branches or twigs, more samples need to be collected than for the compartments with large diameters. For the latter, a single disk, or for the trunk of large trees, only a radial part of a disk, needs to be taken. For the smaller assortments, it is advised to take several samples as in different parts of branches or twigs the moisture content may differ. For small twigs no disks should be taken but rather several specimens of a length of 5–10 cm. Each sample needs to be kept in a plastic bag and sealed properly. If feasible, it can be measured directly in the field before being taken to the laboratory. Each bag has to be marked unambiguously using, e.g., the sample site, tree number, class number, and sample number in that class. The single bags of one tree should be put together into a bigger plastic bag, again marked clearly to allow for identification of the tree in the laboratory. To avoid the loss of moisture, the samples need to be collected shortly after felling the tree, and the plastic bags have to be sealed tightly.

Measurement, Reporting, and Verifications Systems in Forest Assessment

853

Treatment of Samples in the Laboratory In the laboratory the fresh weight of the samples is determined with a precision scale. Samples should be weighed with the plastic bag to account for moisture accumulation in the bag. The weight of the plastic bag has to be subtracted from the fresh weight to receive the correct numbers. After recording the fresh weight, the samples will be oven-dried at a temperature of 101–105  C (Williamson and Wiemann 2010) to remove all moisture from the wood. This is done until the sample achieves a constant weight (24–72 h). Afterwards, the moisture content for each sample can be calculated through the relation of fresh to dry weight. As this is done for the different compartments of the tree, each class can be assigned a different factor to calculate its dry weight and the classes are summed up to obtain the dry weight of the total tree (Samuelsson et al. 2006). Besides estimating the dry weight/moisture content of the samples, they can be used to derive the wood density. Wood density is the relation of the mass of the sample divided by the volume at known moisture content, typically expressed as kg/m3 or g/cm3. Knowledge about species-specific wood densities or ideally about wood densities of different parts of a tree for a single species allows for a more accurate estimation of biomass of that specific tree. Building Biomass Functions With the knowledge obtained in the previous steps, biomass functions for single species or species groups can be derived. A functional relationship between the measured tree attributes and the biomass estimates obtained by the destructive sampling is constructed via regression analysis. The correlation between the single parameters is a quantitative measure to assess the strength of the functional relationships. Depending on the size of the biomass survey conducted, the biomass functions derived may contain one or several independent variables. As a rule of thumb, for larger surveys, e.g., national inventories, more variables should be applied to allow for explaining the higher variability of the population. Examples for such functions are given in section “Biomass Functions.”

Nondestructive Assessment of Aboveground Biomass As stated above, destructive assessments of biomass with actual felling of the trees can generally only be done on a small subsample of trees, if not prohibited at all. For most of the current surveys on biomass of tropical forests, e.g., for the REDD+ mechanism, time and monetary constraints prohibit such a laborious endeavor, and broad-scale inventories are carried out, measuring assessable tree parameters. The survey design and the distribution of sample plots as well as the statistical analysis are described in several textbooks (Hush et al. 2003; Ko¨hl et al. 2006) and in this book (link to Ko¨hl et al.) and will not be repeated here. We focus on the actual measurement in the field to derive the tree attributes needed as input into the biomass functions. This is not meant to be an all-inclusive list of attributes to be

854

D. Plugge et al.

assessed, as information needs always depend on the user of the survey results. We explain the measurement of those attributes already listed above in section “Measurement of Dendrometric Variables.”

Direct Assessment in the Field Most of the needed parameters can be assessed directly in the field using traditional measurement techniques. However, more advanced methods like terrestrial laser scanning (TLS) will also be introduced (see section “Estimating Biomass by Remote Sensing and Terrestrial Laser Scanning”). Before the assessment in the field, threshold values have to be defined, which indicate the range of a minimum or maximum value of an attribute to be assessed. Especially for assessing the regeneration of trees or for measuring shrubs and small trees below a threshold diameter or height it is advised to install subplots in or as satellites to the main sample plot of the survey to keep the workload in a good relation to the amount of biomass these classes represent. On the main plot the most important attributes to be assessed are: • Diameter at 30 cm height (d0.3) (stump height) This diameter measurement is used to derive information on the stem form close to ground level. At 30 cm height most bi- or multipodial trees already have split into their individual stems, and the diameter at this height allows for estimating the biomass of each of the stems. Beside this, many tropical trees have a conical shape that is expressed more intensively closer to the ground than at bigger heights. The d0.3 allows for constructing taper functions especially for the part of the stem between d0.3 and the DBH. The measurement is carried out over the bark using a tape. • Diameter at breast height (d1.3, DBH) The DBH is the most common and most easy to assess attribute in forest inventories. It is used for the grouping of trees into diameter classes and as an independent variable for calculating other parameters like basal area or height. Because of this, some important prerequisites need to be met when measuring DBH. Commonly, DBH is measured at 1.3 m height. In certain cases, DBH is measured at other heights (see Fig. 4). Whenever there is a deviation from 1.3 m height, it needs to be noted carefully. The DBH is either measured over the bark with a caliper or in case of irregular stem forms with a measuring tape. In the case of buttressed trees, a measurement as depicted for the case of a tree with aerial roots is taken (Fig. 4, top right). For the measurement of the buttressed part, please see Ngomanda et al. (2012). • Upper-stem diameters at different heights (dx) The upper-stem diameter is measured at one or several predefined heights along the stem. This measurement allows for the calculation of taper functions. For the measurement several methods exist including climbing, calipers attached to long poles, prisms, or electronic devices. The diameter is measured over the bark. • Bark thickness The thickness of the bark varies largely over different tree species but also alongside the stem. According to information needs, it may be of importance

Measurement, Reporting, and Verifications Systems in Forest Assessment

855

Fig. 4 Measurement of DBH in different conditions (Source: Adapted from Ko¨hl et al. 2006; Zingg 1988)

to subtract the bark from the total volume of a trunk to calculate the merchantable volume or to estimate the tree’s resistance to fire (see Ryan 1982). Bark thickness can be measured using a bark gauge penetrating the tree until the wood section is reached. • Tree height (h) Different tree heights may be of importance according to the information needs of stakeholders. For commercial purposes the merchantable height is of interest, displaying the distance between the ground and the end of the merchantable section of the stem, depending on the use, demarcated by the first big branch. If signs of damage along the bole are visible, these parts have to be subtracted. For biomass surveys in the scope of carbon assessment, the overall height, i.e., the distance between the top of the crown and the ground, is needed. To facilitate later volume and biomass calculations, the bole length is taken, i.e., the distance between the ground and the beginning of the crown. Height can be measured with several devices which all follow the trigonometric principle like Blume Leiss, Vertex, or Criterion. The choice of the instrument depends on the stand density and visibility of the crown. As recording tree height in closed natural forests is more labor-intensive than measuring DBH and good functional relationships between DBH and height exist, typically, the heights of only some of the trees (5–10) at one sample plot are recorded. The amount of trees depends on the forest structure (amount of layers) and therefore variability of heights. • Crown parameters (length, diameter, shape, density) The crown length is recorded during the height measurement. It is the total height subtracted by the bole height. The crown diameter is measured with at least two measurements as crown forms in natural tropical forests seldom resemble a circle. The measurement should include the largest and the smallest diameter. Crown length and crown diameter together with the crown shape are used to calculate the

856

D. Plugge et al.

crown volume, necessary for deriving the crown biomass. Crown shape should be oriented on typical geometric forms (circle, cone, rectangle, or umbrella), and it should be recorded whether the crown form is uniformly shaped or shifted to one side of the tree. Biomass finally depends on the space of the crown volume that is filled with branches and leaves. Therefore, the crown density has to be recorded. This can either be done by the estimation of percentages of the crown filled with biomass or using different classes (e.g., dense, medium, light). • Tree species The tree species must be recorded to apply species-specific biomass functions or to assort a tree into a species group for which biomass functions exist. In general, the botanical names of the species should be recorded. However, for many tropical tree species, only local names can be recorded in the field as hiring botanists for a large-scale forest inventory may be very costly. Whenever there is uncertainty on the botanical name, specimen of leaves and fruits as well as pictures should be taken for later identification. For many tropical countries good compendia exist that refer commonly used local names to botanical species names. • Layer/social position of the tree Biomass changes with the layer (e.g., upper, middle, lower) or social position (e.g., predominant, dominant, codominant, dominated, suppressed) a tree belongs to. This parameter should be recorded to allow for a further grouping of the trees for applying biomass functions. The above described variables display a minimum amount of information that should be recorded for a terrestrial biomass assessment, either destructive or nondestructive. Depending on information needs of user groups, a plethora of other attributes and variables may be required. In the next chapter we are going to introduce some new techniques for biomass assessments in tropical forests that have not been applied on a wide-spread basis so far.

Estimating Biomass by Remote Sensing and Terrestrial Laser Scanning Besides the above described terrestrial measurement methods for individual trees on sampling plots, remote sensing technologies can be used to estimate aboveground biomass of larger areas. As biomass cannot be captured directly by any sensor system, remote sensing systems produce auxiliary information on the spatial biomass distribution. Spatial biomass distribution varies from individual trees over forest stands to landscape level as well as between forest types with specific biomass distribution characteristics. Furthermore, forest management activities also influence the spatial biomass distribution. There are two general types of remote systems: passive and active. Passive systems use the reflection of sunlight, whereas active systems, for example, radar or laser, capture the reflection of self-emitted signals. Further, remote sensing sensors differ in their spectral bands, spatial resolution, and repetition rate as well as in their physical principals. For the assessment of forest biomass, optical sensors with different numbers of spectral bands, radar sensors, and laser systems are used. These systems can be installed either on satellites (spaceborne) or on planes

Measurement, Reporting, and Verifications Systems in Forest Assessment

857

(airborne). Recently, small drones (unmanned aerial vehicles) have become popular for the assessment of small forest patches. Direct and indirect parameters can be extracted from remote sensing data. These can be used directly for biomass estimation or combined with terrestrial measurements. Direct parameters link the signal response to biomass by methods such as regression analysis, classification, or k-NN models. Indirect parameters can be derived from the spatial structure or combination of various remote sensing data sets. These are, for example, forest height, forest closure, forest type, crown size, or 3D characteristics. Biomass is then estimated by models that are based on these indirect parameters (Koch 2010; Saatchi et al. 2011). In the following a short overview on the different sensor types and their potential for biomass assessment is given. Further details on the use and applications of remote sensing systems for tropical forest are given in chapter (link to Baldauf and Jimenez). • Optical systems Optical systems vary in resolution. For example, spaceborne systems vary from coarse, (e.g., MODIS (Moderate-Resolution Imaging Spectroradiometer) with 500 m ground resolution in 29 spectral bands) to medium (e.g., Landsat with 30 m ground resolution in 11 spectral bands) to high (e.g., QuickBird, 0.65–2.62 m ground resolution in 4 spectral bands) resolution. With the resolution the amount of data increases and the repetition rate, i.e., the time interval between two identical flights over the same area, decreases. Airborne aerial pictures have usually very high resolution. All these systems assess the shortwave reflections at the upper canopy and hence depend on the position of the sun and weather conditions. With increasing spatial resolution, the influence of forest structure increases. Overlapping high-resolution pictures, captured, e.g., by low-altitude drones, can be used to derive 3D canopy surface models. Based on these the size of individual tree crowns can be used as an indicator for biomass (Cutler et al. 2012; Lefsky 2010). • Radar systems Radar systems actively use long-wave radio signals which – depending on the wave length – penetrate the canopy to a different extent. Examples for spaceborne systems are the ALOS, RADARSAT and Terra SAR satellites. In radar systems, up to a wavelength-specific point of saturation, biomass can be directly estimated. Radar systems are independent from weather and sun position. Depending on the configuration, coarse to fine resolutions – up to 1 m for small sites – are available. Repeated observations or tandem constellations like TerraSAR X TanDEM allow for the assessment of detailed canopy surface models (Englhart et al. 2011; Mitchard et al. 2012). • Airborne laser systems (ALSs) ALSs are used to assess the 3D structure of forests, while the reflection intensity is of less importance. Each emitted signal is captured as a range of reflections within the canopy, starting from the first reflection at the top of the canopy. Even in dense tropical forests, some of the signals reach the forest floor. Hence, these systems assess the forest density and the height, which can lead to sound biomass estimations (Asner et al. 2012; Dubayah et al. 2010; Maltamo et al. 2014).

858

D. Plugge et al.

• Terrestrial laser scanners (TLSs) TLSs also use laser impulses to measure the 3D structure of tree components. In general, they are operated from ground-based sample plots underneath the canopy. In contrast to airborne laser systems, TLS systems measure per impulse one distance to reflecting objects and no reflection profiles. Hence, they assess the shapes of all tree components visible from their position. With the help of analyzing software, the volume of individual trees or tree components can be estimated and subsequently be linked to biomass (Calders et al. 2015; Kankare et al. 2013).

Biomass Functions In the previous chapters, we have explained (a) how to gather data for the construction of own biomass functions via destructive sampling and (b) how to gather attributes that are commonly needed as input for biomass functions. In this chapter, we want to introduce different types of functions and give an overview of functions that are readily available and widely applied in tropical areas. In general, functions should be region-specific to best fit to the forest of a certain area. Therefore, no universal rule on how to construct the function and what attributes to include exists. However, it is obvious that DBH as the parameter that is the easiest to assess and that correlates strongly with many other tree attributes and therefore also with biomass most likely will be included in such a function. Local functions may rely on DBH as the only variable if the forest stands are rather homogeneous. Regional functions should generally involve more independent variables besides DBH such as height, wood density, or crown attributes to account for higher variability in the forest stands of a region. Following this logic, broad-scale functions should make use of more independent variables to be able to explain an as large part of the variability of the forest population under research as possible. Brown (1997) introduced two different types of functions. The first uses existing volume estimates, whereas the second is in line with the above chapters using treespecific attributes as direct input into biomass regression functions. Whenever information on some parts of the aboveground biomass is lacking, e.g., only the volume of a tree is known but no information on the crown is available, so-called biomass expansion factors (BEFs) are used to relate total aboveground biomass to volume. The IPCC (2006) lists a set of so-called biomass conversion and expansion factors (BCEFs), which are applicable to relating merchantable growing stock volume, net annual increment, or the removal of wood and fuelwood to aboveground biomass. Chave et al. (2014) updated the work of Chave et al. (2005) by collecting data from surveys that have constructed biomass regression functions based on destructive sampling of a total of 4,004 trees in 58 study sites across the tropics. In the supplementary material to their report, Chave et al. (2014) list and describe the different study sites involved. FAO, together with CIRAD (French Agricultural Research Center for International Development) and DIBAF (Department for

Measurement, Reporting, and Verifications Systems in Forest Assessment

859

Innovation in Biological, Agrofood and Forest systems, Tuscia University), is offering an open online database, which hosts a plethora of regional allometric functions (globallametree.org). We invite the reader to refer to this large database and search for functions best fitting to their specific areas. If no functions for an area can be found, Chave et al. (2014) provide a pantropical function for biomass either using DBH, height, and wood density or DBH, wood density, and a measure of environmental stress consisting of several variables if tree height is not available. Wood density plays an important role when predicting biomass with regression functions. For most of the commercially used timber from tropical forests, wood density is known. For other lesser-known species, the IPCC provides a rather large compendium of wood densities for tropical Africa, America, and Asia (IPCC 2006) Furthermore, ICRAF (International Centre for Research in Agroforestry) provides an extensive online database of wood densities (http://db.worldagroforestry.org/ wd) that can be assessed freely. If, in the scope of the biomass assessments, estimates have to be translated into carbon, in general a conversion factor of 0.47 from biomass to carbon is applied (IPCC 2006). This conversion factor may vary in a small range (0.44–0.49) but basically holds true regardless of wood-specific density, tree species, or forest type as it relates to the estimate of biomass in tonnes of dry matter.

Belowground Biomass and Soil Organic Carbon Belowground biomass (BGB, root biomass) and organic carbon in mineral soils (SOC) constitute the two subterranean carbon pools. Land-use and management activities typically have a considerable impact on organic carbon stocks of mineral soils, which have relatively low organic matter compared to organic soils (e.g., peat and muck soils), but predominate in most ecosystems except wetlands (IPCC 2006). Root biomass plays an important role in the carbon cycle as roots transfer considerable amounts of carbon to the ground (Ravindranath and Ostwald 2008). However, our knowledge of biomass allocation to roots lags behind that of its aboveground counterpart (Cairns et al. 1997). Root biomass is a major source of uncertainties in large-scale biomass estimation (Wang et al. 2008) and, therefore, has received an increased focus over the last three decades (e.g., Albuquerque et al. 2015; Cairns et al. 1997; Castellanos et al. 1991; Costa et al. 2014; Fittkau and Klinge 1973; Li et al. 2003; Mokany et al. 2006). In the following sections, we discuss root biomass and SOC assessment methods. As suggested in section “Estimating Biomass Changes Over Time,” the estimation of changes in belowground carbon stocks over a predetermined period of time is out of the scope of this chapter.

Root Biomass Assessment Techniques Estimates of aboveground biomass based on well-established methods are relatively abundant while estimates of root biomass based on standard methods are

860

D. Plugge et al.

much less common (Cairns et al. 1997). The prerequisites for forest inventories to reliably estimate the AGB have been introduced in the previous subchapters; however, to estimate root biomass in a direct way, other means must be sought (Cairns et al. 1997). Consequently, the data available on root biomass are limited (MacDicken 1997). More often, indirect (or nondestructive) methods are used to estimate the pool as a proportion or function of the AGB. This section reviews the current practice for both direct methods (e.g., excavation of roots and core sampling) and indirect methods (root to shoot ratios and allometric equations) used for estimating root biomass. Both methods involve the root biomass assessment up to 30 cm depth. Fine roots of less than 2 mm diameter are often excluded while estimating BGB; the roots below this diameter threshold are considered as soil organic matter (IPCC 2006).

Direct Methods Excavation of roots and core sampling are the direct or destructive methods used in estimating root biomass. The greatest proportion of the total root mass is confined to the upper 30 cm of the soil (Ravindranath and Ostwald 2008). For estimating the root biomass below 30 cm, the monolith method is generally used. The procedure involves excavating a block of soil from a sample plot, separating the roots from the soil, and weighing them (Ravindranath and Ostwald 2008). For the detailed procedures of the direct methods, see Castellanos et al. (1991), Costa et al. (2014), Fittkau and Klinge (1973), MacDicken (1997), Picard et al. (2012), and Ravindranath and Ostwald (2008).

Excavation of Roots This method involves excavating the roots of the trees by digging (or trenching). It is mostly used to estimate the biomass of individual trees. The location of a pit is determined after establishing a Voronoi diagram around the selected tree. Drawing of a Voronoi diagram involves four steps: (i) drawing the segments that link the selected tree to each of its neighbors, (ii) drawing the bisectors for each segment, (iii) connecting the bisectors one to another to construct a space around the tree, and (iv) dividing the space into joined triangles (Fig. 5) (Picard et al. 2012). For medium-sized roots, partial excavation of the Voronoi space, e.g., excavation of an area covered by a triangle in the space, might be enough. Larger roots require the total excavation of the Voronoi space. The size of a pit varies according to the precision needed for the root biomass estimate. The measurement of tree attributes and felling of the selected tree follow the procedures described in section “Destructive Assessment of Aboveground Biomass.” Once the pit is dug, the following procedures are generally applied: separation of the tree roots and non-tree roots, separation of the soil from the roots by washing and sieving, storage of the root biomass of each tree in a separate cloth bag, measurement of the fresh root weights, and collection of a sample (c. 0.5 kg) for each of the tree and non-tree biomass for oven-drying. To determine the ovendry

Measurement, Reporting, and Verifications Systems in Forest Assessment

861

Fig. 5 Steps involved to construct a Voronoi diagram and its subdivisions around a selected tree (Source: Picard et al. 2012)

weight of the biomass, the samples are kept in ovens and dried to a constant weight. The ovendry weight is then extrapolated at the plot level, to hectare, and finally to the stratum area (see section “Upscaling Procedures”).

Core Sampling The establishment of a Voronoi diagram is described in the previous section. Soil cores to a minimum depth of 30 cm are collected, inserting the core samplers (e.g., AMS soil core samplers) into the soil surface. The number of cores obtained depends on the precision required. At a slope, care must be taken so that the sampler is inserted perpendicular to the land surface. The root biomass is extracted by sieving and washing; the extracted roots are weighted and brought to laboratory for oven-drying to determine the dry and constant weight. The dry weight of the roots for the volume of the core sampler is estimated and extrapolated further to estimate the root biomass at plot and hectare level.

Indirect Methods Root biomass accumulation is linked to the dynamics of aboveground biomass (Ravindranath and Ostwald 2008). Root to shoot ratios are routinely used to partition plant biomass into aboveground and root component (Cairns et al. 1997). Allometric regression models have been developed and presented in a myriad of literatures, e.g., by Cairns et al. (1997) and Li et al. (2003).

862

D. Plugge et al.

Root to Shoot Ratios Default values of BGB to AGB ratios for entire forest domains (e.g., tropical, subtropical, temperate, and boreal) can be used to estimate root biomass growth (Tier 1 approach of IPCC). Cairns et al. (1997), Li et al. (2003), MacDicken (1997), and Mokany et al. (2006) provide the ratios as default values. For example, Mokany et al. (2006) recommend the R of 0.235 (low, 0.220; high, 0.327; SE = 0.011; n = 10) for the tropical/subtropical moist forest/plantation (AGB > 125 t/ha). Likewise, Cairns et al. (1997) suggest the R of 0.24 (SD = 0.14, n = 39), 0.26 (SD = 0.07, n = 73), and 0.27 (SD = 0.10, n = 53) for tropical, temperate, and boreal latitudinal zones, respectively. The IPCC (2006) compiles a summary of the default values (Table 3). Table 3 Ratio of belowground biomass to aboveground biomass (R) [tonnes (t) root dry matter (d. m.): (t) shoot d. m.] (From IPCC 2006) Ecological zone Tropical rainforest Tropical moist deciduous forest Subtropical humid forest Tropical moist deciduous forest Subtropical humid forest Tropical/ Subtropical dry forest Tropical/ Subtropical dry forest Temperate oceanic forest Temperate continental forest Temperate mountain systems

Boreal coniferous forest Boreal tundra woodland Boreal mountain systems

Aboveground biomass (AGB) t ha1

R 0.37

125

0.24 (0.22–0.33)

20

0.28 (0.27–0.28)

Conifers 150 Quercus spp. >70 Eucalyptus spp. 150 Other broadleaf 150 t < 80 %). For tropical tree species a wide range of different mating systems have been observed, including different levels of outcrossing, selfing, and mating among relatives. Also asexual reproduction can occur, e.g., by apomixes as detected in Jatropha curcas (Bressan et al. 2013). Initially, it was assumed that tropical tree species were self-compatible and produced seeds primarily by selfing (Federov 1966). The theory was that the low densities of individual trees per species would then lead to low levels of genetic diversity and high inbreeding. Later studies of Bawa (1974), Ashton (1976), and Bawa et al. (1985) using controlled crossings and other studies using gene markers (Table 1) showed that this theory was wrong. Apparently there is a predominance for outcrossing in tropical tree species, and there is a strong selection against inbreeds stabilizing high levels of genetic diversity. Bawa (1974) analyzed the breeding system of tree species in the tropical semi-deciduous forest of Costa Rica. With controlled pollinations and direct observations of the floral biology, he found that from 130 tree species, 68 % were hermaphrodites, 10 % were monoecious, and 22 % were dioecious species. Bawa et al. (1985) observed that 86 % of the hermaphrodite and monoecious species showed self-incompatibility mechanism, which successfully eliminates or avoids selfing in tropical tree species. These mechanisms explain the high rates of outcrossing ðtm ¼ 0:88Þ detected in many tropical tree species (Table 5). The mating system varies among different tree species but also among different populations or individuals of the same tree species (Table 5). Examples for variation among different individuals of the same population are reported for Platypodium elegans (Hufford and Hamrick 2003) and for Magnolia stellata (Tamaki et al. 2009). Temporal variation of outcrossing rates of the same species and population at different reproductive events have also been observed for Platypodium elegans (Hufford and Hamrick 2003), and even variations of outcrossing rates among and within fruits of the same individuals are published (e.g., Acacia melanoxylon, Muona et al. 1991; Eucalyptus rameliana, Sampson 1998; Magnolia stellata, Tamaki et al. 2009). In animal-pollinated tropical tree species, mating systems have been shown to be affected by factors such as the density of reproductive trees (Murawski and Hamrick 1991) and the pollinator behavior (Hirao et al. 2006), by forest fragmentation and the level of spatial isolation of trees (Cascante et al. 2002; Fuchs et al. 2003; Lowe et al. 2005; Aguilar et al. 2008; Feres et al. 2012), and by logging activities (Doligez and Joly 1997; Lacerda et al. 2008; Carneiro et al. 2011). Both the temporal isolation (no overlapping in flowering period of trees) and spatial isolation of flowering (low density of flowering trees) are correlated to high rates of self-pollination for some tree species (Murawski and Hamrick 1991; Dick et al. 2003; Naito et al. 2008; Moraes and Sebbenn 2011). Forest fragmentation and selective logging reduce the density of reproductive individuals and may affect the behavior of pollinators.

Tree species Araucaria angustifolia Araucaria angustifolia Araucaria angustifolia Araucaria angustifolia Araucaria angustifolia Araucaria angustifolia Araucaria angustifolia Araucaria angustifolia Bagassa guianensis Carapa guianensis Carapa guianensis Carapa guianensis Cariniana legalis Cariniana legalis Cariniana legalis Copaifera langsdorffii Copaifera langsdorffii Eucalyptus rameliana Hymenaea courbaril Guaiacum sanctum Guaiacum sanctum Hymenaea courbaril

Type of forest Continuous Exploited Continuous Continuous Continuous Planted Planted Continuous Continuous Continuous Logged Continuous Fragmented Fragmented Fragmented Fragmented Board Continuous Continuous Continuous Fragmented Logged

Density (trees/ha) 201 166 >57 >57 7.7 >200 >200 >50 0.14 2.5 2.0 25.7 0.90 10 %) may germinate readily without pretreatment, while dried and stored seeds require pretreatment (Duguma et al. 1988). This makes it difficult to prescribe a “standard” pretreatment procedure. While mechanical pretreatment is relatively “safe” as long as only the seed coat is influenced and as long as the micropyle site is avoided to avoid damage to the sensitive radicle, bulk pretreatment with hot water and acid (usually sulfuric acid) implies a risk of overtreatment with consequent loss of viability (if high temperature or acid reach the embryo). Since any seed lot contains a variation of resistance to imbibition (seed coat hardness), bulk pretreatment aims at “average” which often results in some seed being insufficiently pretreated (remaining dormant) and others being overtreated (killed/damaged by pretreatment). Chemical inhibitors often occur in fleshy fruits. The inhibitors prevent germination of seeds surrounded by a watery environment in the fruits (Robertson et al. 2006). The fruit inhibitors are mostly removed with the fleshy part during seed extraction. However, remnants of inhibitors may still occur in the seed coats and endocarps of drupes and cause germination delay. Most inhibitors are water soluble and are most effectively removed by through rinsing in running water, e.g., in connection with wet extraction. However, sometimes small fleshy fruits are dried without extraction (e.g., Breynia, Macaranga, some Diospyros species) and some dry fruits (e.g., dry drupes of teak (Tectona grandis)) also contain inhibitors. In these cases rinsing is applied as a pretreatment just before sowing. The fruits can be put in net bags and pretreated in natural running water, e.g., in streams. Photo-inhibition (photo-dormancy – photoblastic seed) is controlled by phytochrome pigment in seed coats or pericarps. The two forms of the pigment can be reversibly converted to the other form (Pfr $ Pr) by exposure to dark, filtered (under canopy), or white light. Photo-dormancy occurs frequently in pioneer trees, where the purpose is to prevent germination under dark (e.g., buried in soil) or shaded conditions (Toole 1973; Mayer and Poljakoff-Mayber 1982; Vazquez-Yanes and Orozco-Segovia 1993; Vazquez-Yanes 1982). The phenomenon has large ecological importance but rarely imposes problems in nurseries or other types of seed sowing, since in practice it is overcome by exposing imbibed seed to light conditions during germination. Covering seed with a thin layer of soil during germination does normally not restrict germination. Temperature inhibition (thermo-dormancy) is known in three forms where the triggering event is cold, hot, or fluctuating temperature. Thermo-dormancy is well known from temperate trees, where (imbibed) seeds (Pinus, Quercus, Fagus) need a cold period prior to germination (Mayer and Poljakoff-Mayber 1982). The same phenomenon has also been found in some highland tropical species, e.g., eucalypts (Bonner et al. 1994; Turnbull and Doran 1987). Another type of thermo-dormancy is encountered in species from fire-prone areas, where a brief exposure to high temperature (in practice performed by a light grass fire on seedbeds) prior to

952

L. Schmidt

watering triggers germination. Eventually, some pioneers benefit from fluctuating temperatures during germination, apparently an adaptation to sense “gaps” as contrary to the more uniform forest climate under a closed canopy (VazquezYanes and Orozco-Segovia 1993). Immature/underdeveloped embryos occur in a number of species. Seeds are here simply dispersed before the embryo has reached full development. The development stage of the embryo at dispersal stage ranges from an undifferentiated zygote cell mass (e.g., Ginkgo) to an almost fully mature embryo. “Pretreatment” is here a prolonged afterripening under warm moist conditions. The phenomenon is known both in tropical and temperate species, e.g., Taxus baccata, Ginkgo biloba, Ilex opaca, and Allanblackia stuhlmannii (Mathew et al. 2009; Phartyal and Thapliyal 2005). Several species have two or more dormancy types, each of which must be broken for germination to occur. Duple dormancy occurs, e.g., in fleshy fruits with hard endocarps (e.g., neem, Azadirachta indica) and triple dormancy in yew (Taxus baccata) including an underdeveloped embryo, a fleshy aril (with inhibitors), and thermo-dormancy. Temperature, light, and inhibitors interfere with the physiological mechanism of germination, which is controlled by hormones. Therefore, germination can often be triggered by applying germination hormones (gibberellins, GA) to imbibed seed. The hormones help overcome some dormancy types and speed up germination (Bhattacharyya et al. 1991; Bewley 1997).

Seed-Borne Pests Being concentrated packages of high nutritional material, seeds are highly attractive to consumers such as insects and fungi. Pests are macroorganisms, mostly insects that physically consume the entire or part of the seeds. Pathogens are disease-causing microorganisms, which infest (from the outside) or infect (from the inside) seed (Neergaard 1979). Most seeds carry on their surface a range of pathogens that have the potential to become destructive (Mohanan and Sharma 1991; Mohanan et al. 2005). However, pests and pathogens are not necessarily fatal for seeds. It depends whether essential parts of the embryo is affected. Even seeds with part of the cotyledons consumed by seed predators may germinate and develop into normal seedlings. Infection or infestation by seed-borne pests implies two problems: (1) damage to seed and plants during germination with possible contamination to other seed or plants of the seedlot and (2) risk of transfer of a pest or pathogen into a pest- and disease-free area. The first problem is dealt with locally either by treatment of seed to eliminate infecting organism or overcoming the problem after germination. It should be noticed that most seeds carry fungal spores and other pathogens, but they often do little damage; seeds have an inert protection, and if seeds germinate fast, they usually overcome possible fungal infection. However, under conditions favorable to fungi and less favorable to germination and seedling development, seed-

Genetics and Forest Seed Handling

953

Fig. 12 Left: Germination test for Dalbergia spp. in tray in lab. Seeds were scarified by surface burning. Fungal attack (Penicillium sp.) initiated at the necrotic site created by the burning treatment. Fungal attack was only prevalent during cold periods and slow germination, while the seeds overcame the fungi when germinated at optimal conditions. Right: Acacia tortilis attacked by bruchid beetles (dark areas pointed with white lines). The beetles survive at very low moisture content and larvae and pupae are able to survive during storage

borne fungi can develop rapidly (Fig. 12). A serious problem is the so-called damping-off disease, which is a fungal attack on germinating seeds and young seedlings typically at the root-stem transition. Surface sterilization, for example, in natrium hypochlorite, is often used under lab conditions. Pretreatment with, e.g., sulfuric acid in hard seed will eliminate any fungal spore and bacteria on the seed coat. However, physical scarification and chemical scarification leave the seed coat ruptured and likely to get reinfested easily. Pretreatment is also therefore usually carried out only just before sowing the seed. The second problem is a special concern for exotics since many introduced organisms do not have natural controlling enemies at their new site. A pest or a pathogen can thus occasionally spread very fast in exotic plantations (Wingfield et al. 2002). Pathogens that use seeds as a vehicle without damaging seeds directly go under the name “seed transmitted.” These include fungi such as rust that are harmless to seeds and seedlings but attacking flowers and buds. Dry and cool conditions prevent development of pest and pathogens, but fungal spores and some insect pupae and eggs may survive adverse conditions only to start activities again when seeds are transferred to germination conditions. Both insecticides and fungicides can be effective pest and disease controls in seedlots and sometimes unavoidable, especially during international transfer. Some key concerns have been possible phytotoxic effect (damage to plants) and risk to persons working with them. If pesticide treatments are deemed necessary, health precautions should be followed and the specific remedy documented on shipping documents.

References Ali SIS (2006) Manual for establishment of seed production areas in Dipterocarp forests in Peninsular Malaysia. In: Malaysia – International Tropical Timber Organisation Joint Project: PD 185/91 Rev.2 (F) – Phase II. Forestry Department Peninsular Malaysia

954

L. Schmidt

Andjic V, Dell B, Barber P, Hardy G, Wingfield M, Burgess T (2011) Plants for planting; indirect evidence for the movement of a serious forest pathogen, Teratosphaeria destructans, in Asia. Eur J Plant Pathol 131:49–58 Ballard LAT (1973) Physical barriers to germination. Seed Sci Technol 1:285–303 Barner H, Willan RL (1985) Seed collection units 1: seed zones, Technical note, no 16. Danida Forest Seed Centre, Humlebaek Barner H, Olesen K, Wellendorf H (1988) Classification and selection of seed sources, Lecture note, no B-1. DANIDA Forest Seed Centre, Humlebaek Barnes RD (1995) The Breeding seedling seed orchard in multiple population breeding strategy. Silvae Genet 44(2–3):81–88 Barot S, Mitja D, Miranda I, Meija GD, Grimaldi M (2005) Reproductive plasticity in an Amazonian palm. Evol Ecol Res 7:1–15 Bawa KS, Webb JT (1984) Flower, fruit and seed abortion in tropical forest trees: implications for the evolution of paternal and maternal reproductive patterns. Am J Bot 71(5):736–751 Bawa KS, Ashton PS, Nor SM (1990) Reproductive ecology of tropical forest plants: management issues, Chapter 1. In: Bawa SK, Hadley M (eds) Reproductive ecology of tropical forest plants, vol 7, Man and the biosphere series. UNESCO, Paris Berjak P, Pammenter NW (2002) Orthodox and recalcitrant seed. In: Vozzo J (ed) Tropical tree seed manual, vol 721, Agriculture handbook. USDA Forest Service, Washington, DC, pp 137–147 Berjak P, Pammenter NW (2003) Understanding and handling desiccation-sensitive seeds, Chapter 22. In: Smith RD, Dickie JB, Linington SH, Pritchard HW, Probert RJ (eds) Seed conservation: turning science into practice. Royal Botanic Gardens, Kew, Richmond, pp 415–430 Bewley JD (1997) Seed germination and dormancy. Plant Cell 9:1055–1066 Bhattacharyya AK, Lahiri AK, Basu RN (1991) Improvement of germinability of Eucalyptus species by pregermination treatments. Indian Forester 117(8):661–663 Boland DJ, Brooker MIH, Turnbull JW, Kleinig DA (1980) Eucalyptus seed. CSIRO, Canberra Bold HC, Alexopoulos C, Delevoras T (1980) Morphology of plants and fungi. Harper & Row, New York, 819 pp Bonner FT, Vozzo JA, Elam WW, Land SB Jr (1994) Tree seed technology training course. U.S. Dept. of Agriculture, Forest Service, Southern Forest Experiment Station, New Orleans Booth TH (1996) Predicting plant growth: where will it grow? How well will it grow? In: Proceedings of the third international conference/workshop on integrating GIS and environmental modeling. National Center for Geographic Information and Analysis, Santa Fe/Santa Barbara, 21–25 Jan 1996 Booth TH (1998) A broadscale land evaluation program to assess the potential for growing particular trees in Africa. Agrofor Syst 40(2):125–138 Booth TH, Searle SD, Boland DJ (1989) Bioclimatic analysis to assist provenance selection for trials. New For 3:225–234 Boshier DH (2000) Mating systems, Chapter 5. In: Young A, Boshier DH, Boyle TJ (eds) Forest conservation genetics: principles and practice. CABI Publishing, Wallingford, pp 63–79 Burger D (1972) Seedlings of some tropical trees and shrubs mainly of South East Asia. Centre for Agricultural Publishing and Documentation, Wageningen Charles-Dominique P (1986) Inter-relations between frugivorous vertebrates and pioneer plants: Cecropia, birds and bats in French Guiana. In: Estrada A, Fleming TH (eds) Frugivores and seed dispersal. Dr. W. Junk Publisher, Dordrect, pp 119–135 Coe M, Coe C (1987) Large herbivores, acacia trees and bruchid beetles. S Afr J Sci 83:624–635 Critchfield WB, Little EL Jr (1966) Geographic distribution of the pines of the world, Miscellaneous publication, 991. U.S. Dept. of Agriculture, Forest Service, Washington, DC Dawson I, Were J (1998) Multiplication, that’s the name of the game: guidelines for seed production in agroforestry trees. Agrofor Today 10(4):19–22, World Agroforestry Centre (ICRAF), Oct–Dec 1998

Genetics and Forest Seed Handling

955

de Vogel EF (1980) Seedlings of cotyledons. Structure, development, types. Pudoc, Wageningen, 466 pp Delaporta SL, Calderon-Urrea A (1993) Sex determination in flowering plants. Plant Cell 5:1241–1251 Dhakal LP, Lillesø JPB, Kjær ED, Jha PK, Aryal HL (2005) Seed sources of agroforestry trees in a farmland context – a guide to tree seed source establishment in Nepal, Development and environment series, 1-2005. Forest & Landscape, Hørsholm Duguma B, Kang BT, Okali DUU (1988) Factors affecting germination of Leucaena (Leucaena leucocephala (Lam.) de Wit.) seed. Seed Sci Technol 16:489–500 Eldridge KG, Davidson J, Harwood CE, van Wyk G (1993) Eucalypt domestication and breeding. Clarendon, Oxford, xix + 288 pp Elliott S, Blakesley D, Hardwich K (2013) Restoring tropical forests: a practical guide. Forest Restoration Research Unit, Chiang Mai University Espinosa-Garcia FJ (1996) Review on alleopathy of Eucalyptus. L’Herit. Boletin de la Sociedad Botanica de Mexico, 58:55–74 FAO, DFSC, IPGRI (2001) Forest genetic resources conservation and management, vol 2: in managed natural forests and protected areas (in situ). International Plant Genetic Resources Institute, Rome Faulkner DS (1989) Introduction to quantitative genetics. Longman Scientific & Technical, New York, 438 pp Finkeldey R, Hattemer HH (2007) Tropical forest genetics. Springer, Berlin, 315 pp Gao J, Queenborough SA, Chai JP (2012) Flowering sex ratio and spatial distribution of dioecious trees in a southeast Asian seasonal tropical forest. J Trop For Sci 24(4):517–527 Garwood NC (1996) Functional morphology of tropical tree seedlings, Chapter 3. In: Swaine MD (ed) The ecology of tropical forest tree seedlings, Man and the biosphere series. UNESCO and The Parthenon Publ. Group, Paris, pp 59–129 Gomez JM (2004) Bigger is not always better: conflicting selective pressure on seed size in Quercus ilex. Evolution 58(1):71–80 Haines R, Nikles G (1987) Seed production in Araucaria cunninghamii – the influence of biological features of the species. Aust For 50(4):224–230 Hansen CP, Kjaer ED (1999) Appropriate planting material in tree plantings: opportunities and critical factors. In: International expert meeting on the role of planted forests for sustainable forest development, Santiago, 6–9 Apr 1999 Harwood CE (1992) Natural distribution and ecology of Grevillea robusta. In: Harwood CE (ed) Grevillea robusta in agroforestry and forestry. Proceedings of the international workshop, 21–28. ICRAF, Nairobi Holl KD, Aide TM (2011) When and where to actively restore ecosystems? For Ecol Manag 261 (2011):1558–1563 Hughes CE (1998) Leucaena – a genetic resource handbook, vol 37, Tropical forestry paper. Oxford Forestry Institute, Oxford Hyde EOC (1954) The function of the hilum in some Papilionaceae in relation to ripening of the seed and the permeability of the testa. Ann Bot 18:241–256 Janzen D (1978) Seedling patterns of tropical trees, Chapter 4. In: Tomlinson PB, Zimmermann MH (eds) Tropical trees as living systems. Cambridge University Press, Cambridge, pp 83–128 Jensen NF (1988) Plant breeding methodology. Wiley, New York, 676 pp Jepma CJ, Munasingh M (1998) Climate change policy: facts, issues and analyses. Cambridge University Press, Cambridge Kannan CS, Sudhakara K, Augustine A, Ashokan PK (1996) Seed dormancy and pretreatments to enhance germination in selected Albizia species. J Trop For Sci 8:369–380 Kleinschmit J, Khurana DK, Gerhold HD, Libby WJ (1993) Past, present and anticipated applications of clonal forestry. In: Ahuja MR, Libby WJ (eds) Clonal forestry II: conservation and application. Springer, Berlin, pp 9–41

956

L. Schmidt

Krauss SL, Sinclair EA, Brussel JD, Hobbs RJ (2013) An ecological genetic delineation of local seed-source provenance for ecological restoration. Ecol Evol 3(7):2138–2149 Leng P, Yamamura H (2006) Fruit set and embryo rescue in crosses using parthenocarpic ‘Mopanshi’ persimmon. Sci Hortic 107:332–336, Elsevier Lillesoe JBL, Graudal L, Moestrup S, Kjær ED, Kindt R, Mbora A, Dawson I, Muriuki J, Ræbild R, Jamnadass R (2011) Innovation in input supply systems in smallholder agroforestry: seed sources, supply chains and support systems. Agrofor Syst 83:347–359 Mathew MM, Munjuga MR, Ndangalasi HJ, Cordeiro NJ (2009) Aspects of the floral and fruit biology of Allanblackia stuhlmannii (Clusiaceae), an endemic Tanzanian tree. J East Afr Nat Hist 98:79–93 Mayer AM, Poljakoff-Mayber A (1982) The germination of seeds. Pergamon, Oxford Mbora A, Barnekov Lillesø JP, Schmidt L, Angaine P, Meso M, Omondi W, Ahenda J, Mutua NA, Orwa C, Jamnadass R (2009) Tree seed source re-classification manual. World Agroforestry Centre, Nairobi McKey D (1975) The ecology of co-evolved seed dispersal systems. In: Gilbert LE, Raven PH (eds) Coevolution of animals and plants. University of Texas Press, Austin, pp 159–191 Mohanan C, Sharma JK (1991) Seed pathology of forest tree species in India – present status, practical problems and future prospects. Commonw For Rev 70:133–151 Mohanan C, Chacko KC, Chandran A, Varma G (2005) Seed health problems in tropical forest tree seeds and their impact on seedling production. In: Working paper of the Finish Forest Research Institute 11 Morellato LPC (2004) Phenology, sex ratio, and spatial distribution among dioecious species of Trichilia (Meliaceae). Plant Biol 6:491–497 Nair KSS (2001) Pest outbreaks in tropical forest plantations, – is there a greater risk for exotic tree species. CIFOR, Jakarta, 82 pp Namkoong G, Barnes RD, Burley J (1980) A philosophy of breeding strategy for tropical forest trees, vol 16, Tropical forestry papers. Department of Forestry, Commonwealth Forestry Institute, University of Oxford, Oxford, 67 pp Neergaard P (1979) Seed pathology. Macmillan, New York Ng FSP (1992) Manual of forest fruits, seeds and seedlings, 2 volume, vol 34, Malayan forest record. Forest Research Institute of Malaysia, Kuala Lumpur Nikles DG, Newton RS (1983) Inventory and use of provenance resource stands of Pinus caribaea Mor. var. hondurensis Barr. and Golf. in Queensland. Silvic VIII 29:122–125 Nyoka BI (2003) Biosecurity in forestry: a case study on the status of invasive forest tree species in Southern Africa. In: Forest biosecurity working papers. Working paper FBS/1E FAO, Rome Obiri JF (2011) Invasive plant species and their disaster-effect in dry tropical forests and rangelands of Kenya and Tanzania. J Disaster Risk Stud 3(2):417–428 Old KM, Wingfield MJ, Yuan ZQ (2002) A manual of diseases of eucalypts in South-East Asia. Center for International Forestry Research, Bogor, 196 pp Opler PA, Bawa SK (1978) Sex ratios in tropical trees. Evolution 32(4):812–821 Overbeek W, Kro¨ger M, Gerber J-F (2012) An overview of industrial tree plantation conflicts in the global South. Conflicts, trends, and resistance struggles. EJOLT report no 3, 100 p Owens JN (1995) Constraints to seed production: temperate and tropical forest trees. Tree Physiol 15:477–484 Owens JN, Blake MD (1985) Forest tree seed production. A review of the literature and recommendations for future research. Information report, PI-X-53. Petawawa National Forestry Institute, Canadian Forest Service Palmberg C (1985) Sampling in seed collection. In: FAO (ed) Forest tree improvement, Forestry paper, 20. Food and Agriculture Organization of the United Nations, Rome, pp 41–45 Pazos GE, Greene DF, Katul G, Bertiller MB, Soon MB (2013) Seed dispersal by wind: towards a conceptual framework of seed abscission and its contribution to long-distance dispersal. J Ecol 101:889–904

Genetics and Forest Seed Handling

957

Phartyal SS, Thapliyal RC (2005) Dispersal germination syndromes of tree seeds in a monsoonal forest in northern India. Seed Sci Res 15(1):29–42 Radke P, Radke A (2004) Plantation improvement using clonal propagation – an overview of the latest technology in Australia. Clonal Solutions Australia Pty Ltd. Prospects for high-value hardwood timber plantations in the ‘dry’ tropics of northern Australia, Mareeba, 19–21 Oct 2004 Rajora P, Mosseler A (2001) Challenges and opportunities for conservation of forest genetic resources. Euphytica 118:197–212 Robertson AW, Trass A, Ladley JJ, Kelly D (2006) Assessing the benefits of frugivory for seed germination: the importance of the deinhibition effect. Funct Ecol 20:58–66 Sakai K, Momose K, Yumoto T, Nagamitsu T, Nagamasu H, Hamid AA, Nakashizuka T (1999) Plant reproductive phenology over four years including an episode of general flowering in a lowland dipterocarp forest, Sarawak, Malaysia. Am J Bot 86(10):1414–1436 Sedgley M, Griffin AR (1989) Sexual reproduction of tree crops. Academic, London Toole VK (1973) Effects of light, temperature and their interactions on germination of seeds. Seed Sci Technol 1:339–396 Trueman SJ (2006) Clonal propagation and storage of subtropical pines in Queensland, Australia. South Afr For J 208(1):49–52 Turnbull J, Doran J (1987) Seed development and germination in the Myrtaceae. In: Langkamp P (ed) Germination of Australian native plant seed. Inkata, Melbourne, pp 46–57 van der Pijl L (1982) Principles of dispersal in higher plants. Springer, Berlin/ Heidelberg/New York Vazquez-Yanes C (1982) Phytochrome control of seed germination in the tropical rain forest pioneer trees Cecropia obtusifolia and Piper auritum and its ecological significance. New Phytol 92:477–485 Vazquez-Yanes C, Orozco-Segovia A (1993) Patterns of seed longevity and germination in the tropical rainforest. Annu Rev Ecol Syst 24:69–87 VECEA (2012) Potential natural vegetation of Eastern Africa. Vegetation and climate change in Eastern Africa. Forest & Landscape, Frederiksberg White TL, Adams WT, Neale DB (2007) Forest genetics. CAB International, Cambridge, 573 pp Willan RL (1984) Provenance seed stands and provenance conservation stands. Technical note, no 14. Danida Forest Seed Centre Williams ER, Matheson AC (1993) Experimental design and analysis for use in tree improvement. CSIRO and ACIAR Wingfield MJ, Coutinho TA, Roux J, Wingfield BD (2002) The future of exotic plantation forestry in the tropics and southern hemisphere: lessons from pitch canker. South Afr For J 195:79–82 Wright J (1976) Introduction to forest genetics. Academic, New York, 460 pp Zobel B, Talbert J (1984) Applied forest tree improvement. Blackburn Press, Caldwell

Forest Seed Collection, Processing, and Testing Lars Schmidt

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flower and Seed Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Planning the Seed Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Collection Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Documentation of Seed Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Collection for Genetic Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Handling, Processing, and Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Handling Between Collection and Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Storing the Seeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Lot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Purity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Weight Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Determining the Moisture Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Viability and Germination Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Estimating Viability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Using Test Results in Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

960 961 961 962 967 970 971 972 972 974 980 983 984 984 985 985 986 987 988 990 991 991

Abstract

This chapter pertains to the techniques of capturing the best genetic quality seeds a seed source can produce at the optimal time of high physiological maturity and maintaining these qualities throughout the handling processes, all at a minimum L. Schmidt (*) University of Copenhagen, Copenhagen, Denmark e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_225

959

960

L. Schmidt

cost. Different collection and processing techniques apply to different species, seed types, situations, and purposes. Yet the collection and processing toolbox contains a number of “standard” methods for most of these groups. Records and documentation help in evaluating “best practice” for future method improvement, and it helps in linking offspring to seed source. Conditions are set for short- and long-term seed storage by their inert storability physiology. The potential storage life of seed may for some robust “orthodox” species be several decades, while no available storage conditions can maintain viability for sensitive “recalcitrant” seed. Seed testing aims at quantifying seed quality parameters such as seed weight, moisture content, purity, and germination. The methods contain a set of standard procedures for preparation and evaluation. Special types of evaluation contain, e.g., health teats and indirect methods of measuring seed viability. Information gathered through seed testing are used as a general quality documentation during seed trade and as a guide for subsequent handling. Keywords

Seed collection • Genetic conservation • Viability • Extraction • Seed cleaning • Moisture content • Target moisture content • Orthodox seed • Recalcitrant seed • Desiccation-sensitive seed • Seed lot • Sampling • Purity analysis • Seed weight • Germination • Maturity indices • Frugivores • Tree climbing • Seed documentation • Heat isolation • Seed testing

Introduction Seeds are carriers of all the genetic information needed to grow into a mature tree. We have seen in connection with the discussion of seed sources and reproductive biology (chapter “▶ Genetics and Forest Seed Handling”) that the quality (desired properties or characters of seed and offspring) is created during the mating process and maintained during maturation. What we collect is what we get, and the way we handle it determines the results. Hence, within the spectrum of biological limitations, seed quality is largely controlled by the way we collect and handle them. Collection and processing are a chain of handling procedures, where each step bears the risk of irreversible damage. Seeds are generally designed to tolerate stress during dispersal and a period before germination. Some seeds are extremely robust and “almost impossible to kill” under normal handling conditions, for example, many pines and legumes. However, at the other end of the scale are some extremely delicate species whose seed will never survive long (desiccation-sensitive or recalcitrant seed), some with thin and fragile seed coats, and some species extremely prone to pest and diseases. Such species must be handled with utmost care during the entire process In a biological context, the waste of genetic material in nature is exorbitant. Long-living trees may produce billions of seed during their lifetime; most of them will succumb. In seed handling, we try to optimize the resources to make all collected seeds germinating and surviving. The ability to germinate is inert in all

Forest Seed Collection, Processing, and Testing

961

seeds, but “filters” that are designed to optimize survival in nature sometimes appear as obstacles to seed handling and germination. Viable seeds can germinate, but it is not always that we can make them germinate. Research has helped overcoming numerous past problems, and progress is still being made; yet still many remain.

Seed Collection Seed collection is about capturing the best genetic quality the seed source can produce, at the optimal time for high physiological quality. Seed processing is about maintaining the physiological quality of these genetically best quality seeds through adapted processing, handling, and storage technology. Seed testing, with its various subcomponents, aims at measuring in a quantifiable manner the physiological seed quality: how many healthy plants can we, under optimal condition, expect from a given quantity of seed? The term “optimal conditions” has an important role in practical propagation, because “optimal” refers to “best practice.”

Flower and Seed Survey If there are no flowers, there will be no seed. Yet, as discussed in section “Seed Processing and Storage,” a good flower crop does not always result in a good seed crop. Predicting seed crop is the first step in planning seed collection, and the earlier a seed crop can be predicted the more time for planning. The duration from flowering to fruiting varies from a few months in some fast pioneers to about a year for many climax forest species. Some gymnosperms may even take up to 2 years from flower differentiation to mature seed (Haines and Nikles 1987). Observation of flowering is especially important in periodic reproducing species such as many rain forest trees. Flowering is registered as +/ flowering and quantitatively as assessment of number of flowers in some key classes, e.g.: 5. Prolific (flowers in large number over the flower color dominating the crown) 4. Abundant (flowers numerous but some parts with less flowers and green leaves visible) 3. Intermediate (flowers numerous, but large parts of the crown with less flowers) 2. Few (sporadic flowers on some branches) 1. None (no or very few flowers can be seen) Details on flower classes vary with species; above guide may fit to potentially heavy flowering Pterocarpus indicus, while large-flowered species such as Sterculia and dipterocarp species always have much fewer flowers (Fig. 1). Flower surveys can be very difficult in closed forest with tall trees and flowers borne at uppermost parts. Some options are:

962

L. Schmidt

Fig. 1 Shed dipterocarp flowers littering the forest floor in Sabah after an event of general flowering

1. Inspection climbing or viewing from high elevation 2. Surveying fallen dehisced flowers beneath the trees 3. Inspection via drone surveys. Still a rather expensive solution but becoming optional with the development of drone facilities and techniques Young fruit surveys are often more difficult because immature fruits are generally less conspicuous. However, immature fruit surveys can be used for qualitative assessment, e.g., examining seed contents in samples of green fruits or cones.

Planning the Seed Collection Planning seed collection includes: 1. Estimation of quantity of seed needed for the forthcoming season based on the assessment of seed stock, turnover rates, and possible new markets/customers/ users 2. Identification of collection sites/seed sources 3. Elaboration of tentative seed collection calendar for the upcoming season 4. Estimation of manpower needed for collection and processing Re. 1. Seed demand for different plantation model stocks depends on planting area, planting density, expected field mortality, expected nursery mortality, and germination percentage (see calculation example in Schmidt 2000). Mortality varies tremendously whether seeds are established in nursery systems or by direct sowing (Schmidt 2008). Species which produce seed only with several years’ interval impose specific problems because surplus seeds must be collected, which can meet the demand during interim years. Excess collection may also be done simply to rationalize collection resources. This can obviously only

Forest Seed Collection, Processing, and Testing

963

be done for orthodox seed. Periodic flowering and fruiting (masting) are usually only evident when flowers and fruits appear (unless there is a very reliable predictability based on weather correlation). Supply and demand of periodic reproducing recalcitrant species is rather opportunistic. Hence, both seed supplier and planters must be geared to take advantage of sudden appearing seed crops in connection with a masting event. Since the quantity of collectable seed during masting years is sometimes overwhelming, alternative establishment strategies should be considered, e.g., direct sowing. Most direct sowing experience is for pioneer species on barren land, but experience has shown that it is also applicable for undercanopy planting of late-successional species (Doust et al. 2006; Cole et al. 2011). Re. 2. The seed sources, where the best genetic quality can be collected, are prioritized. However, in case there is wide variance within the species and planting region, several suitable provenances may be collected. Re. 3. Seasonal climates usually have seasonal fruiting, while less seasonal equatorial climates often exhibit a mixture of reproductive strategies with coherent phenology patterns (cf section “Seed Processing and Storage”). Phenology observations on flowering and fruiting of target species can be used to draw seed collection calendars. However, collection dates often vary from year to year depending on local variation of weather. If target species grow far from the operating seed collection station and it is not possible to get information from the seed source site, indicator specimens can be a shortcut. These are individual trees growing nearby. Since phenology within species is often synchronized, such trees will often give information on development of stands (with several reservations for deviations). Seed should be collected from heavily fruiting trees in the middle of stands and the middle of the season, i.e., edge trees and early and late fruiting individual seed trees are avoided. This is because of concern that such trees may be functionally isolated from good cross-pollination because the potential pollen exchange has been limited due to time or local site isolation. Edge trees will often flower early because of better light exposure and are likely to be pollinated by other edge trees. Maturity criteria for different fruit types must be experienced (Box 1). Fruits of species with orthodox seed usually appear full mature size long time before seeds are physiologically mature. During the time from attaining full maturity size and weight, seeds mature by biochemical processes, zygote development, and maturation drying. In desiccation-sensitive seed, fruits and seed continue to accumulate dry matter (i.e., increase in size and weight) up to full maturity (fruit dispersal) (Finch-Savage and Blake 1994; Berjak and Pammenter 2003). The difference has the practical implication that whereas orthodox seeds may be picked and afterripened sometimes several weeks before their natural dispersal, this is not possible for desiccation-sensitive seed because early collection means inadequate development. Re. 4. Distance, access, collection method, and possible field processing influence labor requirement. Most collection methods require a small working team, e.g., for safety in tree climbing (Anon 1995). Alternative to, or as a supplement to,

964

L. Schmidt

collection teams, some seed-collecting institutions/companies rely on casual workers or contractors. Such network can be highly efficient provided quality measures and quality control (e.g., on a number of and distance between mother trees, collection conditions, and documentation) are implemented. Box 1: Maturity Indices Colour change: Dry fruits: green to yellow / brown

Early dispersal events: Visiting frugivores, fruits on the forest floor

Fleshy fruits: Green => conspicuous

Peduncle detachment (abscission)

Softening of pericarp or sarcotesta

Taste (sweet)

Dehydration – drying of dry fruits. Seed wings => breaking w/o bending

Funicle detachment

Odour, - e.g. nocturnal animal disperser

Splitting open dry dehiscent fruits

Resin exuding, ‘sweating’ fruits

Equipment and Transport The need for equipment and accessories depends on seed types and collection method. If several collection teams are operating, each team should be provided with a “standard kit” containing GPS, local tree flora, seed documentation forms and labels, and writing tools (permanent markers). For most collections tarpaulins, long-handled pruners which can be mounted with saws and secateurs, hand secateurs, flexible saw with throw bag, and drying and depulping devices (latter including water containers and hoses) will be needed. For tree climbing, spurs, ropes, safety belts, and rope clamps are necessary; especially for advanced line

Forest Seed Collection, Processing, and Testing

965

technique, various throwing or ballistic devices (slingshots, crossbows, or the like) are required. Lengthy expeditions with overnight stays may require various camping equipment.

Restrictions and Permits Nowadays seed collection is often restricted by ownership and/or protective rules. Open access to seed sources is generally restricted to seed sources in degraded forests, roadside plantings, or public areas. The vast majority of seed sources are owned by privates or public institutions who would often demand documentation of person and purpose of collection and, in case of larger collections, demand a fee paid. Where national seed supply systems exist, forest departments and private seed suppliers often have their own planted seed sources of their main species. Outside collectors may here be considered intruders. Collection rules usually emphasize that collection must be done without or with minimum damage to the trees. Ground collection does not do any harm to trees, but climbing with spurs and pruning seed-bearing branches inevitably imply some damage. It can therefore be difficult to achieve collection permits for such methods in some protection areas. The use of guns for shooting down fruit-bearing branches obviously requires specific weapon license and permit to use on specific sites. In addition, seed collectors should be aware of possible restrictions for the use of other shooting devices such as slingshots/catapults. Seed collection outside own seed sources is most efficiently complied with an annual renewable contract basis. Contracts should state ownership details, collector details, exact location for collection, approximate quantity of permitted collection under the contract, and possible restriction on methods (vehicle use, equipment, damage to trees, etc.) (Fig. 2). Organizing Post-Collection Handling Post-collection seed handling includes processing (extraction, cleaning, drying) and seed storage. During peak season, where several collection teams may be occupied, loads of seed may be delivered to the central station. Processing must be performed as quickly as possible in order to limit seed deterioration in transit. 1. Each seed lot should be inspected upon arrival. If seeds are moist, they should go to dry or wet extraction depending on the fruit type. 2. Green immature fruits are put to afterripening in shaded conditions. 3. Recalcitrant seed should be treated as fast as possible as they will always deteriorate fast.

Organizing Record Keeping Some of the most crucial seed information refers to collection site and field conditions. Such information must necessarily be recorded in the field. It is practical

966

L. Schmidt

Fig. 2 Large slingshot used for advanced line technique. Beware of what is considered a potential weapon with license requirement

to organize recording by preprepared forms and labels. Because of the ease and storage capacity of modern computers, it is tempting to collect and record large amount of data. However, putting a heavy workload on field recording may be counterproductive if information is not exact. Simple writing conditions in the field may discourage detailed recording. A proposed field recording system is shown in annex A. The essence is that any seed lot can be traced back to its origin. Field staff seed lot numbering systems should be arranged so it gives no confusion with central number system. A central seed collection and distribution system typically consist of certain substations (usually located at different places), each of which has various numbers of collection teams. Each collection team and each substation is allocated an unambiguous identity, e.g., a three-letter acronym (avoid one letter or one number as they easily get confused with other information like species and dates). Each collection team can then use a continuous numbering as they collect seed lots. There is no need to include species name in the seed lot number that just ads problems if the species is incorrectly named in the field. For example, seed lot number Mos-Bat-14-009 means the 9th seed lot collected in year 2014 by Moses collection team from Batu substation. Please notice that the tentative field number is maintained in the final recording system.

Forest Seed Collection, Processing, and Testing

967

Some practical hints, labels and ink, must be water resistant, so reading will persist even if labels are exposed to natural moisture. Labels and seed documentation forms should be part of seed collection kits together with collection equipment, containers, etc. Field staff should be trained in filling in forms. Labels should remain together with fruit and seed throughout processing. Simple clamps to attach labels ease practical handling. Redundant labels, e.g., after extraction where a number of containers are reduced, should be thrown away to avoid later confusion. Once seeds are entered into the central seed system, a new seed lot number is assigned (see section “Documentation of Seed Collection”).

Seed Collection Techniques Collection techniques refer to the physical picking of seed from crowns or beneath trees to seed collectors’ containers. Most physiological and mechanical damage occurs during this process. Since damage cannot be undone, adopting the best collection technique is crucial for the ultimate seed quality. Seed collection is usually one of the most expensive seed handling procedures since it typically involves transport to seed sources and labor-demanding collection. Despite the development of accessories, seed collection is still pretty much a manual work. Collection from the ground, possibly after shaking, is normally much cheaper than collection from the crown, especially where the latter involves arduous climbing. The rule is that the simpler and cheaper methods apply as long as it does not compromise seed quality.

Collection of Fallen Seeds and Fruits Seeds that are not dispersed will eventually fall to the ground. Abscission is the process where peduncle or pedicels develop a weak zone which keeps the fruit attached to the tree, strong enough not to fall by pure gravity force, but still can easily be detached by wind or frugivorous animals (Maurer et al. 2013; Pazos et al. 2013). The fraction of seed falling under the mother tree is thus usually small compared to the total fruit production. In most species, fallen seeds and fruits are readily attacked by pests or predators (some of which may also be dispersers), or in some environment, and particularly for recalcitrant seed, they will quickly be lost to germination. Collection entirely after natural fall is thus rather unreliable. Another problem of collection after natural fall pertains to very small-seeded species like Octomeles and eucalypts. Such seeds are simply not possible to find unless forest floors are covered by tarpaulins or other dense materials. But even under these conditions, seeds are easily blown away. Ground collection usually involves some shaking of branches that simulate the mechanical impact of dispersal release. If shaking is too violent, too many immature fruits will be released, and if they cannot be afterripened, it is wasted. For seed orchards on plane land, special tree shakers are available that can “empty” a tree for fruits in few seconds (McLemore and Chappell 1973). However, for most species, manual shaking of branches by ropes, hooks, or the like from

968

L. Schmidt

Fig. 3 Collection from the crown by shaking branches with throw lines or pole hooks

ground or by climbers provides adequate impact to release mature dispersal ready fruits (Fig. 3). Tree species with long fruiting time impose some special problems because only small quantity of seeds can be collected at any time. Such trees must often be revisited several times during the fruiting season in order to harvest adequate seed (Figs. 4 and 5). Large fruits or seed may be collected manually. Smaller seeds are preferably collected from laid-out nets or tarpaulins. Where ground vegetation impedes this, the nets can sometimes be hung up above soil level. If the forest floor is reasonably clean (e.g., in established seed sources), seeds may be swept, raked, or vacuum collected (Mineau 1973; Riley et al. 2004). These methods imply, however, contamination with soil and debris, possibly also infected with pathogens which, unless the unprocessed mix is used for bulk direct sowing, imply subsequent cleaning operation to eliminate these non-seed materials.

Collecting Seeds from the Crown Seeds are collected from the crown if ground collection is not possible, for example, if fruits are readily being removed by predators/dispersal agents, if fallen seeds tend

Forest Seed Collection, Processing, and Testing

969

Fig. 4 Extraction by drying on wire mesh. The fruit capsules open when dry and seeds fall out themselves. Under humid conditions, extraction is facilitated by kiln drying, where fruits are exposed to a hot air current

Fig. 5 Pterocarpus seed germinating from fruit without extraction (left) and after extraction (right). Extraction is often omitted because of labor and high risk of seed damage

to be easily destroyed by predators or removed/lost by dispersal or fast germination, or if small seeds are blown away if ground collection is attempted. Under these conditions, fruits are collected from the crown either from a ground position or by climbing. Low trees and bushes up to say 6–8 m high do not impose much problems because seeds can be collected by pruning small branches. Sometimes even large trees have fruit-bearing branches at low height that can be pruned. However, outcrossing is often better at higher positions of the crown; relying on the low-hanging fruits is thus not always advisable (Patterson et al. 2001). High fruitbearing branches may be cut down by the help of advanced line technique, where flexible saws are placed over high branches and cut down by alternately pulling the two ends of the ropes. It is, however, difficult to place and operate the saw at more than about 10 m height. In Australia, shooting down tall fruit-bearing branches is common for collection of small-seeded eucalypts (Gunn 2001; Gunn et al. 2004). When access to higher parts of the crown is necessary, there are various options. Hydraulic automobile or truck-mounted platforms/lifts may be used in open flat terrain. Although devices are getting relatively common thanks to their wide use in building industries, the cost of purchase/renting and operation costs still inflict

970

L. Schmidt

restrictions for use in seed collections. Climbing trees for seed collection is, albeit effective, usually reduced to a minimum because of risk and costs. However, the development of lightweight, safe, and strong climbing equipment has made climbing an easier and less risky task (Anon 1995; Blair 1995). Ladders are suitable versatile equipment for climbing low crowns or clear boles and to getting access to crowns with dense branches, which can be used as support. Telescopic ladders up to 6–8 m are available, are lightweight, and take very small space. Aluminum extendable ladders can be used the same way but are more bulky and difficult to carry. Climbing spurs are suitable for trees from about 25 cm diameter to about 70 cm diameter with relatively soft wood or bark where spurs can easily be kicked in. Some hardwoods and palms have extremely hard wood where spur climbing is in practice impossible. Spurs are also difficult to use for smaller diameters because of sideway balance; larger trees are difficult because of foot position and problems of maneuvering the safety strap. For suitable species and sizes, spur climbing is easy as it is a simple “walk” up the trunk by kicking the spurs into the trunk. A safety strop tied to the climber’s safety belt goes around the trunk at all times. When passing branches, a safety strop is placed above the branch before releasing the strop under the branch (Blair 1995). Smaller trees of say 10–30 cm diameter may be climbed using “hitch knot” climbing technique. The climber uses two loops tied around the tree trunk. One is tied to the climber’s safety belt; the other is used for the foot. The climber ascends by lifting himself by the leg and then pushing up the body hitch knot. Hanging in that one, he moves upward the foot knot. The method is safe because the climber is tied to the trunk at any time. It is, however, rather physically demanding. This method is also suitable for climbing short difficult distances between long spacing branches in the crown. Advanced line technique consists of placing a climbing rope over a high branch by some shooting devices (throw line, bow and arrow, catapult). Modern tree climbing technique usually uses special climbing rope where climber moves up and down by the help of foot and hand rope clamps. Both hand and foot clamps are tied via straps to the climber’s safety belt, and movement is both easy and safe. Rope climbing is also convenient for sideways climbing where climbers move around in the crown collecting seeds or pruning seed-bearing branches or fruits.

Documentation of Seed Collection Availability of various data storage and management systems has made data management much easier than before but also rather daunting. The relevant information for seed documentation is the same, whether documentation is in manual ledgers or electronic devices: assignment of seed lot identity code, site of seed collection, conditions, etc., are all relevant. The strength of the computer-based electronic system is that:

Forest Seed Collection, Processing, and Testing

971

1. Large amount of data can be stored at small physical space. 2. Data in one system can be linked to a different system, e.g., (seed source) map or seed testing database. 3. Relevant data can easily be retrieved from a database as long as a person has access to the system. In a database, seed information may be retrieved under different listing criteria, for example, by species or by seed source type. A rather user-friendly system is Microsoft Access #. Databases sort and list on criteria based on digits (numbers or letters) from left to right. A few practical advices ease subsequent data management: 1. Seed lot identity consists of a species code of four capital letters (normally the first two letters of genus and species epithet, e.g., ACMA for Acacia mangium), followed by a seed lot number. Practical system is chronological numbering as seed lots enter the system for each year. Since database listing is normally done from left to right, it is advisable to list year before the number. For example, a seed lot number ACMA-14-089 means the 89th seed lot of Acacia mangium collected in 2014. 2. In addition to GPS coordinates, use region, provenance, and, when applicable, seed zone indications for site identification. This makes listing according to geographical parameters easier.

Seed Collection for Genetic Conservation Conservation of forest genetic resources aims at maintaining the genetic integrity of populations. Since selection and breeding tend to narrow the genetic base (by eliminating less fortunate genes), they also limit the potential gene pool for improvement in the future. Plant breeding often includes wild relatives of domesticated plants because these populations may contain attractive genes (e.g., for disease resistance) which may not be present in breeding populations. Hence, for breeding purposes, maintenance of gene pool is important. In addition, many lesser used species with limited distributions, in particular endemic species, are often endangered on species level. Conservation of forest genetic resources is a mixture of ex situ and in situ conservation. Ex situ conservation plays an increasingly important role, because in situ populations are often difficult to protect. For discussion on the use of ex situ and in situ conservation, reference is made to FAO et al. (2001, 2004a, b). The aim of ex situ conservation is to gather genetic variation from wild populations and either keep them in long-term seed banks (Linington 2003) or raise them in plantations where they can be protected. Conservation collection includes five key challenges: 1. Delineation of the area within which to collect seed. If species include several definable provenances, it may be desirable to maintain provenances separate.

972

L. Schmidt

However, if provenances have been reduced to a very small number of interbreeding trees, it may be sensible to mix close provenances. 2. Identification of representative trees and populations. The history of fragmented populations should as far as possible be assessed. Populations of interbreeding trees would be expected to have greater similarity than isolated populations. What appear as fragmented populations may in some instances be remnants of large recently interbreeding populations; sometimes remnants of ancient coherent populations are separated by, e.g., climate change; in other cases, they are “outliers,” “satellite,” or “island” populations isolated from the main population(s). Such outliers may contain rare alleles not represented in the main population. On the other hand, the outliers may contain very little genetic variation since they may have originated from a small number of accidentally dispersed seeds. Examples of natural fragmented populations are many northern hemisphere species (pines, cypresses, oaks) whose distributions stretch into tropical highlands (Critchfield and Little 1966). 3. A number of mother trees. As a rule of thumb, conservation stands should contain at least 50 widely dispersed mother trees. Of these, the majority should be from the main populations. 4. Maintaining mother tree identity. For fear of losing information, it is often tempting to maintain details, e.g., maintaining single tree collection separate. With modern days’ data management, technology may be suitable. However, if seed lots are supposed to be treated as a unit, e.g., a provenance collection, the time allocated for maintaining separate mother tree collections should be balanced with the possible benefit.

Seed Handling, Processing, and Storage The ultimate destination of seed is to germinate and develop into a healthy plant that grows into a new reproducing tree. The objective of seed handling is to maintain that property from collection to germination. The processes after collection are processing, storage, and distribution. During processing, seeds are exposed to mechanical treatment (extraction and cleaning) and modification of water content (usually drying). During storage, seeds are enforced and maintained in a nonactive (quiescent) stage for a certain time period, where they are potentially deteriorating through natural aging or through infection and infestation. During distribution, seeds may be exposed to adverse conditions different from the storage environment. During all phases, seed should be treated as a living (and thus potentially dying) material with an identity (seed lot number).

Seed Handling Between Collection and Processing Depending on distance between collection site and central processing unit, temporary processing may be necessary in the field. For example, bulky material may

Forest Seed Collection, Processing, and Testing

973

need to be reduced in order to reduce transport cost, and easily deteriorating material may need to be eliminated in order to maintain seed viability. Preprocessing does not differ much from final processing, only that equipment are normally lightweight portable types. For dry seed, lightweight trays, tarpaulins, sieves, and “hammocks” are suitable for sun drying. Precautions should be made against wind blowing away light seed. Often seeds of dehiscent fruits can be released by drying alone after which fruits and branchlets can be removed. Extraction from fleshy fruits is usually more complicated since it requires excess water. However, if many such fruits are collected, depulping equipment, possibly including high-pressure “jet cleaner” equipment, should be considered as “standard” equipment.

Maintaining Seed Viability Any seed that does not germinate immediately will age or deteriorate, which will ultimately lead to the death of the seed (Ellis 1986). Aging for orthodox seed stored under optimal conditions is very slow. Some types of seed deterioration can be repaired during seed germination; others are irreversible. The aim of seed handling after collection is to slow down the aging process as much as possible. This is done by reducing exposure to conditions conducive to deterioration, i.e., high temperature, moisture, and infecting or infesting organisms. For orthodox, seed desiccation reduces and eventually (at around 6 % mc) brings to a halt all metabolic processes (Pammenter and Berjak 1999), both in the seed itself and in possible pests and pathogens. However, both fungal spores and some insects are able to survive very low moisture content. Low temperature has the same effect but also slows internal non-metabolic deteriorations such as disintegration of membranes and nucleotides. Desiccation-sensitive seeds impose special problems because seed viability cannot be maintained by the orthodox way of drying and cooling. For those seeds, there are two ways, which depend on their sensitivity: 1. Drying to lowest safe moisture content and cooling to the lowest safe temperature for the species. This may, for the more resistant species, maintain viability for several months. 2. Letting the seeds germinate at slow rate and keeping them as germinants. Spraying seeds with water keeps them moist and alive.

Maintaining the Seed’s Identity The identity tag of seed is the seed lot number, which can be tracked back to the details of origin and collection details (section “Organizing Record-Keeping”). Because field handling often involves frequent unloading for field processing, tags and labels can easily be lost. Further, seed collection team is often different from processing team, and since seed lot number in the field is temporary, it must be transferred into a proper seed lot number, and after hand over, another potential error can occur. However, some simple routines can overcome most problems:

974

L. Schmidt

1. Field tags should be designed premanufactured with headings and lines (seed lot no., date, etc.) made of durable material (e.g., hard plastic or cardboard) and easy to attach and clip on containers and processing equipment. 2. An identical tag should be put on the outside and one inside of the container (if seeds are moist, it should be put inside a plastic bag). 3. Redundant labels should be thrown away to avoid confusion of their belonging. 4. The field identity number should be maintained with the seed lot after a permanent seed lot number is assigned.

Processing Processing encompasses extraction, cleaning, and possible drying for storage. Processing has various objectives: 1. Bulk reduction by extracting seed or the smallest extractable unit from bulk fruits. The ratio of fruit/total seed weight varies from about 1:1 in some legumes to > 20:1 in some fleshy fruits. That means that bulk and thus storage space can be reduced to at least half by extraction. Cleaning after extraction has also a bulk reduction purpose, viz., to eliminate all non-seed material from the seed lot. 2. Splitting up seeds in individual sowing units. This is practical during subsequent handling. However, occasionally, several seeds are enclosed in locules in an indehiscent pyrene (stone of drupe fruit, e.g., Melia spp.); these cannot be split without destroying the seeds (Fig. 7). 3. Remove possible dormancy influence from enclosing fruits, e.g., germination inhibitors. 4. Remove decomposable fruit material which can influence seed viability and storability. This includes primarily fleshy structures from berries, drupes, and compound fleshy fruits, but also some arils and sarcotestas consist of readily decomposable matters. 5. Dry seed to a lowest safe moisture content to allow storage during a desirable period.

Seed Extraction Seed extraction means separating seeds from enclosing fruits. The procedure depends on fruit type, and as we have mainly two fruit types, dry and fleshy, extraction typically follows one of two lines, dry and wet extraction, respectively. Dry extraction may be pure sun drying, high-temperature kiln drying, or combined with mechanical extraction: 1. Drying alone is adequate for many dehiscent fruit types such as capsules and pods. The fruits are usually collected just before dehiscence to avoid seeds being lost to wind dispersal. Capsules of Cedrela, Swietenia, and most eucalypts, pods of Crotalaria, and follicles of Grevillea and Sterculia belong to this group.

Forest Seed Collection, Processing, and Testing

975

2. Kiln drying is an extension of air-drying also used for dehiscent fruits. It is used, for example, if air humidity is high or for fruits that only dehisce at exceptional high temperature (in nature, e.g., in connection with fire) the so-called serotinous fruits (Banksia, Hakea, and some pines). 3. Sometimes seeds remain trapped inside the fruit or attached to the dehisced fruits via the funicle, e.g., wind dispersed pods (e.g., several Acacia and Albizia species). Beating bags containing fruits with sticks are usually adequate to break funicle attachment, after which seeds can be separated by tumbling. Indehiscent fruits do not open by drying alone but must be disintegrated by some mechanical treatment, e.g., in hammer mills. Fruits of tamarind type (Tamarindus, Inga, Dialium, Pithecellobium, Prosopis) are dry fruits where seeds are embedded in a sticky substance. The dry part may be removed by gentle mechanical treatment (flailing or beating) and the sticky part by wet extraction. Wet extraction is used for removing soft fruit cover from seed of fleshy fruits like berries, drupes, compound fruits, arillate seeds, and seeds with sarcotestas. Soft material tends to stick to endocarps or seed coats. The material tends to decompose (rot) easily after which it can be removed mechanically. The basic procedure is as follows: 1. Fleshy fruits or seed is soaked in water to soften the pulp. It is advisable to use some flow, stirring, or aeration during soaking to prevent fermentation and consecutive formation of deleterious alcohol. 2. Regular stirring and “skimming off” released fruit material (clean seeds tend to stick to the bottom and fruit pulp floats). 3. Final cleaning in running water, possibly after brushing, mechanical stirring, or wet tumbling with abrasion material. Soaking should be no more than 1–2 days since longer-term soaking tends to exert some physiological damage to seed which influences both immediate germination and their storability (Ogunnika and Kadeba 1993; Srimathi et al. 2003; Apetorgbor et al. 2004). Several machines have been developed to ease soft fruit treatment, e.g., coffee depulper and “Dybvig” depulper. Household machines such as blenders and potato peelers are conveniently used for small quantities; rotating cement mixers are versatile machines for larger quantities. High-pressure water in so-called jet cleaners unifies the mechanical impact of the water current with the cleaning effect. Unlike some mechanical rupture, water current is gentle to delicate seed coats. However, due to the jet impact, seeds must be prevented from “blowing” away, e.g., by “packing” them in small-mesh-size wire mesh baskets. Some fruit types are very difficult to disintegrate for seed extraction. In Tamarindus, Prosopis, Dialium, Inga, and similar fruits, the pulp is sticky and difficult to loosen from the seed by traditional water depulping methods. Here biological methods are sometimes suitable, for example, feeding to goats with subsequent extraction from the feces or feeding to termites who will preferably eat the sweet pulp and leaving the clean seed behind (Masilamani and Vadivelu 1993).

976

L. Schmidt

Extracting seeds from very hard fruits sometimes implies damage to the seeds, and seeds are in those cases sometimes sown without extraction, i.e., the entire fruit despite germination is often slower (mechanical or physical dormancy); see Fig. 9.

Seed Cleaning Seed cleaning is a separation process in which non-seed materials such as fruit pieces, flowers, fruit and seed stalks, branchlets and leaves, or soil fragments are removed from the seed lot. The purpose of cleaning is primarily to reduce redundant bulk and secondarily to remove potential material for harboring pests and pathogens. Cleaning thus also includes the possible removal of insect-infested seed. However, even if some impurities may not be harmful to seed, a high degree of purity has some psychological effect in seed trade: customers often regard impurities as poor quality. Separation is based on differences in physical characters (Karrfalt 2008), e.g.,: 1. Size. Separation via sifting, which can be used to eliminate both larger and smaller than seed material by using different mesh sizes 2. Specific gravity. Separation by winnowing or specific gravity separators. For very small differences also flotation in water or other fluids with gravity between the two items to be separated 3. Shape. Separation based on differences in gravity points. Separation on hand screens, rotating belts, and indented cylinders 4. Surface structure. Separation by screens, rotating belts, and vibrating tables The challenge in seed cleaning is that both seed and impurities (debris) encompass a range of physical variation which often overlaps. Some inert matters, e.g., “other seeds,” may be very similar to seeds in some character (e.g., size) but differ in others (e.g., gravity or form). Using one method only will usually leave some inert matter together with the seed lot. However, a combination of two or more methods, each with different degrees of purity, may be used in progressive cleaning (Fig. 6): 1. Each method can be adjusted to give different classes of purity from pure inert matter over mixed inert matter and seed to pure seed. For size, it is separated by different mesh sizes, for gravity by different wind speed, and for shape and surface by, e.g., variation of slopes of screens or belts. 2. Pure debris is discharged, pure seed is kept without further cleaning, and impure fractions are cleaned using a different method until the seed lot is as clean as deemed necessary. Several progressive cleaning operations by different methods may clean seed lots to 100 %. The simplest cleaning procedure is used first followed by some of the more advanced procedures. Small seeds of the target species may deliberately be removed during cleaning since small seeds often have lower vigor than larger seed. However, seed size is also genetically controlled, and in a seed lot consisting of several families of which

Forest Seed Collection, Processing, and Testing

977 71

28

28

2

9% 31%

60%

1st cleaning

Seed lot composition before cleaning

71

88

1

3

10

3

86

11

Trash seed

97

2nd cleaning

1 24

75

100

100

Clean seed

100

Empty seed and light debris Insect damaged seed and heavy debris

100

Filled seed

Fig. 6 Progressive cleaning of seed lot. Each cleaning gives four fractions with increasing purity. Fractions with mostly debris are thrown away, while fractions with only little debris are kept. Purity in this example is less than 100 %

some are relatively small seeded, the general removal of small seed may lead to elimination of small-seeded families and hence reduction of the effective population size (Sorensen and Campbell 1993; Silen and Osterhaus 1979). If elimination of relatively small seed is considered relevant to improve vigor, grading should preferably be done on individual trees’ seed lots before bulking (Fig. 7).

Adjusting Moisture Content For orthodox seed, the lower the moisture content, the longer the potential storage period (Roberts 1973). For dry fruits, drying is an integrated part of seed extraction (section “Seed Extraction”), but for long-term storage, additional drying may be needed. Orthodox seed extracted by wet extraction (e.g., berries, drupes, or sarcotesta seed) will be very moist after extraction and must subsequently be dried. Recalcitrant seed may be dried to lowest safe moisture content for shortterm storage or maintained at higher moisture content to allow them to germinate. Exact moisture content measurement by oven-dry method (section “Determining the Moisture Content”) takes too long time to measure to be practical for processing purposes. Seed moisture meters are convenient for immediate results, but except from a relatively small number of small- to medium-size seed, they are difficult to calibrate for forest seed. However, with one moisture content analysis, and

978

L. Schmidt

Fig. 7 Recalcitrant seed storage. If you cannot prevent germination, let them germinate. Here Lithocarpus lucidus

considering that weight loss during drying is due to water loss, it is possible to estimate seed weight at a predecided moisture content (Box 2). Seed moisture is in equilibrium with air humidity. That means that water will escape from seeds as long as relative air humidity is lower than the equilibrium seed moisture content. If air humidity is 80 %, the equilibrium moisture content is about 12–15 % (depending on seed type and storage material). Once that has been reached, seed cannot be dried any further unless air humidity is reduced, which can be done in one of the two ways: 1. Increasing air temperature. Since hot air can contain more moisture than cold air, the relative humidity is reduced. 2. Removing humidity from the drying air. In dehumidifiers, air is cooled down to make water condensate. When it is heated up again, it contains less relative humidity at the same temperature. For small quantities, e.g., laboratory or test quantities, water is usually absorbed from the air surrounding the seeds by a high absorbent material like silica gel. Orthodox seed that cannot be dried because of too high humidity is metabolically active and will behave like recalcitrant seed, i.e., will sooner or later start to germinate (unless some dormancy regulations prevent them from doing so). If they cannot be dried, it is best to let them germinate at high moisture content, since without germination, seed will rapidly reduce viability.

Forest Seed Collection, Processing, and Testing

979

The base conditions and rate of drying can have an influence on tolerance. Orthodox seeds with high moisture content are more temperature sensitive than dry seed. Direct sun drying should thus be avoided for very moist seed. On the other hand, desiccation damage appears to be smaller for recalcitrant and intermediate seed that is dried fast (Sacande et al. 2004; Peran et al. 2004; Pammenter and Berjak 1999). Adjusting moisture content usually means drying, but in case of too dry recalcitrant seed, rewetting may be relevant. As desiccation damage usually occurs fast, possible rewetting should be done before any irreversible damage has occurred. Desiccation-sensitive seeds tend to store better either at lowest safe moisture content (LSMC) or fully imbibed; intermediate moisture content tends to be less suitable for the maintenance of viability (Walters et al. 2001).

Box 2: Using Target Moisture Content in Practice

1. Moisture content is calculated on a seed sample using the standard formula: Þ x 100 , where MC = moisture content, IW = initial MC ð%Þ ¼ ðIWODW IW weight, and ODW = oven-dry weight. 2. The target moisture content (TMC) is decided (e.g., safe for storage), and target weight is calculated as 100IMC TW ¼ 100TMC x IW, where TW = target weight of seed (at identified moisture content), IMC = initial moisture content, TMC = target moisture content, and IW = initial weight of seed. Example: Initial moisture content of a seed lot is measured before drying. Initial (fresh) weight (IW) = 90 g. Sample weight after oven-drying (ODW) = 77 g. Moisture content (MC) is calculated as

MC ð%Þ ¼

ðIW  ODW Þ x 100 ð90 g  77gÞ x 100 ¼ ¼ 14:4 % IW 90 g

It is decided to dry the seed to a moisture content of approximately 6 %. The target weight (TW) is thus TW ¼

100  IMC 100  14:4 x IW ¼ x 90 ¼ 82 g 100  TMC 100  6

A sample of 90 g of seed will thus have a moisture content of 6 % if dried down to 82 g. OBS: The seed sample (here 90 g) must be kept in a net bag or the like during drying under the same conditions as the bulk seed. The bag is weighed regularly during seed drying, and drying is concluded when the target weight has been reached.

980

L. Schmidt

Storing the Seeds The main purpose of seed storage is to establish a buffer between seed collection and planting. Storage may be short term between collection season and sowing season, or it may be long term covering several years’ consumptions and thus make annual seed collection expeditions unnecessary. The latter may be a necessity, for example, for periodic reproducing species or for expensive collections such as provenance or conservation collections to remote areas (Huth and Haines 1996). For orthodox seeds with long storage potential, the envisaged storage period should be included in the processing plan: drying to very low moisture content and cool storage essentially adds to the processing and storage costs, and such preparation and conditions are unnecessary for short-term storage of seeds that will be sown within short time. Hence, seed lots may sometimes be divided into short- and longterm storage.

Storing Orthodox Seeds Orthodox seeds are desiccation tolerant to at least to 6–7 % fresh weight basis. At that moisture content, seeds are metabolically dormant, i.e., there is no respiration or any other life functions, like a switched off engine with no fuel consumption but maintaining all the potential to run. But certain aging inevitably takes place in seed during storage. Generally, the lower the storage moisture content, the slower the deterioration and the longer the potential viability. Cold storage also prolongs viability. In addition, cold storage prevents insect activities, which may be destructive under ambient conditions. Since moisture content is in equilibrium with air humidity, seeds may reabsorb moisture from the air. Seeds with high moisture content or surrounded by humid air are exposed to fungal attack. Therefore, they should be stored in sealed plastic bags after drying (Fig. 8). Most orthodox seed stored at ambient temperature with a moisture content of 6–7 % will remain viable for at least 1–2 years. For longer-term storage, mc may be reduced to 5 % or lower, and storage under cold room (5–7  C) or deep freezer conditions may prolong viability for several decades. Decline in seed lot viability means that some seeds in the lot have deteriorated beyond repair. Viable seeds are able to germinate under optimal conditions (test conditions) even if also they have undergone some aging. However, if viability declines to say 50 % of the original, it is recommended to discard the whole seed lot since the 50 % deemed viable may show very low germination under field condition (reduced vigor). Storing Recalcitrant Seeds Recalcitrant seed will not survive desiccation to low moisture content where metabolism ceases, yet most recalcitrant seed undergoes some degree of maturation drying (Berjak and Pammenter 2002). Recalcitrant seed represents a continuum where some species will lose viability in few days, while others may be stored at “lowest safe moisture content” for several weeks or even months. Lowest safe moisture content varies from more than 40 % through intermediate seeds of

Forest Seed Collection, Processing, and Testing

981

Fig. 8 Vacuum packed storage of dry orthodox seed. CATIE Costa Rica

15–20 % to true orthodox seed that tolerates less than 5 % mc (Berjak and Pammenter 2002, 2008).

Storage Facilities Seed storage typically fluctuates seasonally with large space needed immediately after seed harvest and minimum space at the onset of the normal nursery sowing season or, supposed seed being dispatched throughout the year, immediately before new seed harvest. For most species, the major part of seeds is dispatched within 1 year; only irregularly collected species, e.g., from remote provenances (including conservation species) or periodic fruiting species, are collected and dispatched with longer standard annual cycles. Cold storage is relying on electric cooling devices, where every degree of temperature reduction and every cubic unit of cooling space cost electricity and money. Some practical measures can be used to reduce storage cost in general and cold storage in particular: 1. Avoid unnecessary excess seed collection; seed viability always declines during storage, and for common species, it is difficult to sell old seed if fresh seed is available. 2. Use 3–4 levels of storage temperatures: I. ambient storage, II. cool storage in air-conditioned room (18–22  C), III. refrigerated storage (household refrigerators or “walk-in” refrigerated storage rooms, 5–8  C), and IV. deep freezer (household freezer, 18  C to 20  C). Use only low temperatures where necessary for long-term storage. 3. Use storage capacity fully. Construct walk-in cool store rooms with compartments that can reduce the total cooled space during off season or construct

982

L. Schmidt

smaller cool rooms for long-term storage. For over-season storage, move to refrigerators or small cool rooms and switch off unused cold room. 4. Use minimum space and good insolation for walk-in cool stores. Store rooms are supplied with shelves for storage of seed containers on each side and a walking area in the middle. Walls and ceilings should be covered with 10–15 cm thick insolation material (mineral wool, Styrofoam, or the like with a low thermal conductivity (k). The smaller the k value, the larger the corresponding thermal resistance (R) value). 5. Construct cooled rooms or place cooling devices in basements of buildings and establish natural shade (trees) around the building to reduce air temperature. For long-term storage, store seed at highland stations with natural lower temperature. Since storage requirements for desiccation sensitive must allow some respiration to maintain viability, they are preferably stored (for short time) separate from orthodox seed. Net trays are suitable because seeds can be sprayed in case of desiccation, and roots and shoots of pre-germinated seeds are not harmed.

Storage Containers Storage containers are used for isolating seeds from adverse storage conditions and contaminants, i.e., maintaining conditions (moisture content, purity, away from pest and pathogens) achieved during processing. Since dry orthodox seeds do not have respiration, they are best stored in airtight containers. Nowadays there are multitudes of plastic storage bags and containers of different sizes available in almost any medium to large town of the world. Many are suitable for seed storage. However, a few considerations are appropriate: 1. Household “zipper bags” are suitable to keep seeds dry, but many insects can penetrate the material. They are also prone to mechanical damage during handling. Hence, rather thick material is recommended. 2. Seeds containing insects that may continue their activity during storage may be stored in CO2 directly in the sales quantities (Sary et al. 1993) in strong sealable plastic bags of standard consignments of weight, e.g., 50 g, 100 g, 200 g, etc., depending on seed size and usual customer requests. In order to avoid mechanical damage, the bags should be stored in containers (preferably flat, wide boxes where it is easy to see the content without emptying the whole container). 3. Large seed or large quantity of seed may be stored as loose, bulked seed in large containers. Upon seed order, any quantity of seed can be weighed out and dispatched. Seeds are usually stored in medium-size containers (1–5 l depending on seed size) to avoid them being opened too frequently and hence potentially being exposed to moist air. Because of their different storage physiology, desiccation-sensitive seed cannot be stored airtight. Hence, containers should be open to allow adequate air exchange,

Forest Seed Collection, Processing, and Testing

983

yet preventing excessive drying. Most species can be stored in loosely closed (but not tied) plastic bags for short time. However, conditions should be inspected regularly.

Seed Testing Testing of tropical forest seed does not principally differ from testing of horti- or agricultural seed or temperate seed for that sake, in terms of parameters. The standard parameters of all seed testing are seed weight, purity, moisture content, and viability or germination. However, tree fruit/seed morphology and physiology sometimes show forms and properties not common in other seed types. For example, large winged or fleshy fruit and seed are more common in forest trees because their base point of dispersal is at a certain height. Many tree fruits and seeds are large, and recalcitrance/desiccation sensitivity is much more common in tree crop seed. Because forest seed is used in much smaller quantity than agricultural seed, simply because the plant progeny develops into a larger plant, sampling of forest seed often differs. Eventually, collection methods of forest seed are quite different from collection of other seeds. All these factors imply some special adaptation of tree seed sampling and testing. Seed testing consists of a series of standard tests (purity, seed weight, moisture content, viability/germination) on a sample of a seed lot (ISTA 2006; AOSA 2014). Two of the tests, viz., purity and seed weight, are nondestructive, meaning that other tests can subsequently be carried out on the same seeds. Moisture content and germination/viability tests are destructive, meaning that they cannot be used for another type of test subsequently. Hence, for practical matters, the order of tests is as follows: purity ! seed weight ! moisture content ! germination/viability. Each test procedure carries a certain number of replications, which allows calculations of statistical parameters of average and standard deviations. According to ISTA, the following number of replications applies: Purity: 2 Seed weight: 8 Moisture test: 2 Germination: 4 It is important that the conditions of the seed sample do not change prior to or during the procedure of the testing. The former applies mainly to packing and possible storage of samples prior to the test and the latter primarily to the duration of seed tests. Purity and seed weight give immediate result; moisture content implies oven-drying for 24 h; some viability tests (cutting, X-ray, and TTZ) show immediate or fast result, while germination tests may take from a couple of weeks to several months.

984

L. Schmidt

Seed Lot Seed testing refers to a seed lot which is a consignment of seed from the same seed source, harvested and treated as a unit, i.e., harvested within a short time and processed by the same method. A seed lot is normally referred to with a common seed lot number, which is an identity tag assigned in the field to a consignment seed, and to which a seed documentation form is filled in with details on seed source, number of mother trees, and treatment (ref. seed documentation). Each seedcollecting unit/company has its own seed lot numbering system.

Sampling Seed samples for testing should be a representative for the whole seed lot, so that any information from the sample can be taken as information of the seed lot (ISTA 2006; AOSA 2014; Morrison 1999). If it is not a representative, the result will not reflect the seed lot conditions, and the whole exercise would be in vain. If a seed lot is completely uniform, any sample from any place (bag, container) would be a representative. However, it is quite difficult to create and maintain uniformity even over short periods of time. Seeds that vary in size, shape, water content, or other distinctive characters will tend to stratify themselves during handling and maintain such stratification. For example, relatively heavy round seed will tend to collect at the bottom of a container, while relatively flat, lighter seed and debris would collect at the top. The larger the difference, the more pronounced the stratification. Any seed character correlated with the morphological stratification will thus also be stratified. Further, any seed will be affected by its surroundings. Hence, seeds that happen to be positioned close to the top may dry or absorb moisture at a different rate than seeds at the inside of the seed container. Eventually, large seed lots may be stored in different containers, which may imply slightly different storage conditions. Several comprehensive manuals exist on sampling procedures indicating both overall principles and details for a number of common species (ISTA 1986; Kruse 2004). The two ways to overcome the problem of heterogeneity in seed lots are mixing and multiple sampling. Usually both methods are applied. Mixing should use a method that diminishes stratification and promotes uniformity. For example, shaking a container typically aggravates stratification, and some types of mechanical mixers risk just to inverse stratification. The best mixing is usually achieved by turning seed lots upside down several times, for example, by shoveling. Taking subsamples from different places in the seed lot compensates for possible nonuniformity of the seed lot. The number of samples from each seed lot depends on the uniformity and seed lot size. If the seed lot is stored in several containers, samples should as a minimum be taken from three levels of each containers (top, middle, bottom). Sample size should correspond with the container size (larger samples from large containers). Special sampler probes are available for small

Forest Seed Collection, Processing, and Testing

985

seeds. Larger seeds are usually collected by hand. All subsamples are eventually poured into a common sample, which is mixed and possibly divided before tests are carried out (ISTA 2006). Standard seed testing recommends sampling sizes for a number of species (ISTA 2006). Most listed species are small to medium size. Sampling becomes more difficult for large seeds.

The Purity Analysis A purity analysis indicates how many percentages (on weight basis) of a seed lot consist of potentially germinable seeds of the target species. For example, a purity percentage of 90 % means that 90 % of the weight is potentially germinable seeds of the target species, and 10 % is impurities. Potentially germinable means that the seed is not broken to a degree, where it can obviously not germinate. The impurities may, in some instances, be broken down in “other seeds” and inert matter. The purity test is theoretically simple. The problem is (sometimes) to define what is a “pure” seed and what is inert matter. As a rule, covers and adherences (whether of fruit or seed origin), which are normally part of the seed, are considered part of the pure seed as long as they cover or adhere (Poulsen et al. 1998). Mechanical damaged or broken seeds are also considered “pure” as long as there is a fair chance that they will germinate. Hence, a “pure” seed is not an unambiguous size or entity. For example, a pure seed may be a whole fruit (with possible extra-fruit adherences like wings) or the morphological seed that can theoretically be extracted. In practice, purity tests are carried out by manually separating pure seed fraction and inert matter and weighing each fraction on a scale. It is customary to indicate any fraction with two decimals (see ISTA 2006). The purity is then calculated by calculating the weight of the pure seed as a percentage of the total sample, i.e.: pure seed % ¼

weight of pure seed fraction x 100 weight og total fraction

Pure seed is used in all subsequent seed testing methods, e.g., seed weight, moisture, and germination or viability test. Hence a seed that is considered “pure” (albeit possibly damaged) in the purity test should be weighed in seed weight test, oven-dried in moisture content analysis, and sown in the germination test. In this way, possible bias in purity tests will be corrected in, e.g., the viability/ germination test, when calculating how many seed (number) can be expected from a given seed lot (weight) (Fig. 9).

Weight Determination Seed weight (sometimes specified as pure seed weight, psw) is the average weight of seed in a seed lot. The pure seed definition is again used as the baseline.

986

L. Schmidt

Fig. 9 Pure seed definition. Seeds of Pterocarpus spp. are enclosed in a modified indehiscent pod, and seeds are extremely difficult to extract without damaging them. Fruits are sometimes de-winged to reduce bulk. Pure seed may thus cover the whole range from entire fruits to morphological seed. See also Fig. 5

There will normally be eight replications of about 100 seeds from where the average is calculated (ISTA 2006). In standard seed testing, seed weight is normally indicated in a number of seed per 100 g (for very small seeds sometimes per 10 g). In biological seed information, the weight of 1,000 pure seeds is preferred. The two figures are interchangeable: 1000psw ¼

1000 1000 ; mber of seeds per gram ¼ number of seeds per gram 1000psw

Even though seed size and morphology show only moderate variation within species, the measured seed weight in seed testing may vary considerably for some species. This is partly due to the variation in extraction (pure seed) and partly due to variation of moisture content. For example, in large-seeded legumes of Sindora and Afzelia, the aril weighs almost the same as the seed without aril. Both seeds with and without aril are considered pure seed, but a seed lot where the aril has been removed during extraction has only half the seed weight compared to those with aril still attached. Water weighs more than dry matter; hence moist seeds are heavier than dry seed. The influence of variation in mc upon weight is small for dry orthodox seed but can be considerable for recalcitrant seed with high natural moisture content.

Determining the Moisture Content Seed moisture content (mc) indicates how much water is in the seed. It is used to predict storability under a given set of conditions. In practice, mc is measured as weight loss after oven-drying at high temperature until all water has evaporated. ISTA recommends two replications of mc tests. For each sample, mc test is carried out as follows: 1. Weighing out sample of approximately 100 g 2. Cutting or grinding seed into smaller pieces 3. Weighing empty container

Forest Seed Collection, Processing, and Testing

987

4. Pouring in grinded sample and weighing exactly (together with container), weight indication with 2 decimal grams 5. Drying samples 18 h at 103  C in drying oven 6. Weighing dried samples (together with container) Moisture content (fresh weight basis) is calculated as follows: Moisture content % ¼

Water content ðweight before minus weight after ovendryingÞx 100 Weight before

Viability and Germination Test There is a slight difference between germination and viability although they both refer to life processes which should ultimately lead to germination of the seed. Germination is proven directly in germination tests, whereas viability indicates the probability that seeds are alive and germinable. Germination test should in that context be preferred since it gives the direct evidence of what we want to know. However, in many instances, viability tests are suitable, e.g.: 1. Recalcitrant seed where seeds are short-lived and the result of a germination test cannot be used as information of the seed lot since deterioration takes place during the test period 2. Seed with very long germination time (several months) 3. Seeds with complicated dormancy requiring lengthy pretreatment, e.g., afterripening

Laboratory Tests A standard germination test is carried out on (usually four replications) samples of 50 seeds. Germination should be conducted under optimal conditions of moisture and temperature, and seeds should receive optimal pretreatment before germination test to break possible dormancy (ISTA 2006). Under laboratory conditions, seeds are also kept free from pest and pathogens. This means that a germination test shows germination under optimal conditions, i.e., the germination potential (which is the only thing meaningful to test), which is most likely higher than will be achieved under nursery or field conditions. A standard test indicates germination percentage with one figure (% germination = average of the replications). Tree seeds are often quite heterogeneous and germination speed often much slower than crop seed. These factors often complicate the testing procedure. 1. The official seed germination criteria imply that seeds germinate and seedlings develop normally. Seeds must therefore be kept until early seedling stage where they have radicle and at least two unfolded leaves which can be evaluated. This can spread over many weeks, sometimes even months. For slow germinating and

988

L. Schmidt

growing seeds, it can be difficult to keep laboratory conditions optimal for prolonged periods. Often germination criteria are devaluated so that germination is considered complete once radicle and shoot have formed. Germination is recorded at intervals, typically every week. In practice, it may be necessary to remove germinated seedlings to avoid them from harboring diseases (typically fungi) during the germination period. 2. Germination test must be concluded after a certain duration. If germination still takes place at that time, the test will not capture the total number of germinable seed. It is therefore customary to examine non-germinated seeds after the test by a simple viability test (e.g., a cutting test). Since germination is counted regularly over a period, it is possible to calculate germination speed. This is done in one of the two ways: 1. The number of days taken to reach 50, 75, 90, or other definite percentage of the total final germination 2. Germination percentage after a certain definite time, e.g., 21 days

Germination Test Under Field Conditions Official seed tests are carried out under controlled conditions so that results are comparable and replicable. This is not possible under field conditions. However, there are two conditions where field germination tests may be applicable: 1. For very slow germinating and developing plants which are difficult to test under laboratory conditions, because they get infected by fungi 2. Recalcitrant seed which cannot be submitted to laboratory tests Field germination test has the advantage that it can be integrated with nursery raising seedlings, i.e., the produced plants can be potted and used for ordinary planting.

Germination Tests for Recalcitrant Seeds A test result should reflect the conditions of the product. However, since germination of recalcitrant seeds often declines so rapidly that a test result will only show the conditions as they were and not as they are at the end of the test, they are usually of limited value. However, a seed test can be an assurance in case customers complain about quality of already dispatched seed. A cutting test will often give a useful clue about viability (see below). Since these seeds have high moisture content, dead and live seeds are usually easy to distinguish (dead seed usually dry and shrunken).

Estimating Viability Viability is an assessment or a measure of whether a seed is alive or not without it manifesting its life processes of germination. Viability tests are used primarily to

Forest Seed Collection, Processing, and Testing

989

Fig. 10 TTZ test of seed of Samanea saman shows red staining of live tissue and no staining of dead tissue. Only fully stained seeds are considered viable (Photo from Yu and Wang 1996)

save time of the test, and there must thus be a reasonable correlation between estimated viability and actual germination in case the whole germination test was carried out. Viability tests are often used for forest seed because many of these seeds have very slow germination and complicated dormancies and for recalcitrant seed, viability declines so fast that a seed lot has deteriorated before the result of a germination test would be available. Viability tests are sometimes carried out on non-germinated seeds in a germination test to see if the seeds would be likely to germinate if given longer time. Viability tests include four main methods (Fig. 10): 1. Cutting test is a simple longitudinal cutting of seed to inspect the interior. Empty seeds, insect-infested seed, and rotten seed are easy to distinguish. In filled seed, the embryo is evaluated visually. With a bit of experience, it is possible to distinguish between healthy and dead embryos. Healthy embryos are milky white to slightly greenish; they are intact without insect bites, mold infection, or other visible damages. 2. Topographic tetrazolium (TTZ) is a chemical test that reveals live metabolic cells by red staining. Uniform bright red areas of the seed are considered alive; unstained (white) areas are dead. Failed staining of essential parts of the embryo (radicle, cotyledon, plumule) is interpreted as nonviable seed, while seeds may be viable even with some necrotic tissue of the cotyledon. Manuals on TTZ interpretation exist (Moore 1985; Yu and Wang 1996; Enescu 1991; AOSA 2010). A few problems remain on correlating TTZ viability test with germination: (1) Immature seeds are alive (staining) but not germinable. (2) Dormant seeds are alive (staining) but may not germinate due to intrinsic factors. (3) A few species (e.g., Madhuca) have natural red embryos which make it difficult to distinguish live and dead tissue. (4) Any respiring part including infecting organism (bacteria and fungi) may stain tissue red (but the trained eye can usually distinguish fungus red from plant tissue red).

990

L. Schmidt

3. X-ray is primarily used for two types of tests, viz., to distinguish (1) insectinfested seed from intact fresh seed and (2) filled seed (seeds with embryo) from empty seeds. It is also possible to see mechanical damage on embryos, e.g., breaking of cotyledons from the main axis. It is not possible to separate aged seed from fresh seed. A thorough description of X-ray on seeds is available in Simak (1991), Saelim et al. (1996), and Craviotto et al. (2004). 4. Hydrogen peroxide (H2O2) test is a type of fast-germination test, which is not carried to the stage of seedling development but evaluated as viable after radicle protrusion. Soaking in 1 % hydrogen peroxide helps in breaking some types of physiological dormancy and accelerates the germination process (Laedem 1984; Bhodthipuks et al. 1996). Since viability tests are evaluated visually on the individual seeds, they are difficult to apply on very small seeds like eucalypts (Boland et al. 1990).

Other Tests Health tests (phytosanitary, infection, and infestation) are not normally part of the routine test but carried out, e.g., in connection with trade, particularly export. Phytosanitary test examines possible infection by fungi and infestation by insects. X-ray can be used to reveal insect infections and microscopy to examine fungal infection. In case species identification of infective organisms is deemed necessary, samples are grown under conditions conducive to their development, so that the species can be identified, for insects on macro-morphology and for fungi by microscopy (Sutherland et al. 2002). Vigor tests are germination tests under stressed conditions, which are compared to tests under normal “optimal” conditions. The background is the theory that the ability to germinate under stressed conditions declines faster than the ability to germinate under optimal conditions. Stress (vigor) test can be carried out under suboptimal temperature regime, moisture, or physical conditions like growing through a hard crust (gravel layer). The strength of vigor tests is that they often show a more realistic germination under field conditions than the laboratory test, because field conditions are rarely optimal. The weakness is that it is often difficult to quantify a given stress condition; too high stress (e.g., very high or very low temperature) gives predictable zero germination (and is thus nor relevant). Only stress normally encountered in the field (e.g., temperature +/ 10  C compared to optimum) makes sense (AOSA 1983). Accelerated aging (AA) is a type of vigor test that examines the relative aging of a seed lot. The philosophy behind is that if seed life can be prolonged by desiccation and cold temperature, then the opposite, humid and warm, would lead to rapid seed deterioration. Following that argumentation, if two seed lots are exposed to AA for a certain period (typically a couple of weeks), an already aged seed lot would show a more rapid decline in germinability than a good seed lot (TeKrony 2005). AA can in theory be used to predict storage life of seed (Roberts 1973). However, it is still

Forest Seed Collection, Processing, and Testing

991

disputable whether seeds undergo the same type of aging during an AA process as the one normal for dry-stored seeds. For example, fungi often play a significant role during AA because the humid warm conditions are conducive for fungal development, but they are not active during normal aging in dry storage.

Using Test Results in Practice Seed test may be used to guide future handling and dispatch of seed. 1. Seed lots with low purity percentage may be upgraded by further cleaning, i.e., by removing impurities. 2. Seed lots with low seed weight may be upgraded by removing light seed during a seed cleaning procedure grading on weight (specific gravity cleaning) or size (sifting). 3. High moisture content of seed lots may lead to a decision of further drying. 4. Results of seed health test may lead to various pest and pathogen control measurements, e.g., storage in CO2 (insect-infested seed) (Sary et al. 1993) or treatment with fungicides (Mohanan and Sharma 1991). Small seed lots infected by fungi may be surface sterilized by lab methods, e.g., hydrogen peroxide (H2O2), sodium hypochlorite (NaHCl), or 75 % alcohol (Bonner et al. 1994). 5. Vigor test may be used to determine the relative deterioration of seed lots and hence influence the order of dispatch. Especially AA may be used to predict storage life of seed. 6. Seed lots with low germination percentage may be upgraded by cleaning if viability is linked to some physical differences that can be separated, e.g., insectinfested or light seed. If germination goes under a certain limit (e.g., for stored seed, half of the fresh weight germination percentage), the whole seed lot should be dispatched, because the low germination indicates that deterioration is far progressed. Seed buyers will often be reluctant to buy seed lots with very low germination.

References Anon (1995) A guide to good climbing practice. The Arboricultural Association, Stonehouse AOSA (1983) Seed vigor testing handbook. Contribution no 32 to the handbook on seed testing. Association of Official Seed Analysts, Beltsville AOSA (2010) Tetrazolium testing handbook. Society of Commercial Seed Technologists and the Association of Official Seed Analysts, Washington, DC AOSA (2014) Rules for testing seeds. Association of Official Seed Analysts, Ithaca, Washington, DC Apetorgbor MM, Turco E, Cobbinah JR, Ragazzi A (2004) Potential factors limiting viability of Milicia excelsa (Welw.) C. C. Berg seeds in plantation establishment in West Africa. Zeitschrift fur Pflanzenkrankheiten und Pflanzenschutz 111(3):238–246

992

L. Schmidt

Berjak P, Pammenter NW (2002) Orthodox and recalcitrant seed. In: Vozzo J (ed) Tropical tree seed manual, vol 721, Agriculture Handbook. USDA Forest Service, Washington, DC, pp 137–147 Berjak P, Pammenter NW (2003) Understanding and handling desiccation-sensitive seeds. In: Smith RD, Dickie JB, Linington SH, Pritchard HW, Probert RJ (eds) Seed conservation: turning science into practice. Royal Botanic Gardens, Kew, pp 415–430, Chapter 22 Berjak P, Pammenter NW (2008) From Avicennia to Zizania: seed recalcitrance in perspective. Ann Bot 101:213–228 Bhodthipuks J, Saelim S, Pukittayacamee P (1996) Hydrogen peroxide (H2O2) testing for viability of tropical forest tree seed. In: Bhodthipuks J, Pukittayacamee P, Saelim S, Wang BSP, Yu SL (eds) Rapid viability testing of tropical tree seed, vol 4, Training course proceedings no. ASEAN Forest Tree Seed Centre Project, Muak Lek, pp 11–15 Blair DB (1995) Arborist equipment: a guide to the tools and equipment of tree maintenance and removal. International Society of Arboriculture, Champaign Boland DJ, Brooker MIH, Turnbull JW, Kleinig DA (1990) Eucalyptus seed. CSIRO, Canberra Bonner FT, Vozzo JA, Elam WW, Land SB Jr (1994) Tree seed technology training course. U.S. Dept. of Agriculture, Forest Service, Southern Forest Experiment Station, New Orleans Cole RJ, Holl KD, Keene CL, Zahawi RA (2011) Direct seeding of late-successional trees to restore tropical montane forest. For Ecol Manage 261:1590–1597 Craviotto RM, Arango MR, Salinas AR, Gibbons R, Bergmann R, Montero MS (2004) A device for automated digital x-ray imaging for seed analysis. Seed Sci Technol 32(3):867–871 Critchfield WB, Little EL Jr (1966) Geographic distribution of the pines of the world. US Dept. Agric., For Service. Miscel. Publ. 991, Washington, DC Doust SJ, Erskine PD, Lamb D (2006) Direct seeding to restore rainforest species: microsite effects on the early establishment and growth of rainforest tree seedlings on degraded land in the wet tropics of Australia. For Ecol Manage 234:333–343 Ellis JB (1986) Quantifying seed deterioration. In: McDonald MB, Nelson CJ (eds) Physiology of seed deterioration, CSSA special publication no 11. Crop Science Society of America, Madison, pp 101–123 Enescu V (1991) The tetrazolium test of viability. In: Gordon AG, Gosling P, Wang BSP (eds) Tree and shrub seed handbook. International Seed Testing Association, Zurich, Chap 9 FAO (2006) Quality declared seed systems. FAO plant production and protection Paper 185, FAO, Rome, 242 pp FAO, DFSC, IPGRI (2001) Forest genetic resources conservation and management. Vol. 2: In managed natural forests and protected areas (in situ). International Plant Genetic Resources Institute, Rome FAO, DFSC, IPGRI (2004a) Forest genetic resources conservation and management. Vol. 1: Overview, concepts and some systematic approaches. International Plant Genetic Resources Institute, Rome FAO, DFSC, IPGRI (2004b) Forest genetic resources conservation and management. Vol. 3: In plantations and genebanks (ex situ). International Plant Genetic Resources Institute, Rome Finch-Savage WE, Blake PS (1994) Indeterminate development in desiccation sensitive seeds of Quercus robur L. Seed Sci Res 4:127–133 Gunn B (2001) Australian tree seed centre operations manual. Australian Tree Seed Centre, Canberra Gunn B, Agiwa A, Bosimbi D, Brammall B, Jarua L, Uwamariya A (2004) Seed handling and propagation of Papua New Guinea’s tree species. CSIRO Forestry and Forest Products, Canberra Haines R, Nikles G (1987) Seed production in Araucaria cunninghamii – the influence of biological features of the species. Aust For 50(4):224–230 Huth JR, Haines R (1996) The effects of de-winging seeds of hoop pines (Araucaria Cunninghamii) on seed viability and longevity. In: Yapa AC (ed.) Recent Advances in tropical

Forest Seed Collection, Processing, and Testing

993

tree seed technology and planting stock production (Yapa, AC ed.). Proc. Intl. Symp. ASEAN Forest Seed Centre, Muak Lek/Saraburi, pp 45–50 ISTA (1986) ISTA handbook on seed sampling. International Seed Testing Association, Zurich ISTA (2006) International rules for seed testing. International seed testing association, Zurich Karrfalt RP (2008) Seed harvesting and conditioning. Chapter 3 in. USDA. The woody plant seed manual. Agriculture handbook 727. United States Department of Agriculture, Forest Service, Washington, DC, pp 57–84 Kruse M (2004) ISTA handbook on seed sampling. International Seed Testing Association, Zurich Laedem CL (1984) Quick tests for tree seed viability. British Columbia Ministry of Forests and Lands Research Branch, Victoria Linington SH (2003) The design of seed banks. In: Smith RD, Dickie JB, Linington SH, Pritchard HW, Probert RJ (eds) Seed conservation – turning science into practice. The millennium seed bank project. Royal Botanic Gardens, Kew, pp 593–636 Masilamani P, Vadivalu KK (1993) Effect of seed extraction methods on germination and vigour of honey mesquite. Madras Agric J 84(8):512–514 Maurer KD, Bohrer G, Medvigy D, Wright SJ (2013) The timing of abscission affects dispersal distance in a wind-dispersed tropical tree. Funct Ecol 27(1):208–218 McLemore BF, Chappell TW (1973) Mechanical shaking for cones harmless to Slash Pines. J For 71(2):96–97 Mineau A (1973) A new machine for collection off the ground; the beech-mast ‘vacuum-cleaner. Bulletin Technique, Office National de Foret, France 5:21–23 Mohanan C, Sharma JK (1991) Seed pathology of forest tree species in India – present status, practical problems and future prospects. Commonw For Rev 70:133–151 Moore RP (1985) Handbook on tetrazolium testing. International Seed Testing Association, Zurich Morrison RH (1999) Sampling in seed testing. Phytopathology 89:1084–1087 Ogunnika CB, Kadeba O (1993) Effect of various methods of extracting on germination of Gmelina arborea seeds/fruits. J Trop For Sci 5(4):473–478 Pammenter NW, Berjak P (1999) A review of recalcitrant seed physiology in relation to desiccation tolerance mechanisms. Seed Sci Res 9:13–37 Patterson B, Vaillancourt RE, Potts BM (2001) Eucalypt seed collectors: beware of sampling seedlots from low in the canopy. Aust For 64:139–142 Pazos GE, David F, Greene DF, Katul G, Bertiller MB, Soon MB (2013) Seed dispersal by wind: towards a conceptual framework of seed abscission and its contribution to long-distance dispersal. J Ecol 101:889–904 Peran R, Pammenter NW, Naiker J, Berjak P (2004) The influence of rehydration technique on the response of recalcitrant seed embryos to desiccation. Seed Sci Res 14:179–184 Poulsen KM, Parratt MJ, Gosling PG (eds) (1998) ISTA tropical and sub-tropical tree and shrub seed handbook. International Seed Testing Association, Zurich Riley JD, Craft IW, Rimmer DL, Smith RS (2004) Restoration of magnesian limestone grassland: optimizing the time for seed collection by vacuum harvesting. Restor Ecol 12(3):311–317 Roberts EH (1973) Predicting storage life of seeds. Seed Sci Technol 1:499–514 Sacandé M, Jøker D, Dulloo ME, Thomsen KA (2004) Comparative storage biology of tropical tree seeds. IPGRI, Rome Saelim S, Pukittayacamee P, Bhodthpuks J, Wang BSP (1996) X-radiography testing for viability of tropical forest seed. In: Bhodthipuks J, Pukittayacamee P, Saelim S, Wang BSP, Yu SL (eds) Rapid viability testing of tropical tree seed, vol 4, Training course proceedings no. ASEAN Forest Tree Seed Centre Project, Muak Lek, pp 17–31 Sary H, Yameogo CS, Stubsgaard F (1993) The CO2 method to control insect infestation in tree seed, Technical note no 42. Danida Forest Seed Centre, Humlebaek Schmidt L (2000) Guide to handling of tropical and subtropical forest seed. Danida Forest Seed Centre, Humlebaek

994

L. Schmidt

Schmidt L (2008) A review of direct sowing versus planting in tropical afforestation and land rehabilitation, Development and environment, nr. 10. Forest & Landscape Denmark, Copenhagen University, Hørsholm Silen R, Osterhaus C (1979) Reduction of genetic base by sizing of bulked Douglas-fir seed lots. Tree Planters Notes 30(1):24–30 Simak M (1991) Testing of forest tree and shrub seeds by X-radiography. In: Gordon AG, Gosling P, Wang BSP (eds) Tree and shrub seed handbook. International Seed Testing Association, Zurich, Chap 14 Sorensen FC, Campbell RK (1993) Seed weight – seedling size correlation in coastal Douglas-fir: genetic and environmental components. Can J For Res 23:275–285 Srimathi P, Ramadane T, Malarkodi K, Natarajan K (2003) Seed extraction in Jamun (Syzygium cumini Skeels). Progress Hortic 35(2):221–223 Sutherland JR, Diekmann M, Berjak P (2002) Forest seed health, IPGRI technical bulletin no. 6. IPGRI, Rome TeKrony DM (2005) Accelerated aging test: principles and procedures. Seed Technol 27(1):135–146 Walters C, Pammenter NW, Berjak P, Crane J (2001) Desiccation damage, accelerated ageing and respiration in desiccation tolerant and sensitive seeds. Seed Sci Res 11:135–148 Yu SL, Wang BSP (1996) Tetrazolium testing for viability of tree seed. In: Bhodthipuks J, Pukittayacamee P, Saelim S, Wang BSP, Yu SL (eds) Rapid viability testing of tropical tree seed, Training course proceedings no 4. ASEAN Forest Tree Seed Centre Project, Muak Lek, pp 33–58

Trade and Transfer of Tree Seed Lars Schmidt

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 996 Dispatch of Seeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 997 Centralized and Decentralized Forest Seed Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 999 Legal Regulations Regarding Forest Reproductive Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1001 Annex: Seed Lot Information Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1002 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1003

Abstract

When seed producers and seed users are geographically or functionally separated, seeds are transferred from producers to users. In market-oriented systems, transfer includes the pricing of seed, which reflects the procurement cost and seed quality. Physiological quality is documented via the seed testing records. Genetic quality is documented as documents on origin or seed source. New types of tree planting by smallholders imply special problems in distribution and supply systems since production systems for tree seeds have large areas while many consumers have small space for planting. A centralized forest seed supply contains large central units with good facilities for production and procurement but is far from seed users. Alternative decentralized systems with many small producers may have problems meeting high standards of seed quality and dealing with central regulations.

L. Schmidt (*) University of Copenhagen, Copenhagen, Denmark e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_191

995

996

L. Schmidt

Keywords

Seed trade • Seed transfer • Transport • Shipment • Seed supply system • Desiccation-sensitive seed • National seed centers • Legal regulations • Phytosanitary certificate • Ownership of genetic material • Formal system • Informal systems • Quality declared seed

Introduction The use or selection of species for various plantation programs is often influenced by the availability of seed and the ease of handling and propagation. In spite that the collection and propagation cost may make up only a fraction of the total plantation cost, tree planters often tend to select what is cheap and easiest. Seed supply may easily be caught in a vicious cycle: no seed sources => no seed => no planting => no seed sources. Opposite to the progress in seed handling, a network of adapted seed sources and effective distribution chain to end users can contribute to make species accessible and worth growing for those with land to plant. In this way, forest seed supply is a keystone to environmental rehabilitation, biodiversity conservation, and economic progress. Trade and transfer happen when producers and consumers (users) of seed are different people. Sometimes they are linked by middlemen, such as seed suppliers, who physically move seed in the demanded quantity from producers to consumers. Since the quantity of forest seed is relatively small compared to agriculture and horticulture seed, forest seed producers and suppliers are often identical. The ultimate consumers are tree planters, e.g., farmers and plantation owners. However, since most trees are planted and not sown directly, a nursery segment typically occurs between the supplier and tree planter, sometimes being an integrated part of one or the other and sometimes being an independent plant producer. Production and consumption of forest seed (as most other goods) are in most countries (at least those with a mixed market economy) a complicated network, usually with many stakeholders at each level. Forest seed supply has much similarity with agricultural seed supply, but there are some fundamental differences: 1. Because trees have long juvenile periods and individual trees are large, seed sources with high-quality seed require large production areas, and once the trees start fruiting, the seed production is often very high. Forest seed production is thus often concentrated on relatively large production units of plantation owners. 2. Consumers with limited land areas at their disposal use small quantity of plants of any species and very few tree seed. Plantation owners use large quantities but sometimes irregularly depending on plantation rotation.

Trade and Transfer of Tree Seed

997

Dispatch of Seeds Seed should reach end users with minimum loss of physiological quality (germination capacity). Shipping implies a risk of physiological deterioration during transit, especially for sensitive seed, long distances, and poor transport conditions. Conditions as those used during seed storage should as far as possible be maintained during shipment. That means primarily dry and free from insects and pathogens. Shipment implies inevitably some mechanical handling of the consignment; package material that can withstand such handling is essential. Seed should be put in transparent sealed bags that allow inspection through the material. If the consignment is prone to inspection that implies opening bags, it is advisable to put the sealed bags into strong zipper bags that can easily be re-closed to avoid loss. For ordinary dry orthodox seed and relatively fast shipments, cold shipment is unnecessary. The main problem is to avoid overheating during transport and transits. Temperature inside car boots and non-air-conditioned store rooms under tropical suns can rise to critical high levels. Cool transport may be necessary during shipment of sensitive moist material. Cold consignment shipment is often available from specialized transport companies when necessary. Several goods (primarily food) are shipped cool, and some of that network may be used. Alternatively, sensitive material may be packed with dry ice (dry to prevent water from sipping out when melting). The duration of transport should, as a rule, be as short as possible. Depending on transport type, notification of the receiver on dispatch is often the most efficient way to get the seed fast “through the system.” Upon receipt, seed may be sown immediately or stored temporarily in cool store until sowing. All seed consignments should be documented (labeled) with seed lot number, collection information, and seed testing information (seed weight, purity, moisture content, and viability). Test data should, as a rule, not be older than 9 weeks (Karrfalt 2008). In case additional information is needed, the seed lot can be traced back to the seed source via the seed lot number. Since various types of seed dormancy may reduce germinability or the seed may require special germination conditions, it is customary to accompany a traded seed lot with germination and propagation instructions. End users may cover user groups from plantation owners to small farmers. Their demand varies in terms of species, quantities, economic resources, and technical background. They may also cover large geographical areas. In order to reach each customer group, the dispatch system must be able to address different demands, including giving technical instructions on different levels. A small quantity of seed supply to small farmers appears to be a widespread bottleneck in forest seed supply. As a group, small farmers are often of the largest stakeholders in tree planting, yet each planter has only a small demand. The pathway is the distribution of larger quantities to larger local centers (e.g., village nursery or agricultural centers) who

998

L. Schmidt

Fig. 1 Distribution of seeds of agroforestry species in small quantities. Left: local dealer sells the tree seeds in small bags together with propagation and growth information, Nepal (Photo by Iben Nathan). Right: seeds of Erythrina velutina distributed in “six packs” to farmers (Brazil)

may then further distribute to small farmers via local shops, mobile salesmen, or the like (Moestrup et al. 2007) (Fig. 1). International trade and transfer of important commercial species has nowadays, thanks to the development of long-distance communication via Internet, made it significantly easier to communicate requests and supply data across boundaries (WAC undated). The general globalization of economy with easy money transactions and cheaper air transport fees has made the exchange of most goods much smoother than a few decades ago. However, the international seed supply chain still suffers from some obstacles: 1. Exported and imported live material is often subject to regulation, to avoid both the import of potential dangerous pathogens (importing countries) and export of potential valuable genetic resources (exporting countries). As the potential value of genetic material (species) is a relatively new concern, rules on how to get these permits are often not clear. The materials for research purpose are often exempted from the strict regulations. 2. The transit and loading of material can be critical for short-lived material such as desiccation-sensitive seed. Much seed deteriorates “while waiting.” 3. The production of valuable and sought-after seed is usually in remote and difficult accessible areas. With the general shrinking of forest resources, the choice of seed sources for most species is getting smaller. So despite the improved transport, impediments such as ownership, regulations, and restrictions can make seed collections more difficult. 4. As subject to regulations and general pollicization of international relations, informal joint seed collection expeditions are becoming increasingly regulated.

Trade and Transfer of Tree Seed

999

5. Large seed specialist institutions such as Danida Forest Seed Centre, Commonwealth Forest Network, Australian Tress Seed Centre, and CIRAD-Foret who used to mediate international transfer of forest seed have either closed or the seed supply part has been heavily cut back. Few other organizations (e.g., WAC (ICRAF)) mediate international transfer.

Centralized and Decentralized Forest Seed Supply National seed centers with central seed supply and distribution were established with the rationale that the production of quality forest seed needs a large area network of seed sources and good processing and storage facilities. Secondly, quality seed supply requires the knowledge of improvement and tree breeding  activities that are closely linked to high-level research. However, while these centers have often proved effective on seed production (collection, processing, storage), they have also often had problems getting seed distributed especially to smaller consumers like farmers. National seed centers are also sometimes less attractive to large companies who often have their own independent seed supply system for their key species. Since the knowledge base on tree seed is often concentrated in national centers, the roles are often diverted between different activities: 1. 2. 3. 4.

Seed production, procurement, and supply Seed research and tree breeding Training, extension, and awareness raising Policy formulation and certification

Re. 1. Seed production pertains to the selection and establishment of seed sources (section 2); procurement pertains to collection, processing, and storage (section 5); and supply pertains to dissemination and seed sale for various plantation programs (section 8.1). These are largely commercial activities where centers are usually obliged to operate on market conditions by selling seed, paid for by consumers, which may be private (from smallholders to plantation companies) or public enterprises. Re. 2. Seed research and tree breeding are part of producing quality seed and should hence ideally be financed by seed sale (which should be more expensive for high-quality seed). However, only for very few short rotation species can research and breeding activities be expected to be implemented in better seed quality in the short term. External (e.g., government) funding for research activities may be obtained independent on commercial seed trade, but the product of, e.g., an advanced seed orchard is a seed-producing stand, which has a direct link to seed production and sale.

1000

L. Schmidt

Re. 3. Training, extension, and awareness are considered public duties fulfilling a national objective of increasing productivity by using the best genetic plant material. This is thus supposed to go beyond the narrow commercial interest of seed supply involving stakeholders not necessarily linked to the seed business yet indirectly an advertisement. Re. 4. Policy formulation and regulations aim at supplying public goods that are supposed to influence seed quality on a national level. This could include regulations on trade and transfer, certification of seed sources, and documentation schemes. Where national centers have been established in larger countries, it has usually been structured with a national center or main office and a number of smaller regional centers (e.g., National Tree Seed Centre, Tanzania, Indonesia Forest Seed Project). This has been with a logistic rationale that these countries contain large ecological variation, and because of the envisaged local adaptation, seed users are recommended to use seed sources that are relatively close to the planting site. Hence, in practice, most seed is collected and distributed by local/regional centers. A survey of the dispatch pattern from national and regional centers has usually shown that the “coverage” of seed dispatch is usually few 100 km around the centers with little supply to more remote regions. Keeping large quantities of seed in store for interim use may, from a seed supply point of view, be a sensible way of coping with biological fluctuations of seed availability and to rationalize collection. However, the price of running cold stores is high, and where power supply is unreliable as is the case in some tropical countries, they are not always very effective. Since recalcitrant seed cannot be stored under any conditions for a very long time, a swift and efficient communication system between seed users and suppliers, leading to higher degree of collection upon demand, may be a suitable way to minimize the requirement for seed storage. National seed centers in tropical countries have often been heavily supported by overseas donors and subsequently relying on governmental support (National Tree Seed Centre (Tanzania), Central Forest Seed Company (Vietnam), Kenya Forestry Research Institute (KEFRI)). It has often proven difficult just to cover operation costs  not to mention background research by seed sale alone. Customers are often skeptic to the claimed superiority of the seed center’s quality seed and unwilling to pay for it if there are other seeds on the market. Such other seeds may come from local small suppliers, who can collect seeds at low costs from public or private seed sources. Another problem of national tree seed centers is when they act in a mixed establishment with other private stakeholders. Here their roles in policy development may tend to favor own seed sources and seed supply without being founded on objective quality criteria.

Trade and Transfer of Tree Seed

1001

Legal Regulations Regarding Forest Reproductive Material Legal aspects pertaining to trade and transfer have various purposes: 1. Quality guarantee to seed users 2. Preventing pest and disease to be spread with seed 3. Protecting ownership of genetic material Re. 1. Quality guarantee is in some cases a mere legal verification of the seed documentation scheme pertaining to genetic, physical, and physiological seed quality. In particular, in connection with genetic quality (provenance, source, collection base), whose validity is hard to prove or disprove due to the often long generation time of trees, particular verification or certification schemes are pertinent. Certification schemes have existed for a long time in Europe and the USA (Mangold and Bonner 2008; OECD 2011), whereas only few tropical countries have implemented control of forest tree germplasm, and in case some certification system exist, it is normally limited to larger plantation species and not covering, e. g., agroforestry tree seed. A main problem for the implementation of high-quality seed certification schemes is that it requires several verifications and inspections on various levels, all of which adds to the price of the seeds and in several cases effectively becomes a barrier to those farmers that should be benefitted with the proven high-quality seed (Nyoka et al. 2011). In addition, genetic quality with reference to particular seed sources is documented for only few species and it is far from complete. A less formal system, quality declared seed (QDS), which aims at providing flexibility in implementation while still retaining the basic principles of quality assurance, which can achieve the confidence of seed users, has gained some support (FAO 2006). Re. 2. Phytosanitary precautions are a key concern for exotic species where natural controlling enemies are often absent and where pest and diseases can consequently spread rapidly and cause great loss. The problem is obviously only for the importing countries, but requirement compels exporting countries or seed suppliers to issue phytosanitary certificates which state that seeds are free from pest and pathogens. Importing countries would often require seed to go through quarantine where seeds are examined. Re. 2. Ownership to genetic material has appeared out of concern of “gene mining” where rich countries/companies may collect material from poorer countries and then market and profit from it without leaving any benefit to the country of origin. The concern is mainly pertaining to crop plants and plants with medical potential, where potential patentation might make it problematic for countries of origin to utilize resources in the future. These legal aspects are an obstacle to smooth seed transfer, which is further hampered by bureaucratic conditions.

1002

L. Schmidt

Annex: Seed Lot Information Outline Seed lot no.

Supplier

Species name (botanical):

Provenance name:

Common name : Country :

Seed source information Seed source location:

Region/state  'N/S, longitude:

Geographical coordinates: Latitude: Altitude:

Country  'E/W; UTM

Mean annual rainfall (mm):

Rainfall regime:

Summer

Y:

Uniform

Winter

Bimodal

pH:

Soil type: Stand type:

, X:

m.a.s.l

Plantation;

Natural stands

Seed source type:

Unclassified,

Provenance seed stand,

Type:

Selected stand,

Seed production area,

Breeding seedling seed orchard (BSO),

Farmland seed source

Seed orchard

Other information:

Collection data Collection date: Genetic representation: Number of parent trees collected from: Average spacing between parent trees (m): Phenotypic selection of seed trees: Selection criteria:

Height,

Straightness,

Yes

No

Branching habit,

Health,

Others,

Test results Date of (latest) test

Germination percentage: %

Purity: Moisture content:

Viability: %

Measured by:

1000 grain seed weight:

TTZ Cutting X-ray Other:

No. of viable seeds per gram:

Seed treatment Seeds treated with:

Pretreatment: Scarification, method and duration:

Date of treatment Stratification, method and duration

Recommended seed handling before sowing Soaking in water, duration:

Date :

Leaching, duration Manual extraction, method: Signature

Other, Inoculation:

Mycorrhiza, species / type: Rhizobium, species/ type: Frankia, species/ type:

Trade and Transfer of Tree Seed

1003

References FAO (2006) Quality declared seed systems. FAO plant production and protection paper 185. FAO, Rome, 242 pp Karrfalt RP (2008) Seed testing. Chapter 5 In: USDA. The woody plant seed manual. Agriculture handbook 727. United States Department of Agriculture, Forest Service, Washington, DC, pp 98–116 Mangold RD, Bonner FT (2008) Certification of tree seeds and other woody plant materials. In: Bonner FT, Karrfalt RP (eds) The woody plant seed manual, vol 727, Agriculture handbook. United States Department of Agriculture, Forest Service, Washington, DC, pp 117–124 Moestrup S, Schmidt L, Nathan I (2007) Guidelines for distribution of tree seed in small bags: small quantities and high quality. Center for Skov, Landskab og Planlægning/Københavns Universitet, Hørsholm. http://forskning.ku.dk/search/?pure=da%2Fpublications%2Fguideline s-for-distribution-of-tree-seed-in-small-bags(bde97030-a1c3-11dd-b6ae-000ea68e967b)%2Fe xport.html Nyoka BI, Ajayi OC, Akinnifesi FK, Chanyenga T, Mng’omba SA, Sileshi G, Jamnadass R, Madhibha T (2011) Certification of agroforestry tree germplasm in Southern Africa: opportunities and challenges. Agrofor Syst 83:75–87 OECD (2011) OECD forest seed and plant scheme. http://www.oecd.org/agriculture/code/ 47439648.pdf WAC (undated) Seed suppliers directories. http://www.worldagroforestry.org/our_products/data bases/tssd

Tropical Nursery Concepts and Practices Diane L. Haase, R. Kasten Dumroese, Kim M. Wilkinson, and Thomas D. Landis

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Propagation Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Minimally Controlled Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Semicontrolled Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fully Controlled Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Modifying Light and Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Growing Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Physical Properties of Growing Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical Properties of Growing Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biological Properties of Growing Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Growing Media Ingredients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Developing and Mixing Growing Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Treating Growing Media Ingredients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Testing Growing Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Containers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Container Characteristics Affecting Plant Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Economic and Operational Factors Affecting Container Choice . . . . . . . . . . . . . . . . . . . . . . . . . Ways to Preclude Problems If Using Polybags and Polytubes . . . . . . . . . . . . . . . . . . . . . . . . . . . Water Quality and Irrigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water Quality and Quantity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water Testing and Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1007 1007 1007 1009 1009 1010 1011 1011 1012 1012 1013 1016 1017 1018 1019 1020 1021 1023 1024 1025 1026

D.L. Haase (*) State and Private Forestry, USDA Forest Service, Portland, OR, USA e-mail: [email protected] R.K. Dumroese USDA Forest Service, Rocky Mountain Research Station, Moscow, ID, USA K.M. Wilkinson Gibsons, BC, Canada T.D. Landis Native Plant Nursery Consulting, Medford, OR, USA # Springer-Verlag Berlin Heidelberg (outside the USA) 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_142

1005

1006

D.L. Haase et al.

Determining When and How Much to Irrigate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Irrigation Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plant Nutrition and Fertilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sources of Mineral Nutrients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fertilizer Application Rates and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring Nutrition Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reducing the Environmental Effects of Fertilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Beneficial Microorganisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nitrogen-Fixing Bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mycorrhizal Fungi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problem Prevention and Holistic Pest Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1027 1028 1030 1031 1032 1034 1035 1035 1036 1037 1038 1039

Abstract

A tropical nursery produces quality plants by providing a favorable environment and meeting the plants’ needs. Nursery propagation structures are designed to mitigate the factors that limit plant growth on a given nursery site. The ideal types of nursery structures are determined by site characteristics, nursery objectives, crop size, species, length of crop rotation, and the number of crops grown per year. Growing medium and container type are also important considerations in crop production. The physical, chemical, and biological characteristics of a growing medium affects seedling health, root development, and growth, and influences nursery operations such as irrigation and fertilization needs. Growers often use different types of growing media for seed propagation, rooting cuttings, and for transplanting larger plants and many mix their own using a combination of organic and inorganic ingredients. The best containers increase seedling root health, encourage good form and shoot-to-root ratios, and lead to good outplanting performance. Different species will require different types of containers based on the types of leaves and root systems they possess. Sufficient quantities of good-quality water must be available throughout the year to supply all the various uses at the nursery. Irrigation system design and application must meet the needs of a diverse species of plants and cater to their changing needs during different phases in their growth and development. In addition to water, plants require adequate quantities of mineral nutrients in the proper balance for basic physiological processes and to promote rapid growth and development. Nutrition and outplanting performance can also be improved by using beneficial microorganisms such as nitrogen-fixing bacteria and mycorrhizal fungi in the nursery. Providing optimum growing conditions for nursery plants and practicing good hygiene and sanitation in the nursery are important tools to prevent problems with diseases and pests. Many problems are triggered by stresses that can be avoided or corrected by good horticultural practices. Monitoring the crop and keeping records enables early detection and treatment of problems that do arise. Keywords

Propagation Environment • Irrigation • Fertilization • Growing Media • Nursery Containers • Beneficial Microogranisms

Tropical Nursery Concepts and Practices

1007

Introduction The nursery environment is a place to start, grow, and protect locally adapted plants until they are healthy, strong, and large enough to meet the challenges of outplanting sites and achieve project goals (Evans 1996). The work nurseries do to produce high-quality plants for reforestation helps perpetuate these species. Good nursery work makes successfully establishing plants more effective, affordable, and more likely to happen, whether nursery clients are foresters, farmers, restoration ecologists, community groups, or others. Crop growth relies on optimal levels of light, water, nutrients, temperature, humidity, and space in the nursery. These levels are achieved in the nursery through good design of propagation environments, choices of growing media, and use of appropriate container shapes and sizes. Plant care includes water quality and irrigation management, plant nutrition, working with beneficial organisms, and holistic prevention and management of pests and diseases. This chapter includes key concepts and processes for growing tropical plants in nurseries based on proven techniques, practices, and the best science available at the time of this writing. An understanding of some of these concepts and principles will make it easier to operate a nursery successfully, to serve clients, and to meet project objectives in the field.

Propagation Environments Many environmental factors influence growth and production of nursery plants. Photosynthesis and transpiration are the primary processes affected by environmental factors. Photosynthesis is the conversion of atmospheric carbon dioxide to carbohydrates in the presence of chlorophyll, the green pigment in leaves, using the energy in sunlight. Photosynthesis is a “leaky” process because, to allow the intake of carbon dioxide, water vapor is lost through pores, or stomata, on the leaf surfaces, a process called transpiration. To maximize the photosynthesis necessary for plant growth, nursery managers must reduce the factors that limit photosynthesis and/or increase the factors that promote photosynthesis, usually accomplished by using propagation environments. Propagation environments can be as simple as a garden plot where water and fertilizer are applied, or as complex as high-tech greenhouses that also modify all atmospheric factors. Although some species are grown from seeds, others in the same nursery might have to be grown from rooted cuttings. So, a good tropical plant nursery should be designed with various propagation environments in which plants of similar requirements and growth stages can be grown.

Minimally Controlled Environments A minimally controlled environment is the simplest and least expensive of the propagation environments. The most common type is an open growing compound.

1008

D.L. Haase et al.

Fig. 1 Open compounds, like at this Republic of Palau forestry department nursery, are common in tropical climates (Photo by Tara Luna)

It consists of an area where plants are exposed to full sunlight and is usually nothing more than an irrigation system and a surrounding fence (Fig. 1). Nurseries use open compounds for plant propagation and for areas to expose crops previously grown inside structures to ambient conditions during hardening. Plants can be grown on elevated platforms, benches, or pallets to improve air pruning of the roots, or directly on a layer of gravel (to provide drainage) that is covered with landscape fabric (to control weeds). Irrigation is provided by handwatering, or by sprinklers for smaller containers or driplines for larger ones. The compound usually needs to be fenced to minimize animal damage, and, in windy areas, a shelterbelt of trees around the compound can protect from desiccation and improve irrigation coverage. Artificial ponds are another type of minimally controlled environment. They are used for growing riparian, coastal, and wetland plants and can be used to provide specific habitats for certain wetland plants, such as saline conditions for plants adapted to coastal habitats. Wetland ponds can be aboveground reservoirs, such as shallow tanks or cattle-watering troughs, or they can be constructed with heavy plastic liners either in an excavated area or at ground level using a raised perimeter. These simple propagation environments require only periodic flood irrigation; nutrients are often added as controlled-release fertilizer mixed with the medium or through the use of organic-based medium (described below).

Tropical Nursery Concepts and Practices

1009

Semicontrolled Environments Semicontrolled propagation environments modify only a few of the limiting factors in the ambient environment, such as heavy rains, wind, or excessive temperatures. Cold frames are the most inexpensive semicontrolled propagation structures and are easy to build and maintain. Because conditions inside can stay relatively warm and moist, cold frames can be used for seed germination or rooting cuttings. The cold frame needs to be built so that it is weather tight and so the top can be opened daily to allow for ventilation, watering, and the easy removal of plants. The cover must be able to be attached securely to the frame to resist wind gusts. Heavy plastic film is an inexpensive covering but usually lasts only a single season. Hard plastic or polycarbonate panels are more durable and will last for several years; recycled windows work well too. In the tropics, cold frames usually need to have shade cloth suspended above them to help moderate temperatures. In a cold frame, plants grow best at 18–29  C. If air temperature exceeds 29  C, the top must be opened to allow ventilation. Hoop houses and polyethylene (“poly”) tunnels are versatile, inexpensive options. They are usually constructed of semicircular frames of polyvinyl chloride (PVC) or metal pipe covered with a single layer of heavy polyethylene and are typically quite long. The cover on hoop houses is changed or removed during the growing season to provide a different growing environment, eliminating the need to move the crop from one structure to another. Shadehouses are the most permanent of semicontrolled propagation environments. In the tropics, shadehouses are commonly used to propagate plants under conditions of intense sunlight (Fig. 2a). When used for growing, shade houses can be equipped with sprinkler irrigation and fertilizer injectors. When the shade is installed on the sides of the structure, shadehouses are very effective at protecting crops from wind and therefore help to reduce transpiration. Shadehouses can also be built with local materials (Fig. 2b).

Fully Controlled Environments Fully controlled environments control most or all of the limiting environmental factors. These propagation environments have the advantage that most crops can be grown faster with more uniform quality than those grown in propagation environments with less control. Examples include growth chambers (high-cost option used almost exclusively for research) and greenhouses. Tropical nurseries with large forestry and restoration programs often make use of greenhouses. These benefits must be weighed against the higher costs of construction and operation. The more complicated a structure is, the more problems that can develop. All greenhouses are transparent structures that allow natural sunlight to be converted into heat. At the same time, greenhouses are poorly insulated and require specialized cooling and ventilation systems to regulate temperatures. Jacobs et al. (2014) describe engineering considerations necessary when designing greenhouses for tropical plant production.

1010

D.L. Haase et al.

Fig. 2 Shadehouses protect plants from intense sunlight, rain, or wind and can be constructed of wood frames with lath or metal frames with shadecloth (a) or created with locally available materials (b) (Photos by Tara Luna)

One specialized, fully controlled environment is for rooting cuttings, the most common type of vegetative propagation. These “rooting chambers” create specific conditions to stimulate root initiation and development. Because cuttings do not have a root system, rooting chambers must provide frequent misting to maintain high humidity to minimize transpiration. Root formation is stimulated by warm temperatures and moderate light levels; these conditions maintain a high level of photosynthesis. Therefore, many rooting chambers are enclosed with polyethylene coverings that, in addition to maintaining high humidity, keep the area warm. If the chambers are outside, the covering further protects cuttings from rain and drying winds.

Modifying Light and Temperature To reduce light intensity and the resultant heat, growers apply shadecloths to propagation environments. Shadecloths are rated by the amount of shade they produce, ranging from 30 % to 80 %. Black has been the traditional color because it is relatively inexpensive, but now shadecloth comes in white, green, and reflective metal. Because black absorbs sunlight and converts it into heat that can be

Tropical Nursery Concepts and Practices

1011

Fig. 3 Compared with white or reflective shadecloth, black shadecloth can absorb heat and radiate it into the propagation environment (Illustration by Jim Marin)

conducted into the propagation structure (Fig. 3), black shadecloth should never be installed directly on the covering of any propagation structure, but instead needs to be suspended above it to facilitate air movement. Although more expensive than black shadecloth, white or aluminized shade fabrics are better for tropical environments and will do a much better job of cooling the propagation environment while still keeping light levels high. Applying a series of shadecloths, each with a lesser amount of shade, during a period of time is a good way to gradually harden nursery stock and prepare it for outside conditions. Thermometers that record the maximum and minimum temperatures during the day are simple and economical instruments that can help growers monitor subtle microclimates within any propagation environment.

Growing Media A growing medium can be defined as a substance through which plant roots grow and extract water and nutrients. Selecting a good growing medium is fundamental to good nursery management and is the foundation of a healthy root system. Growing media for use in container nurseries is available in two basic forms: organic-based (sometimes called “soil-free”) and soil-based. Compared with soilbased media, organic-based media (a base of organic materials that may be compost, peat, coconut coir, or other organic materials, mixed with inorganic ingredients) promotes better root development.

Physical Properties of Growing Media Physical characteristics of a growing medium determine how well it holds water and allow gas exchange with roots. Four characteristics – water-holding capacity, aeration, porosity, and bulk density – are inter-related and greatly influence plant

1012

D.L. Haase et al.

development. In general, nursery managers want a medium with high porosity (micropores and macropores). Micropores increase water-holding capacity because they hold water against the pull of gravity until plants use it. Macropores are filled with air after excess water has drained away because of gravity and contribute to aeration of the medium by providing pathways for oxygen to reach roots and carbon dioxide generated by root respiration to dissipate. The ratio of micropores and macropores varies by the types and sizes of the growing medium ingredients (discussed below). Thus, for water-loving plants, a desired medium has more micropores to increase water-holding capacity, but a drought-tolerant plant, or cuttings, would be grown in a medium with more macropores to prevent waterlogging and rotting. Soils with more micropores usually have a higher bulk density (the weight per volume of the medium) whereas well-drained media, especially those with pumice or perlite, have lower bulk densities. A medium with low bulk density can be compressed into a container so that the macropores are destroyed, thus resulting in a higher bulk density with subsequently more micropores and a greater water-holding capacity. An ideal growing medium is lightweight enough to facilitate handling and shipping while still having enough weight (a sufficient bulk density) to provide physical support to the plant.

Chemical Properties of Growing Media The most important chemical properties of a growing medium are fertility, pH, and cation exchange capacity (CEC). Plants rely on the growing medium to meet their increasing demand for mineral nutrients for growth. Many nursery managers prefer media with inherently low fertility to discourage damping-off during the establishment phase and add soluble fertilizers to media throughout the remainder of the growing season, whereas others blend controlled-release fertilizer into the medium or top dress the medium during the growing season. If fertilizers are difficult to obtain or cost prohibitive, organic amendments such as manure or compost can be included in the growing medium. The availability of those nutrients depends on the pH of the growing medium (Fig. 4) and most plants tend to grow best at pH levels between 5.5 and 6.5. Exceptionally high or low pH levels also affect the abundance of pathogens and beneficial microorganisms. CEC refers to the ability of a growing medium to chemically hold positively charged ions. The CEC of a growing medium reflects its nutrient storage capacity, thus it provides an indication of how often fertilization will be required. Because nutrient leaching occurs during irrigation, container nurseries prefer a growing medium with a very high CEC.

Biological Properties of Growing Media Growing media ingredients may contain pathogenic bacteria or fungi; those that do should be sterilized or pasteurized before use (see below). Organic-based growing media are preferred in nurseries because they are generally pest free. Although peat

Tropical Nursery Concepts and Practices

1013

Fig. 4 The pH of the growing medium affects availability of all mineral nutrients. In organic-based growing media, maximum availability occurs between 5.5 and 5.6 (Illustration modified from Bunt (1988))

moss is not technically sterile, it rarely contains pathogens or weed seeds when obtained from reliable sources. Vermiculite and perlite are rendered completely sterile by hot temperatures ( 1,000  C) during manufacturing. Well-prepared composts are generally pest free because sustained, elevated temperatures during composting kill most pathogens.

Growing Media Ingredients Once the functions and characteristics of growing media are understood, an effective and affordable growing media can be developed. A typical growing medium is a composite of two or three ingredients selected to provide certain physical, chemical, and/or biological properties. Mixtures of organic and inorganic ingredients are popular because these materials have opposite, yet complementary, properties (Table 1). Here are some of the common components of growing media: • Peat Moss – The horticultural and uniform properties of Sphagnum peat moss make it the only peat moss used in plant nurseries. Most peat moss comes from Canada, some comes from New Zealand, and the one known tropical source is

1014

D.L. Haase et al.

Table 1 Different chemical and physical properties of some common materials used to create growing media (Buamscha and Altland 2005; Johnson 1968; Lovelace and Kuczmarski 1994; Newman 2007) Bulk Component density Organic ingredients Sphagnum peat Very low moss Bark Low Coir Low Sawdust Low Rice hulls Low Compost Variable Inorganic ingredients Vermiculite Very low Perlite Very low Sand Very high Pumice Low

Porosity: water

Porosity: air

pH

Cation exchange capacity

Very high

High

3–4

Very high

Low High High Low Variable

Very high High Moderate Moderate Variable

3–6 6–7 3–6 5–6 6–8

High Low Low Low High

Very high High Moderate

High High Very low

6–8 6–8 Variable

High Very low Low

Low

High

6–8

Low

Indonesia (Miller and Jones 1995). Therefore, it is expensive and problematic to import peat for most tropical nurseries. Some nurseries may use peat as a transition component, comparing peat’s properties to local materials, such as composts or coir, to develop local alternatives for growing media while moving forward with plant production. • Compost – Composts are an excellent sustainable organic component for any growing medium and significantly enhance the medium’s physical and chemical characteristics by improving water retention, aeration porosity, and fertility (Fig. 5a). Some compost has also been found to suppress seed- and soil-borne pathogens. Compost quality can vary considerably among source materials and even from batch to batch; each material and batch should be tested before general use. Nursery managers should be able to find a sustainable source of organic matter that can be composted and used as a growing media component. Mature compost should not produce an unpleasant odor or heat before incorporating into a growing medium. Landis et al. (2014a) provide a detailed description on producing compost for tropical nursery use. • Coconut Coir – A byproduct of processing coconut husks, coir, sometimes called coir dust or coco peat, has proven to be an excellent organic component (Noguera et al. 2000) and is readily available in some tropical locales (Fig. 5b). Physical and chemical properties of coir can vary widely from source to source (Evans et al. 1996). • Sawdust – Raw sawdust can negatively affect nutrient availability, especially nitrogen, but its properties can be improved with composting (Fig. 5c; Miller and Jones 1995). However, the inherent differences in chemical properties among tree species make the suitability of sawdust extremely variable.

Tropical Nursery Concepts and Practices

1015

Fig. 5 Common organic ingredients of growing media include coconut coir (a), sawdust (b), or composted tree bark (c) (Photos (a) and (c) by Tara Luna, photo (b) Thomas D. Landis)

• Rice Hulls – Rice hulls or husks have been used as a component for many years in Indonesia (Miller and Jones 1995). Several nurseries have used composted, screened, and hammer-milled rice hulls in place of composted bark (Landis and Morgan 2009). • Vermiculite – Vermiculite is a popular component because it has many desired qualities, such as a very low bulk density, an extremely high water-holding capacity, a neutral pH, a high CEC, and small amounts of potassium and magnesium. Vermiculite is sold based on its particle size, which determines the relative proportion of aeration and water-holding porosity. • Perlite – Perlite particles have a unique closed-cell structure so that water adheres only to their surface; they do not absorb water as peat moss, coir, or vermiculite do. Therefore, growing media containing perlite are well drained and lightweight. Perlite is also rigid and does not compress easily, which promotes good porosity. One safety concern is that perlite can contain considerable amounts of very fine particle sizes that cause eye and lung irritation during mixing. Wetting the material while mixing and wearing dust masks and goggles can reduce this risk. • Pumice and Cinder – The porous nature of pumice particles improves aeration porosity but also retains water within the pores. Pumice is the most durable of the inorganic ingredients and so resists compaction. Cinder is another type of volcanic rock and a common component in volcanic areas.

1016

D.L. Haase et al.

• Sand – The composition of sand varies widely. When considering if local sand is a suitable component, the type of sand and its particle sizes must be determined. For example, some silty river sands with small particle size can negatively affect growing media by making them excessively heavy and not contributing to improved aeration. • Field Soil – Field soil is not recommended for growing nursery plants. If circumstances require a nursery to include some field soil in their media while more affordable and sustainable alternatives are being developed, dark topsoil should be selected. Soil-based mixes are safest for transplant media when transplanting into larger containers, such as polybags or larger containers (>3.5 L). The properties of soil-based mixes make them unsuitable for smaller containers, and the risk of disease makes them unsuitable in media for germinating seeds or rooting cuttings. Soil should comprise no more than 10–20 % of the transplant media by volume although some nurseries use up to 30 %.

Developing and Mixing Growing Media Every nursery manager needs to be able to experiment and find suitable, local, affordable ingredients to create good growing media. Three general types of growing media are used in container nurseries: 1. Seed Propagation. For germinating seeds or establishing germinants (sprouting seeds), the medium must be sterile and have a finer texture to maintain high moisture around the germinating seeds. 2. Rooting Cuttings. Cuttings are rooted with frequent misting, so the growing medium must be very porous to prevent waterlogging and allow good aeration necessary for root formation. 3. Transplanting. When smaller seedlings or rooted cuttings are transplanted into larger containers, the growing medium is typically coarser. Because of the diverse characteristics of various growing media ingredients, a growing medium can be formulated with nearly any desired property. The physical, chemical, and biological properties of each growing medium strongly interact with nursery cultural practices, particularly irrigation, fertilization, and container type. When considering a new growing medium, first test it on a small scale with several different species and evaluate its suitability before making a major change to the whole crop. A variety of commercial mixes that feature combinations of organic and inorganic ingredients are available. Many brands also contain a wide variety of amendments including fertilizers, wetting agents, hydrophilic gels, and even beneficial microorganisms. Many media are formulated for crops other than tropical plants and may do more harm than good; always check the label to be sure of exactly what is in the mix. Many nursery managers prefer to mix their own custom growing

Tropical Nursery Concepts and Practices

1017

media. In addition to saving money, custom mixing is particularly useful in small nurseries where separate mixes are needed to meet propagation requirements of different species. Some media may also include amendments – supplemental materials that contribute less than 10 % of the mixture including fertilizers, lime, surfactants, hydrogels, and mycorrhizal inoculum. Some of these materials may be undesirable because they are formulated for other crops and are detrimental to native plant growth. If amendments are added to the growing medium, it is important that they be added uniformly and tested on a small scale before widespread usage. Whitcomb (2003) emphasized that improper media mixing is one of the major causes of variation in container plant quality. Small batches of growing media ingredients can be mixed by hand and larger batches can be mixed on any clean, hard surface using shovels. Some organic ingredients repel water when dry, so frequently misting the media with water at regular intervals during mixing improves water absorption. Do not compress or compact during mixing. Nursery managers that regularly require larger quantities of custom growing media should consider purchasing a mixer. A cement mixer works well as long as care is taken to avoid excessive mixing, which breaks down the size and texture of ingredients. When handling growing media, workers need to follow safety precautions to protect from dust and infections. Perlite dust is of particular concern because of potential for silicosis, an inflammation that occurs over time when dust that contains silica is inhaled into the lungs and open wounds should be covered to prevent infections.

Treating Growing Media Ingredients Some growing media ingredients may need to be leached, pasteurized, and /or screened before use to reduce potential damage to plants. Using fresh water to leach out salts may be necessary for materials such as coir, sand, sawdust from mills near the ocean, and composts with excessive soluble salt levels (Carrion et al. 2006; Landis and Morgan 2009). Pasteurization, especially of organic ingredients, can prevent the introduction of pests, weeds, and diseases into the nursery (Fig. 6). Most inorganic components are inherently sterile. Heat generated during the composting process will kill pathogens and other pests, but field soil should be pasteurized. Heat pasteurization is the most common way of treating growing media and includes moist heat from steam, aerated steam, or boiling water or dry heat from flame or electric pasteurizers or microwave ovens. Small pasteurizing equipment is available for nurseries and some nurseries have developed their own pasteurization process using fire or solar heat. Some ingredients, such as soil, sand, and cinder, may require screening or sifting to achieve the desired particle size. It may be necessary to sift twice, once with a small mesh to eliminate material larger than desired, and a second time with a larger mesh to remove material smaller than desired.

1018

D.L. Haase et al.

Fig. 6 Necessary temperatures for heat pasteurization vary depending on the target pest. Effective control is achieved if the target temperature is held for 30 min (Illustration from Landis et al. (2014a) by Jim Marin)

Testing Growing Media To preclude surprises, nursery managers test compost and growing media well in advance of use and retain the results to compare with new or experimental batches (Grubinger 2007) and to develop and refine suitable alternative mix(es) with similar favorable properties. One easy and effective test is a plant bioassay (Grubinger 2007). Put a sample of the growing medium in the containers that will be used in the nursery, sow an abundantly available, fast-growing species into the medium, and observe how the planting performs during a few weeks. If the mix works, it is ready to try in the nursery. The salinity (salt level) of the growing medium is a key parameter affecting the development and health of roots. Salts may come from growing media ingredients, irrigation water, and from added fertilizers. Routinely measuring electrical conductivity (EC) monitors the amount of nutrients and salts present to ensure they are in the appropriate ranges for the species grown (Table 2). Excessively high salt levels can damage or even kill succulent young plants. For more details on proper technique with EC meters, see Landis and Dumroese (2006). For more formal testing, growing media samples can be sent to a soil-testing laboratory (private, local extension office, or university) for testing. A measurement of pH, soluble salts (electrical conductivity), and nutrients should be requested (Grubinger 2007). Results can vary among laboratories depending on their

Tropical Nursery Concepts and Practices Table 2 Electrical conductivity (EC) guidelines for artificial growing media (Timmer and Parton 1982)

EC range (μS/cm) 0–12,000 1,200–2,500 2,500–3,000 3,000–4,000 >4,000

1019

Salinity rating general guidelines Low Normal High Excessive Lethal

μS/cm = microSiemens per centimeter

Fig. 7 Nurseries use a variety of containers to produce different species and stocktypes (Photo by Diane L. Haase)

procedures, so it is best to select one laboratory and use them for testing year to year, provided that the data appear accurate and consistent.

Containers A suitable container could be anything that holds growing media, drains, allows for healthy root development, does not disintegrate before outplanting, and allows for an intact, healthy root system to be removed with a minimum of disturbance to the plant. Most nurseries grow a wide variety of species and therefore several different containers are required (Fig. 7). In general, the following points hold true regarding container type: Plants that develop shallow, fibrous root systems, as most forbs do, grow better in shorter containers. Plants with long taproots, such as many kinds of trees, grow better in taller containers. And, plants with multiple, thick, fleshy roots, and species with thick, fleshy rhizomes grow better in wide containers. Many types of containers are available and each has its advantages and disadvantages concerning plant development, economics, and efficiency under operational conditions (Landis et al. 2014b). It is a good idea to try new containers for

1020

D.L. Haase et al.

each species on a small scale before buying large quantities. Several containers types are used in container plant nurseries and can vary considerably in attributes and size.

Container Characteristics Affecting Plant Development • Volume – Container volume dictates how large a plant can be grown in it and this varies by species, target plant size, growing density, length of the growing season, and growing medium used. Larger containers occupy more growing space and take longer to produce a firm root plug so therefore are more expensive to produce, store, ship, and outplant but the benefits, however, may outweigh the costs if the outplanting objectives are more successfully satisfied. • Height – Container height determines the depth of the root plug, which may be a consideration on dry outplanting sites (where a deep root system that can stay in contact with soil moisture is desired) or sites with shallow soils (where only a short root system can be planted). • Diameter – Broad-leaved trees, shrubs, and herbaceous plants generally need a larger container diameter so that irrigation water applied from above can penetrate the dense foliage and reach the medium. The container diameter must also be large enough to accept the seeds. • Shape – Containers are available in a variety of shapes and most are tapered from top to bottom. Most containers are round but some are square and maximize the growing space used in the nursery. Container shape is important as it relates to the type of outplanting tools used and the type of root system of the species grown. • Density – The distance between plants is important because it affects the amount of light, water, and nutrients that are available to individual plants (Fig. 8a). In general, plants grown at closer spacing grow taller and have less stem diameter than those grown farther apart (Fig. 8b). Plant leaf size greatly affects growing density. Broad-leaved species grow better at fairly low densities, whereas smaller leaved and needle-leaved species can be grown at higher densities. Trays holding individual containers provide some flexibility in density because, as the plants grow, containers can be rearranged to allow greater space among plants (Fig. 8c). • Root Control – Some plants have aggressive roots that quickly reach the bottom of the container and may spiral or become rootbound. Many containers have vertical ribs to force the roots downward and prevent spiraling. Chemical pruning involves coating the interior container walls with chemicals that inhibit root growth. Several companies have developed containers that feature air slits on their sides to promote pruning and mitigate root deformation (Fig. 9). • Drainage – Containers must have a bottom hole or holes large enough to promote good drainage and encourage “air pruning.” The drainage hole must also be small enough to prevent excessive loss of growing medium during the container-filling process. • Color and Insulation – Color and insulating properties of the container affect medium temperature, which directly affects root growth. Black containers can

Tropical Nursery Concepts and Practices

1021

Fig. 8 Next to volume, density (spacing) is the most important characteristic in container choice (a). Plants grown too close together become tall and spindly and have less stem diameter (b). Trays with removable containers are popular because they allow flexibility in spacing between plants (c) (Adapted from Dumroese et al. 2008)

quickly reach lethal temperatures in full sun whereas white ones are more reflective and less likely to have heat buildup.

Economic and Operational Factors Affecting Container Choice • Cost and Availability – In addition to the purchase prices, remember to consider associated expenses for various container types, such as shipping and storage costs. Nursery managers in the tropics often face high shipping costs and long

1022

D.L. Haase et al.

Fig. 9 Plants with aggressive roots often exhibit spiraling and other deformities after outplanting. If rootbound, roots often do not grow out beyond the original plug (a). Containers coated with copper will chemically prune roots (b) and other containers are available with lateral slits to reduce spiraling and encourage air pruning on the side of the plug (c). (Illustrations adapted from Dumroese et al. 2008).

shipping times. Also consider the potential for long-term availability to ensure that ample supplies can be secured in the future. • Durability and Reusability – Containers must maintain structural integrity and contain root growth during the nursery period. The intense heat and ultraviolet rays in container nurseries can cause some types of plastics to become brittle and break, although many container plastics now contain ultraviolet (UV) inhibitors. While some containers are designed to be used only once, others can be reused for 10 or more crop rotations. The purchase cost of reused containers can be amortized over their life span after adjusting for the cost of handling, cleaning, and sterilizing the containers between crops. • Return from Customers – Reusing containers is important; it saves money and resources and protects the environment from waste. Charging a refundable container deposit (similar to bottle deposits for beverage containers) encourages clients to return containers to the nursery. All containers should be washed and sterilized before reuse in the nursery; even though disease symptoms may not be apparent, disease organisms accumulate in non-treated containers and reduce growth (Dumroese et al. 2002). • Ease of Handling – Containers are typically moved several times during crop production, so handling is a major concern from logistic and safety standpoints. Collapsible or stackable containers may have lower shipping and storage costs but labor is required to prepare them for filling and sowing. Large containers are increasing in popularity, but they become heavy when saturated with water. Weight must be considered for shipping and field planting.

Tropical Nursery Concepts and Practices

1023

• Ability to Cull, Consolidate, and Space – One advantage of tray containers with interchangeable cells is that cells can be rearranged. During thinning, empty cells can be replaced with those containing plants. During roguing, diseased or undesirable plants can be replaced with healthy plants. For species that germinate during a long period, plants of the same size can be consolidated and grown under separate irrigation or fertilizer programs. Thus, consolidation can save a considerable amount of growing space. Another unique advantage is that cells can be spaced further apart by leaving empty slots; this practice is ideal for larger leaved plants and for promoting air circulation later in the season when foliar diseases can become a problem.

Ways to Preclude Problems If Using Polybags and Polytubes Bags made of black polyethylene (poly) plastic sheeting are the most commonly used nursery containers in the world because they are inexpensive and easy to ship and store (Fig. 10a). Polybags often produce seedlings with poorly formed root systems that spiral around the sides and bottoms of the smooth-walled containers. This problem worsens when seedlings are held over and not outplanted or transplanted at the proper time. In cases in which converting to hard plastic containers would be operationally or financially impractical, ways exist to improve container production using polybags. Some of these cultural modifications include (Landis 1995): • Managing container seedlings as a perishable commodity with a limited “shelf life.” This concept is particularly critical in tropical nurseries where seedlings grow year round. If seedlings cannot be outplanted when their roots fill the container, then they must be transplanted into a larger container. Holding over polybag seedlings is not an option. • Using polytube containers (a polybag open at both ends, sometimes called a polysleeve) instead of polybags (Fig. 10b). These containers can usually be obtained from the same supplier as polybags or cut from a continuous roll with no bottom (Jaenicke 1999). Poly tubes eliminate much of the root spiraling. Poly tubes will hold growing media if they are properly filled and placed on elevated screen-bottomed trays to promote air pruning of roots (Fig. 10c). • Using copper-coated polytubes or polybags. Plants grown in copper polybags produce a much finer, fibrous, non-circling root system that is well distributed throughout the containers. • Switching from soil-based to “artificial” or organic-based growing media (based on composts, bark, or other materials instead of soil). • Carefully transplanting germinants or direct-seeding into the polytube containers to avoid root deformations.

1024

D.L. Haase et al.

Fig. 10 Polybags are inexpensive containers that can produce good plants (a) but root spiraling is often serious. Open-bottomed polytubes (b) in trays (c) can help solve that problem (Photos by Tara Luna)

Water Quality and Irrigation Water is the single most important biological factor affecting plant growth and health. Determining how, when, and how much to irrigate is a crucial part of nursery planning and day-to-day operations. Adequate watering is particularly important with container plants because they can dry out quickly, but excessive watering can lead to root disease and contribute to other problems with seedling growth.

Tropical Nursery Concepts and Practices

1025

Tropical nurseries typically grow a wide range of species with different water requirements, and these water requirements change as the plant moves through the three phases of growth (establishment, rapid growth, and hardening; described below). The nursery might have various propagation areas and corresponding irrigation zones that provide for the changing needs of plants during these phases. The best design for any irrigation system comes from understanding the needs of the plants, the factors that affect water availability, and the details of how, when, and why to water. Tropical plant nurseries can use water from several different sources, including rivers, ponds or reservoirs, rainwater, groundwater, and municipal sources. New nurseries need to evaluate the quantity, quality, and seasonal availability of all potential water sources. For surface or groundwater, a hydrologic survey and analysis of local water rights needs to be conducted before nursery development. Surface water sources that have flowed through agricultural land need to be tested for waterborne pests or herbicides and may need to be treated. Rainwater is an attractive source of high-quality water for tropical nurseries if enough can be collected from the roofs of buildings and stored in tanks until needed.

Water Quality and Quantity The amount of water to grow a crop varies tremendously between humid and arid locations. Remember that a nursery also needs water for operational requirements other than irrigating crop and that a nursery that starts small may choose to expand. Therefore, ensure an abundance of water is available to meet present and future needs. Even in cases with access to a steady, reliable, high-quality water source, an emergency backup system is always a good idea. A prudent investment is a backup water storage tank containing sufficient water to meet the nursery’s needs for at least 1 week. Water quality needs to be a primary consideration during nursery site evaluation. For irrigation purposes, water quality is determined by the quantities of salts and pests in it. A salt is defined as any chemical compound that releases charged particles (ions) when dissolved in water. Some salts are fertilizers while other salts can reduce growth or even cause injury or death. Excessive dissolved salts in irrigation water can clog nozzles and accumulate in growing media and eventually harm plant tissue. The most characteristic symptom of high salinity is reduced growth and burn or scorch of leaf margins or tips (Fig. 11). Excessive dissolved salts result from local climatic or geologic influences, saltwater intrusion, high fertilization rates, or poor irrigation practices. Test results for salinity are traditionally expressed as electrical conductivity (EC); the higher the salt concentration, the higher the EC (Table 3). The EC can be checked at the nursery using a conductivity meter, or by sending water samples to a local laboratory. Tropical nurseries that use irrigation water from surface water sources such as ponds, lakes, or rivers may encounter problems with pests: weeds, pathogenic fungi, moss, algae, or liverworts. Recycled nursery irrigation water should also be

1026

D.L. Haase et al.

Fig. 11 Agricultural water quality is determined by the level of soluble salts because they can accumulate and eventually “burn” seedling foliage (Photo by Thomas D. Landis)

Table 3 Water-quality standards for nursery irrigation water (Modified from Landis et al. 1989a; Robbins 2011)

Quality index pH Salinity (μS/cm) Sodium (ppm) Chloride (ppm) Boron (ppm) Fluoride (ppm) Iron (ppm)

Optimal 5.5–6.5 0–500 1.00a >1.00

a

Sensitive species may be damaged at lower levels

analyzed. Many weed seeds and moss and algal spores are small enough to pass through the irrigation system and can cause problems. Chlorination and some specialized filtration systems may remove many disease and pest organisms from irrigation water. Irrigation water, especially in agricultural areas, may be contaminated with residual pesticides, and these sources need to be tested for pesticide contamination during the nursery site selection process.

Water Testing and Treatments Water should be tested during the site selection process, once the nursery is established, and again at yearly intervals. A complete analysis of irrigation water consists of a salinity evaluation listing the concentrations of sodium, chloride, and boron, which are reported in parts per million (ppm) and the three standard waterquality indices: EC, toxic ion concentrations, and pH (Table 3). It should also be tested for the presence of pathogenic fungi during the site selection process and later if a problem is observed. Collect an irrigation water sample in a clean plastic bottle with a firm, watertight lid. Let the water run for several minutes and then rinse the sample bottle well before collecting the sample.

Tropical Nursery Concepts and Practices

1027

Establishing the nursery on a site with tested, good-quality water is the best way to preclude water-related problems. If existing water quality is poor, methods such as deionization and reverse osmosis can treat and improve irrigation water, but they are often prohibitively expensive and not feasible for most nurseries. To correct or safeguard against minor problems with otherwise good-quality water, however, chlorination and filtration are low cost and highly effective for container nurseries. Chlorination can kill fungi, bacteria, algae, or liverworts introduced through the irrigation system. A simple method is to mix household bleach (5.25 % sodium hypochlorite) at a rate of 18 ml per 1,000 L. This low dose (about 1 ppm) is not phytotoxic to a wide range of plant species (Cayanan et al. 2008) – but always test it first on a small subset before treating the entire crop. Filtration removes suspended or colloidal particles, thus preventing problems such as plugging or damaging irrigation equipment, as well as removing unwanted pests such as weed seeds or algae spores. Granular medium filters can remove fine sand or organic matter and are constructed so that they can be backflushed for cleaning. Surface filters include screens or cartridges of various mesh sizes to remove suspended material; screens must be physically removed and cleaned whereas cartridge filters are not reusable and must be regularly replaced. Handreck and Black (1984) recommend using filters small enough to remove particles greater than 5 μm in diameter, which will take care of most suspended materials.

Determining When and How Much to Irrigate When irrigating container crops, it is important to apply enough water such that some drips out the bottom of the container, but not so much that water streams out the bottom. The general rule is to apply approximately 10 % more water than is needed to completely saturate the entire growing medium profile during irrigation. It is absolutely necessary to regularly monitor the moisture status of growing media. In small containers, the limited volume of moisture reserves means that critical moisture stresses can develop quickly. Visual and tactile assessments are the most common method of monitoring irrigation effectiveness. Monitoring can also include formal or informal assessments of container weight. In addition, various tools, such as tensiometers, electrometric instruments, balances, commercial moisture meters, or pressure chambers can be used to monitor irrigation efficacy (Landis et al. 1989a). Any equipment-based method must also be supported by actual observation (visual and tactile) and the grower’s experience. A quantifiable technique, such as container weight (the difference between the container weight at field capacity and some target weight, for example 75 % of field capacity), ensures that proper irrigation frequency can be repeated by all staff for a particular crop for consecutive growing seasons (see Dumroese et al. 2012). The amount of irrigation to apply varies during the growing season because of the three stages of plant development: establishment, rapid growth, and hardening (described in the chapter, “▶ Planning and Managing a Tropical Nursery”).

1028

D.L. Haase et al.

The establishment phase is the time from when sown containers are placed in the growing area until true leaves form or cuttings begin to root. The growing medium needs to be brought to field capacity (some water dripping from the bottom). Thereafter, watering needs during establishment should be monitored carefully and tailored to the needs of the species. On the one hand, inadequate irrigations will allow seeds to dry out, decreasing germination success or causing total crop loss. On the other hand, excessive irrigation may create excessively wet conditions that promote damping-off and delay germination. Until the seeds germinate and begin to grow, water must be applied with the goal of replenishing the moisture in the thin surface layer of the medium. This practice is usually best accomplished by periodic misting or light irrigation with a very fine spray nozzle, which also protects germinating seeds from being moved or damaged by the force of the water. These fine sprays can also be used to control the temperature around germinating seeds; misting just enough to dissipate heat around the seedling. During the rapid-growth phase, the plant experiences a large increase in shoot size, which increases the amount of water lost through transpiration so irrigations must be longer and more frequent. Water use can double or even triple during the rapid growth phase. No plants should ever be allowed to dry out completely. Nursery managers need to be aware of the varying water requirements for different species and adjust irrigation practices accordingly. The rapid growth phase is also the time when liquid fertilizers are most concentrated and water loss through transpiration is high, so growers must monitor for salt accumulation. Once plants near the target size, the hardening phase begins. Manipulating irrigation frequency is an effective way to initiate the hardening of plants before shipment and outplanting. Because seedling growth is tied to moisture stress levels, growers can slow shoot growth and increase general resistance to stress by inducing mild water stress. This “drought stressing” procedure consists of withholding irrigation for short periods of time until the plants can be seen to wilt slightly or until some predetermined moisture stress is reached. For some species, this process may be repeated several times. After this stress treatment, the crop is returned to a maintenance irrigation schedule.

Types of Irrigation Systems The best method of applying irrigation water depends on the water requirements of the species being grown and on the size and complexity of the nursery (Table 4). In general, four methods (hand watering, microirrigation, overhead irrigation, and subirrigation) are the most commonly used, and individual nurseries may use one or more of them depending on the particular crops. Hand watering is often the most practical irrigation strategy for small nurseries, nurseries producing a wide diversity of species with radically different water requirements, or nurseries in the startup phase. Although the task may appear easy, it is challenging to uniformly apply the proper amount of water to diverse species of plants in a diverse suite of containers at different growth stages. Nursery

Tropical Nursery Concepts and Practices

1029

Table 4 Advantages and Disadvantages for four types of irrigation systems for container nurseries (Landis and Wilkinson 2014) Advantages Hand watering • Requires inexpensive equipment that is simple to install • Is flexible and can adjust for different species and container sizes • Irrigators have a daily connection to the crop and can scout out diseases or other potential problems • Allows water to be directed under plant foliage, reducing risk of diseases Microirrigation • Water is delivered directly to the root zone of plants (not to foliage, where it may cause disease) • Use of water is very efficient; less than 10 % of applied water is wasted • Delivery is uniform; an even amount of water is applied to each container • Infiltration rate is good (because of slow delivery) • The amount of leachate is low

Sprinkler irrigation • Relatively simple and inexpensive to design and install • A variety of nozzle patterns and application rates are available • Water distribution patterns can be measured with a cup test

Subirrigation • Although commercial products are available, subirrigation systems can be constructed from affordable, local materials • Foliage remains dry, reducing the risk of foliar diseases • Water use is efficient (up to 80 % less water use than overhead watering systems) • Application among plants is very uniform • Lower fertilizer rates are possible • Reduced leaching of mineral nutrients is possible • Drainage water can be captured for reuse or recycling • No splashing disrupts or displaces mulch, germinants, or medium • Provides the ability to irrigate different size containers and different age plants concurrently • Is efficient in terms of time and labor requirements following installation

Disadvantages • Is time consuming and labor intensive • Involves a daily responsibility including weekends and holidays • Requires skill, experience, and presence of mind to do properly • Presents a risk of washing out or compacting growing medium

• Designing the system and installing individual emitters for each plant is difficult and time consuming • It is not generally efficient to install for plants grown in containers smaller than 4 L in size • Each irrigation station must run a long time because of slow water delivery • Emitters can plug easily (water filtration and frequent irrigation system maintenance is required) • With drippers, it is difficult to verify water delivery visually; often, problems are not detected until it is too late • Foliar interception makes overhead watering ineffective for large-leaved crops • Irrigation water can be wasted because of inefficient circular patterns • An increased risk of foliar diseases is possible from excessive water on leaves • For overhead sprinklers, nozzle drip from residual water in lines can harm germinants and young plants • For basal sprinklers, irrigation lines must run along the floor, creating obstacles for workers and equipment • Overhead or hand watering may be required to ensure sufficient surface moisture until seeds germinate • No leaching occurs, so it cannot be used with poorquality water because salt buildup would occur • Less air pruning of roots occurs • Risk of spreading waterborne diseases is greater • High humidity within plant canopy is possible

1030

D.L. Haase et al.

managers need to ensure that irrigators have a conscientious attitude and are properly trained to work effectively with water application. For nurseries that grow plants in containers 4+ L in size, microirrigation can be an efficient watering method. Microirrigation usually involves polyethylene pipe fitted with microsprayers (sometimes called “spitters” or “spray stakes”), drippers inserted individually into each container, or smaller lateral tubing to reach all areas on the bench. This system makes very efficient use of water because it is applied directly to medium in each container. Many types of overhead irrigation systems exist, ranging from fixed sprinklers to moving boom systems. Fixed overhead sprinkler systems consist of a series of parallel irrigation lines, usually constructed of plastic polyvinyl chloride pipe, with sprinklers spaced at uniform intervals to form a regular grid pattern. The most expensive but efficient type of overhead sprinkler irrigation is the moveable boom, which applies water in a linear pattern. Moveable booms are generally considered too expensive for smaller nurseries but should be considered for large operations. For more information, see Landis et al. (1989a). A full discussion of types of irrigation designs and calculations is available in Stetson and Mecham (2011). Wide leaves combined with the close spacing of plants in a nursery create a canopy that intercepts most of the water applied through overhead irrigation systems, reducing water use efficiency and creating variable water distribution among plants. These problems can be precluded by subirrigation systems, which offer a promising alternative for tropical plant nurseries. Subirrigation has been successfully used to grow many native plants (Pinto et al. 2008; Dumroese et al. 2006; Davis et al. 2008). In subirrigation systems, the bottoms of containers are temporarily immersed in water on a periodic basis (for example, for a few minutes). The water is then drained, leaving the growing medium thoroughly wet while the leaves remain dry. The discarded water can be retained and reused; a worthy feature when the supply of good-quality irrigation water is restricted.

Plant Nutrition and Fertilization Plants require adequate quantities of mineral nutrients in the proper balance for basic physiological processes and to promote rapid growth and development. Young plants rapidly deplete mineral nutrients stored within their seeds, and cuttings have limited nutrient reserves. Therefore, nursery plants must rely on root uptake of nutrients from the growing medium. An important nutrition concept in plant nutrition is Liebig’s Law of the Minimum, which, when applied to plants, states that growth is limited by the mineral nutrient in shortest supply. Just as important as the absolute quantities of nutrients available to plants is the balance of nutrients. The proper nutrient balance is relatively consistent among plant species. Healthy plant tissue contains approximately 100 parts of nitrogen to 50 parts of phosphorus, to 15 parts of potassium, to 5 parts of magnesium, or to 5 parts of sulfur. On a practical basis, most nurseries use complete fertilizers that contain a balance of all mineral nutrients.

Tropical Nursery Concepts and Practices

1031

Sources of Mineral Nutrients Plants produced in tropical nurseries may acquire nutrients from several different sources, including the growing medium, irrigation water, beneficial microorganisms, and fertilizers. Many tropical nurseries use organic-based (soil-less) growing media that are essentially infertile, which enables nursery managers to apply the correct type of fertilizer, in the correct amount, and at the correct time. Native soils and composts contain higher nutrient concentrations than commercial growing media but rarely enough for the fast growth rate and nutrient balance desired in nurseries. To achieve the desired plant growth and health, fertilizers are the most common source of mineral nutrients. Many different types of fertilizers are used and vary according to their source materials, nutrient quantities, and mechanisms of nutrient delivery. Fertilizers can be broadly organized into two types: organic and synthetic (Landis and Dumroese 2011). Because of the variability involved, it is difficult to compare organic and synthetic fertilizers but some generalizations can be made (Table 5). Organic fertilizers can be defined as materials that are naturally occurring and have not been synthesized. Animal or plant wastes are what most people consider to be organic fertilizers and can be applied to crops directly or developed into a wide variety of other processed fertilizers. One of the attractions of these types of organic fertilizers is they are renewable and widely available. The second major category of organic fertilizers includes minerals and other materials that come directly from the earth. Minerals like sodium nitrate are commonly used in many blended organic fertilizers because they are soluble and have a high nutrient content. Like all types of mining, however, obtaining natural minerals is an extractive process and nonrenewable in the long term. Some native plant nurseries prefer organic fertilizers because they are less likely to burn crops, have lower risk of water pollution, and provide more hospitable conditions for beneficial microorganisms. The main drawbacks of organic fertilizers are that they are more

Table 5 Comparison of attributes of organic and synthetic fertilizers Factor Mineral nutrient analysis Range of mineral nutrients Nutrient release rate Compatibility with beneficial microorganisms Cost Handling Ecological sustainability Water pollution risk Other benefits

Organic Low All Slower Yes

Synthetic High One to many Faster At low levels

More Bulkier Yes Low Improves soil texture and encourages soil microbes

Less More concentrated No High Easier for research/control

1032

D.L. Haase et al.

Table 6 Comparison of advantages and disadvantages of two major types of synthetic fertilizers used in tropical plant nurseries Factor Nutrient release rate

Soluble fertilizer Very fast

Number of applications

Multiple – must be applied at regular intervals Good, but dependent on irrigation coverage Easy and quick

Uniformity of application Adjusting nutrient rates and ratios Nutrient uptake efficiency Leaching and pollution potential Potential for fertilizer burn (salt toxicity) Product cost Application costs

Controlled-release fertilizer Much slower – dependent on type and thickness of coating Usually once per season, but additional top-dressing is an option Can be variable if incorporated, resulting in uneven growth Difficult

Poorer

Better

Higher

Lower

Low if applied properly

Low, unless prills damaged during incorporation or in high temperatures Higher Lower

Lower Higher

expensive, and their lower nutrient content and solubility result in slower plant growth. Synthetic fertilizers are popular because they are relatively inexpensive, readily available, and have higher nutrient content compared with organic products. In populated tropical areas, synthetic fertilizers can be found at garden supply shops and through horticultural dealers, but inaccessibility and transport costs may be a limitation in remote areas. In the humid tropics, storage of synthetic fertilizers becomes a challenge because they readily absorb moisture from the air. Synthetic fertilizers can be divided into two classes: (1) soluble products that release nutrients quickly when dissolved in water and (2) slow-release or controlled-release fertilizers that release nutrients slowly over time. Both types have their advantages and disadvantages, which need to be considered before deciding upon a fertilization system (Table 6). Granular fertilizers that are used on lawns or in agriculture are not recommended for native plant nurseries.

Fertilizer Application Rates and Methods Fertilizer application rates depend on the growing environment and other factors such as container volume, type of growing media, growth stage (establishment, rapid-growth, hardening [described in the chapter, “▶ Planning and Managing a Tropical Nursery”]), and irrigation frequency (for example, Dumroese et al. 2011). Very small containers require lower rates applied frequently whereas larger containers can tolerate high application rates applied less frequently. In general, three types of fertilizers (liquid, controlled-release, and organic) are commonly used in nurseries.

Tropical Nursery Concepts and Practices

1033

Injecting liquid fertilizer solution into the irrigation system is a practice called “fertigation”. Fertilizer injectors range from simple, low-cost siphons for hand watering to sophisticated pumps for automated sprinklers. Because it can be designed to apply the proper mineral nutrients at the appropriate concentration for each growth stage, fertigation has several advantages compared with other types of fertilization. Remember that fertilizers are salts and that injecting liquid fertilizers adds to the base salinity level of the irrigation water; adding enough fertigation solution so that it barely drips from the bottoms of containers should avoid problems with salt accumulation in the medium, but if a salty crust appears at the bottom drainage holes, leaching the medium with regular irrigation water (“clearwater flush”) will reduce salt accumulation. Every fertilizer injector must be installed with a backflow preventer to eliminate the possibility that soluble fertilizer could be sucked back into the water line and contaminate drinking water. Controlled-release fertilizers (CRF) can be either topdressed (sprinkled onto the surface of the medium), if care is taken to ensure that each container or cell receives an equal number of prills or incorporated into the growing medium. If growers mix CRF into their growing medium, care must be taken to ensure uniform distribution and to prevent damaging the prill coating. If the coating is fractured, then the fertilizer releases immediately causing severe salt injury. Managers can begin by using the rate recommendations provided by manufacturers (low, medium, high) if they have an idea about their particular crop; regardless, rates should be evaluated for their effect on individual plant growth and performance. Organic fertilizers can be solid or liquid, commercially prepared, or natural. Composts could be incorporated into growing media but they must be fully mature to prevent fertilizer burn. One of the challenges of using liquid organic fertilizers is how to achieve the high soluble nitrogen levels necessary for rapid growth rates. Highquality nursery crops can be grown with organic fertilizers but, because their nutrient analysis is relatively low (Table 7), production schedules may have to be adjusted. Growers need to be aware of the different nutrient requirements during each growth phase (described in the chapter, “▶ Planning and Managing a Tropical Nursery”) and adjust fertilizer prescriptions accordingly, especially during hardening. These adjustments are particularly important for nitrogen (especially the ammonium form of nitrogen), which tends to be a primary driver of plant growth and development. Because artificial growing media, such as coir or pumice, are inherently infertile, fertilization should begin as soon as the seedlings or cuttings become established. Some commercial brands of growing media contain a starter dose of fertilizer that should be considered when determining fertilizer rates. Homemade soil mixes that have been amended with compost or other organic fertilizers may not need fertilization immediately, so observe plant growth and establish small trials to be certain. Some tropical plant species require very little fertilizer while others must be “pushed” with nitrogen to achieve desired growth rates and reach target specifications. Gaining experience and keeping good records about growing a particular species is the best way to develop species-specific fertilizer prescriptions. Managers should never wait for their crops to show deficiency symptoms before fertilizing

1034 Table 7 Percentages of nitrogen, phosphorus, and potassium supplied by a variety of organic materials (Diver et al. 2008)

D.L. Haase et al.

Organic fertilizers Bat guano (fresh) Bat guano (old) Blood meal Bone meal (steamed) Cottonseed meal Eggshells Fish emulsion Fish meal Greensand Hoof and horn meal Kelp meal Soybean meal Worm castings Manure: Cow Horse Pig Sheep Poultry

Nitrogen 10 2 10 1 6 1.2 4 5 0 12 1.5 7.0 0.5 2 1.7 2 4 4

Phosphorus 3 8 0 11 2 0.4 1 3 0 2 0.5 0.5 0.5 2.3 0.7 1.8 1.4 4

Potassium 1 0 0 0 1 0.1 1 3 7.0 0 2.5 2.3 0.3 2.4 1.8 1.8 3.5 2

because it can be difficult to alleviate the problem and produce the crop on schedule. Crop monitoring and testing, experience, and knowledge about the growth phases are the best guides for determining fertilizer timing and rates. Refer to Jacobs and Landis (2014) for further information on managing nutrition in tropical nurseries.

Monitoring Nutrition Practices Growers who fertigate need to periodically check the EC of the applied fertigation water and the growing medium solution. Measuring the fertigation water as it is applied to the crop can confirm that the fertilizer solution has been correctly calculated and that the injector is functioning properly. Simple handheld EC meters are fairly inexpensive. Normal readings in applied fertigation should range from 0.75 to 2.0 μS/cm. The typical range of acceptable EC values in the growing medium for most native plant species is 1.2–2.5 μS/cm. If the EC is more than 2.5, it is a good idea to leach out the salts with clean irrigation water. Testing plant foliage is the best way to monitor plant nutrition and responses to fertilization because it provides an exact measurement of nutrients that the plant has acquired (Landis et al. 2005). By examining tissue nutrient concentrations and simultaneously monitoring plant growth, it is possible to identify if and when specific nutrients are deficient or excessive. Foliar samples must be collected in a systematic manner and sent to a reputable laboratory for processing (see section “Testing Growing Media” discussed earlier). The analyzed nutrient concentration

Tropical Nursery Concepts and Practices Table 8 Estimated ranges of foliar nutrient levels for healthy tropical plants (based on data compiled by Drechsel and Zech (1991)) on field-grown, broadleaved, tropical tree species). Nutrient ranges can vary greatly among species, therefore, trials are recommended to determine the best ranges for specific species

Nutrient Macronutrients (%) Nitrogen Phosphorus Potassium Calcium Magnesium Sulfur Micronutrients (ppm) Iron Manganese Zinc Copper Boron Molybdenum

1035

Range of foliar levels in healthy plants 1.5–3.5 0.10–0.25 0.60–1.8 0.50–2.5 0.15–0.50 0.10–0.30 50–250 35–250 10–40 5–20 15–50 0.10–1.0

values can be compared with some known set of adequate nutrient values to determine which specific elements are deficient (Table 8). Small growth trials are another good way to monitor plant nutrition and fertilization needs. These trials are especially informative for tropical plant species because so little published information is available. Detailed documentation of growing conditions, fertilizer inputs, and resulting plant response can help formulate future fertilizer prescriptions for a specific species.

Reducing the Environmental Effects of Fertilization Regardless of the method of fertilizer application or the type of fertilizer used, runoff of excess fertilizers is a major environmental concern. Nutrients, notably nitrate and phosphate, leach easily from container nurseries and can pollute groundwater or adjacent streams. Managers should choose the types of fertilizers and schedule their applications to minimize potential pollution concerns. Because nitrate and phosphate are extremely soluble in water, growers should irrigate only when necessary and then apply only enough water so that only small amounts drain from the containers. This approach also makes sense from an economic standpoint, because the desire is to have most of the applied fertilizer taken up by crop plants rather than lost in runoff.

Beneficial Microorganisms In natural ecosystems, the root systems of most plants have microbial partnerships with mycorrhizal fungi and, if applicable, with nitrogen-fixing bacteria. These partnerships enable plants to survive and grow even in harsh conditions. Without microsymbiont partners, plants remain stunted and often die. In the nursery,

1036

D.L. Haase et al.

microsymbionts can be introduced by “inoculating” the root systems of plants with the appropriate beneficial microorganisms to form effective partnerships. Plants that have been inoculated in the nursery will be outplanted with microbial partnerships in place and often are better able to survive in the field.

Nitrogen-Fixing Bacteria Nitrogen (N), one of the most important nutrients for plant growth, is abundant in the Earth’s atmosphere as N2, but it must be converted to either nitrate (NO3) or ammonium (NH4+) before most plants can use it. In nature, N2-fixing bacteria convert (“fix”) N2 from the air into a form usable to plants. When the growing roots of a plant capable of forming a partnership with rhizobia come in contact with a compatible strain of N2-fixing bacteria in soil or growing media, the rhizobia bacteria will enter (“infect”) the roots. Nodules then form on the plant’s roots where the contact occurred. The bacteria live and multiply in the nodules on the host root system, providing N from the atmosphere to their plant host. Two types of N2-fixing bacteria form symbiotic partnerships with plants: rhizobia (consisting of several genera) and the genus Frankia. Inoculants are live N2-fixing bacteria cultures that are applied to seeds or young plants, imparting the beneficial bacteria to the plant’s root system. Inoculants for N2-fixing bacteria tend to be very specialized. Care must be taken to select appropriate and effective N2-fixing partners for specific plant species. Pure-culture inoculants of N2-fixing bacteria usually come in small packets of finely ground peat moss. Not all manufactured inoculants are selected and matched to native species, however, so be sure to check the source. Manufactured products usually come with application instructions; these directions need to be followed. Crude inoculant can be made from nodules collected from the roots of healthy, established host plants. For rhizobia, a brown, pink, or red color inside is usually a good indicator that the millions of bacteria in the nodule are actively fixing N2. For Frankia, desirable nodules will be white or yellow inside. Grey or green nodules should be avoided, because they likely are inactive. As soon as possible after collection, put the nodules in a blender with clean, chlorine-free water. About 50–100 nodules blended in about 1 L of water are sufficient to inoculate about 500 plants. Inoculant for N2-fixing bacteria is commonly applied when seedlings are emerging, usually within 2 weeks of sowing, or just after cuttings have formed roots. The inoculant is watered into the growing media or soil in which seedlings are growing. After 2–6 weeks, these four signs indicate that the plant has formed a symbiotic partnership with N2-fixing bacteria: (1) plants begin to grow well and are deep green despite the absence of added N2 fertilizer (Fig. 12a); (2) root systems give off a faint but distinctive ammonia-like scent; (3) nodules are visible on the root system; and (4) when a nodule is broken open, its inside is pink, red, or brown (for rhizobia) (Fig. 12b), or yellow or white (for Frankia).

Tropical Nursery Concepts and Practices

1037

Fig. 12 The 6-week-old native Acacia koa seedlings (right) were inoculated with rhizobia at 2 weeks of age; the seedlings on the left were not inoculated (a). Nodules from an Acacia koa seedling showing pink inside, signifying nitrogen is being fixed (b) (Photo (a) by Craig R. Elevitch and photo (b) by J.B. Friday)

Mycorrhizal Fungi “Myco” means “fungus” and “rhiza” means “root;” thus “mycorrhizae” means “fungus-roots.” Most of the world’s plants depend on their partnership with mycorrhizal fungi to thrive. The host plant’s roots provide a substrate for the fungi and supply food in the form of simple carbohydrates. In exchange, the mycorrhizal fungi offer increased water and nutrient update, stress and disease protection, and increased vigor and growth. Mycorrhizal fungi are not “one size fits all,” but they often are “one size fits many.” Also, one plant can partner simultaneously with several species of mycorrhizal fungi, and a plant may change partners over time as it grows and adapts to its environment (Amaranthus 2010). Three types of mycorrhizae are important to tropical native plant nurseries.

1038

D.L. Haase et al.

• Arbuscular Mycorrhizal (AM) Fungi – AM fungi are essential for most tropical trees and other plants and for many annual crops and grasses. AM fungi are not visible on plant roots to the unaided eye and must be observed under a microscope. Inoculant for AM fungi is sometimes collected from root systems of AM host plants or soil underneath them and incorporated into growing media. Another method is pot culture inoculant, in which a specific AM fungus species is acquired either commercially or from a field site as a starter culture and then incorporated into a sterile growing medium. A host plant, such as corn, sorghum, clover, or an herbaceous native plant, is grown in this substrate. After the host plant roots have spread throughout the medium, their shoots are removed and the substrate, now rich in roots, spores, and mycelium, is chopped up and incorporated into fresh growing medium (Habte and Osorio 2001, Miyasaka et al. 2003). Commercial sources of AM fungi inoculant are also available, usually containing several species or strains. • Ectomycorrhizal (ECM) Fungi – ECM fungi only affect a small percentage of tropical species, including pines, eucalypts, poplars, oaks, dipterocarps, and some legumes. Nurse plants and soil spores have been used historically, while spores collected from the fruiting bodies of mushrooms and pulverized in a blender or pure culture inoculant in a peat-based carrier are usually recommended for nurseries. • Ericoid Mycorrhizal (ERM) Fungi – Plants that form partnerships with ericoid mycorrhizal fungi are able to grow in exceptionally nitrogen-poor soils and harsh conditions, including bogs, alpine meadows, tundra, and even in soils with high concentrations of certain toxic metals. Similar to ECM fungi and AM fungi, ericoid mycorrhizal fungi must come in contact with the host plants roots to form partnerships. Ericoid mycorrhizal inoculant is available as commercial cultures or from soil near healthy host plants. The product or soil is mixed into nursery growing medium.

Problem Prevention and Holistic Pest Management Holism is the theory that systems are not a group of isolated parts, but rather should be viewed as a whole. Holistic pest management is an integrated and preventative approach that considers the overall health of the plant and the nursery environment to prevent problems and to manage them wisely if they arise. Holistic pest management includes problem prevention through cultural mechanisms, early detection and evaluation, and management measures as needed to suppress pests and ecologically balance their populations. Holistic pest management can reduce reliance on pesticides (Dumroese et al. 1990). For holistic pest management, it is important to remember the “disease triangle” concept that illustrates the interrelationships among the pest, host, and environment. All three factors are necessary to cause biotic disease.

Tropical Nursery Concepts and Practices

1039

The holistic approach to nursery pest management involves a series of four interrelated practices, which ideally function together (modified from Wescom 1999): • Problem Prevention through Cultural Measures – includes good sanitation, proper scheduling, management of the nursery environment, and promotion of plant health through proper irrigation and fertilization. • Problem Detection and Diagnosis – is accomplished through regular monitoring, recordkeeping, and accurate problem identification. • Problem Management – includes, if necessary, timely and appropriate pest suppression measures and balancing pest populations with beneficial organisms and pest predators. • Ongoing Process Evaluation – is to learn from experience by assessment and improved effectiveness of pest management approaches. A complete description of holistic pest management is beyond the scope of this chapter. For a good overview of this concept, consult Landis et al. (1989b, 2008) and Dumroese (2012), and for specific focus on tropical nurseries, refer to Landis et al. (Landis et al. 2014c). Acknowledgements This chapter draws heavily on Wilkinson et al. (2014) and we thank Brian F. Daley, Douglass F. Jacobs, David P. Janos, and Tara Luna for their contributions.

References Amaranthus M (2010) Personal communication. Mycorrhizal Applications, Grants Pass, Oregon, USA Buamscha G, Altland J (2005) Pumice and the Oregon nursery industry. Digger 49(6):18–27 Bunt AC (1988) Media and mixes for container grown plants. Unwin Hyman, London, p 309 Carrion C, Abad M, Fornes F, Noguera V, Maquieira A, Puchades R (2006) Leaching of composts from agricultural wastes to prepare nursery potting media. Acta Horticult 697:117–124 Cayanan DF, Zheng Y, Zhang P, Graham T (2008) Sensitivity of five container-grown nursery species to chlorine in overhead irrigation water. HortScience 43:1882–1887 Davis AS, Jacobs DF, Overton RP, Dumroese RK (2008) Influence of irrigation method and container type on growth of Quercus rubra seedlings and media electrical conductivity. Native Plants Journal 9:4–12 Diver S, Greer L, Adam KL (2008) Sustainable small-scale nursery production. National Center for Appropriate Technology (NCAT) Sustainable Agriculture Project, Butte. https://attra.ncat. org/attra-pub/summaries/summary.php?pub=60. Accessed 11 Nov 2011 Drechsel P, Zech W (1991) Foliar nutrient levels of broad-leaved tropical trees: a tabular review. Plant Soil 131:29–46 Dumroese RK (2012) Integrated nursery pest management. In: Cram MM, Frank MS, Mallams KM (eds) Forest nursery pests (tech cords). US Department of Agriculture, Forest Service, Agriculture Handbook 680, Washington, DC, pp 5–12 Dumroese RK, Wenny DL, Quick KE (1990) Reducing pesticide use without reducing yield. Tree Planters’ Notes 41(4):28–32

1040

D.L. Haase et al.

Dumroese RK, James RL, Wenny DL (2002) Hot water and copper coatings in reused containers decrease inoculum of Fusarium and Cylindrocarpon and increase Douglas-fir seedling growth. HortScience 37:943–947 Dumroese RK, Pinto JR, Jacobs DF, Davis AS, Horiuchi B (2006) Subirrigation reduces water use, nitrogen loss, and moss growth in a container nursery. Native Plants Journal 7:253–261 Dumroese RK, Luna T, Landis TD (eds) (2008) Nursery manual for native plants: volume 1, a guide for tribal nurseries. US Department of Agriculture, Forest Service, Agricutlure Handbook 730, Washington, DC, p 302 Dumroese RK, Davis AS, Jacobs DF (2011) Nursery response of Acacia koa seedlings to container size, irrigation method, and fertilization rate. J Plant Nutr 34:877–887 Dumroese RK, Landis TD, Luna T (2012) Growing native plants in nurseries: basic concepts. US Department of Agriculture, Forest Service, Rocky Mountain Research Station, General Technical Report RMRS-GTR-274, Fort Collins, p 84 Evans J (1996) Plantation forestry in the tropics. Clarendon, Oxford Evans MR, Konduru S, Stamps RH (1996) Source variation in physical and chemical properties of coconut coir dust. HortScience 31:965–967 Grubinger V (2007) Potting mixes for organic growers. University of Vermont Extension, Brattleboro. http://www.uvm.edu/vtvegandberry/factsheets/pottingmix.html. Accessed 21 Aug 2011 Habte M, Osorio NW (2001) Arbuscular mycorrhizas: producing and applying arubscular mycorrhizal inoculum. Honolulu, HI: University of Hawai‘i, College of Tropical Agriculture and Human, p 47 Handreck KA, Black ND (1984) Growing media for ornamental plants and turf. Kensington, Australia: New South Wales University Press, p 401 Jacobs DF, Landis TD (2014) Plant nutrition and fertilization. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 233–250 Jacobs DF, Landis TD, Luna T, Haase DL (2014) Propagation environments. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 89–99 Jaenicke H (1999) Good tree nursery practices: practical guidelines for research nurseries. International centre for research in agroforestry. Majestic Printing Works, Nairobi, p 93 Johnson P (1968) Horticultural and agricultural uses of sawdust and soil amendments. Paul Johnson, National City, p 46 Landis TD (1995) Improving polybag culture for sustainable nurseries. Forest Nursery Notes (July 1995) pp 6–7 Landis TD, Dumroese RK (2006) Monitoring electrical conductivity in soils and growing media. Forest Nursery Notes (Summer 2006), pp 6–10 Landis TD, Dumroese RK (2011) Using organic fertilizers in forest and native plant nurseries. Forest Nursery Notes 31(2):9–18 Landis TD, Morgan N (2009) Growing media alternatives for forest and native plant nurseries. In: Dumroese RK, Riley LE (eds) National proceedings: forest and conservation nursery associations – 2008 (tech coords). US Department of Agriculture, Forest Service, Rocky Mountain Research Station, Proceedings RMRS-P-58, Fort Collins, pp 26–31 Landis TD, Wilkinson KM (2014) Water quality and irrigation. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 207–230 Landis TD, Tinus RW, McDonald SE, Barnett JP (1989a) The container tree nursery manual. Volume 4, seedling nutrition and irrigation. US Department of Agriculture, Forest Service, Agriculture Handbook 674, Washington, DC, p 119

Tropical Nursery Concepts and Practices

1041

Landis TD, Tinus RW, McDonald SE, Barnett JP (1989b) The container tree nursery manual. Volume 5, the biological component: nursery pests and mycorrhizae. US Department of Agriculture, Forest Service, Agriculture Handbook 674, Washington, DC, p 171 Landis TD, Haase DL, Dumroese RK (2005) Plant nutrient testing and analysis in forest and conservation nurseries. In: Dumroese RK, Riley LE, Landis TD (tech coords) National proceedings, forest and conservation nursery associations – 2004. US Department of Agriculture, Forest Service, Rocky Mountain Research Station. Proceedings RMRS-P-35, Fort Collins, pp 76–83 Landis TD, Luna T, Dumroese RK (2008) Holistic pest management. In: Dumroese RK, Luna T, Landis TD (eds) Nursery manual for native plants: a guide for tribal nurseries. Volume 1: nursery management. USDA Forest Service. Agriculture Handbook 730, Washington, DC, pp 262–275 Landis TD, Jacobs DF, Wilkinson KM, Luna T (2014a) Growing media. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 101–121 Landis TD, Jacobs DF, Wilkinson KM, Luna T (2014b) Containers. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 123–139 Landis TD, Luna T, Wilkinson KM, Dumroese RK (2014c) Problem prevention and holistic pest management. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 273–291 Lovelace W, Kuczmarski D (1994) The use of composted rice hulls in rooting and potting media. Int Plant Propag Soc Comb Proc (1992) 42:449–450 Miller JH, Jones N (1995) Organic and compost-based growing media for tree seedling nurseries. World Bank Tech. Paper No 264, Forestry Series. The World Bank, Washington, DC, 75 p Miyasaka SC, Habte M, Friday JB, Johnson EV (2003) Manual on arbuscular mycorrhizal fungus production and inoculation techniques. University of Hawai’i at Mānoa, College of Tropical Agriculture and Human Resources, Honolulu, p 4 Newman J (2007) Core facts about coir. Greenhouse Manag Proc 27(2):57 Noguera P, Abad M, Noguers V, Puchades R, Maquieira A (2000) Coconut coir waste: a new and environmentally friendly peat substitute. Acta Hort 517:279–286 Pinto JR, Chandler R, Dumroese RK (2008) Growth, nitrogen use efficiency, and leachate comparison of subirrigated and overhead irrigated pale purple coneflower seedlings. HortScience 43(3):897–901 Robbins J (2011) Irrigation water for greenhouses and nurseries. Publication FSA6061. University of Arkansas, Horticulture Department, Little Rock, p 6 Stetson LE, Mecham BQ (2011) Irrigation, 6th edn. Irrigation Association, Falls Church, p 1089 Timmer VR, Parton WJ (1982) Monitoring nutrient status of containerized seedlings. In: Proceedings, Ontario ministry of natural resources nurseryman’s meeting, 1982 June, Thunder Bay/Toronto, Ontario Ministry of Natural Resources, pp 48–58 Wescom RW (1999) Nursery manual for atoll environments. SPC/UNDP/AusAID/FAO Pacific Islands forests and trees support programme, RAS/97/330. Working Paper 9. p 53 Whitcomb CE (2003) Plant production in containers II. Lacebark Publications, Stillwater, p 129, 1 Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) (2014) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service. Agriculture Handbook 732, Washington, DC, p 377

Planning and Managing a Tropical Nursery Kim M. Wilkinson, Thomas D. Landis, Diane L. Haase, and R. Kasten Dumroese

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Planning a Tropical Nursery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Determine Community Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Define Target Plants for the Nursery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Start with a Small Pilot Nursery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Select a Nursery Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Defining the Target Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Case Study: Defining the Target Seedling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Field Testing the Target Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nursery Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Propagation Protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Propagule Collection, Processing, and Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Crop Growth Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monitoring Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems with Holdover Stock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Facilities Schedules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scheduling Multiple Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Harvesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Packing and Shipping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1044 1045 1046 1046 1048 1049 1050 1058 1059 1060 1062 1063 1064 1067 1067 1069 1069 1069 1071

K.M. Wilkinson (*) Gibsons, BC, Canada e-mail: [email protected] T.D. Landis Native Plant Nursery Consulting, Medford, OR, USA e-mail: [email protected] D.L. Haase State and Private Forestry, USDA Forest Service, Portland, OR, USA e-mail: [email protected] R.K. Dumroese USDA Forest Service, Rocky Mountain Research Station, Moscow, ID, USA e-mail: [email protected] # Springer-Verlag Berlin Heidelberg (outside the USA) 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_94

1043

1044

K.M. Wilkinson et al.

Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Record Keeping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problem Solving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1072 1073 1074 1076

Abstract

Every tropical nursery will have a unique design based on local needs and resources. Assessing the needs and desires of the community and potential clients is essential to define nursery objectives and determine what plants and services the nursery will provide. Starting with a small pilot nursery is a valuable step to understand how to propagate and manage plant production. A pilot phase also helps to assess costs, resources, and feasibility of operating a nursery. Nursery site selection based on environmental and socioeconomic considerations is critical. Once a site is chosen, planning the layout and flow of work supports nursery efficiency and efficacy. The Target Plant Concept is a feedback system between nursery managers and their clients in which information from the project site is used to define and refine the best type of plant material to grow. The client and nursery manager define project objectives, site conditions, limiting factors, mitigating measures, species and genetics, stocktypes, outplanting windows, and outplanting techniques to determine the ideal plant for that particular project. The nursery then produces these target plants that are evaluated after outplanting to refine target specifications for future crops. Nursery crop planning ensures that quality plants are ready when conditions at the outplanting site are optimal; in other words, ready when the clients need them. Crop planning involves visualizing, planning, and scheduling all the growth phases of each crop, from collection of propagules through to the delivery of field-ready plants. Propagation protocols for each species define requirements during all phases of crop growth. Record keeping helps track developments, minimize problems, and expand successes in future seasons. During harvesting, packing, and shipping, it is critical to handle plants with utmost care to maintain a high level of quality. Moisture, temperature, and physical stresses are damaging and cumulative and should be avoided. Even with the best nursery management, unforeseen problems will arise. Knowing how to approach problems systematically will ensure the nursery can overcome challenges. Keywords

Target Plant Concept • Nursery Site Selection • Recordkeeping

Introduction Tropical forests represent almost half of the world’s forested area and much of the world’s biodiversity, but these forests are under severe and unrelenting pressure for conversion to other uses. Converted to agriculture and pasture, these formerly

Planning and Managing a Tropical Nursery

1045

forested areas are often abandoned; the result is a loss of productivity, carbon sequestration, water quality, biodiversity, and other important ecosystem services (Davis et al. 2012). Tropical nurseries, by providing plants, can be a key link in restoring tropical ecosystems, often integrating goals of economic resiliency, cultural renewal, and other community needs with ecosystem protection and restoration (Wilkinson and Daley 2014). Tropical nurseries effectively and affordably increase the number of plants available to foresters, farmers, gardeners, restoration ecologists, community groups, and others. These nurseries can be small, operated by a single person growing a few native and culturally important plants for neighbors. They can also be managed by larger forestry corporations or programs producing genetically improved, nonnative species to be intensively managed in plantations (e.g., Stape et al. 2001). Forest plantations reduce dependence on harvesting of native forests and can be managed to improve biodiversity by reexamining monetary and ecological values and the benefits of diversification of species, particularly native species (Davis et al. 2012). Hopefully, a unifying thread for tropical nurseries, regardless of size, is a focus on efforts that will be of the most benefit to the land and the local people. Good nursery management supports genetic and sexual diversity to ensure conservation of plants and the species that depend on them (Landis et al. 1993). Genetic diversity is essential for future adaptability and resilience in the face of climate change (Williams and Dumroese 2013). Good nursery work helps to protect local biodiversity and ecosystem resilience and contributes to species conservation, cultural diversity, economic resilience, enhanced human livelihoods, and greater health and productivity of the land. And, proper nursery management has a pronounced effect on survival and growth of tropical species (e.g., Mexal et al. 2002). Therefore, in this chapter, we present the basic tenets of planning and managing a tropical forest nursery. Although our focus is less on the forest products industry and more toward restoration of tropical forests, the concepts are universally applicable. In the first section, “Planning a Tropical Nursery,” we present important steps in planning (or refining) a tropical nursery. One of these steps, describing the plants to be grown, is discussed in much greater detail in the second section, “Defining the Target Plant,” where we present the Target Plant Concept as a holistic approach to identifying the appropriate plant materials to be used to meet client objectives. In the third section, “Nursery Management,” we highlight critical steps in plant production and their required planning and management and stress the importance of record keeping toward improving the efficiency and economics of nursery operations.

Planning a Tropical Nursery Every tropical nursery will have a unique design based on distinct and local needs, goals, and resources. The initial planning phase is an opportunity to step back and thoughtfully clarify the vision and goals of the nursery, research how those visions

1046

K.M. Wilkinson et al.

and goals can be achieved, and begin organizing and coordinating the factors needed to reach these goals. For example, the initial thought may be to build a large structure, such as a greenhouse, to grow a particular suite of plants. Deeper exploration of the nursery’s goals along with some research on how the species are best grown, however, may reveal how to create a site-appropriate design of several different, smaller-scale environments that are ultimately more economical, efficient, and effective for producing plants (Wilkinson and Landis 2014). Most nurseries are founded on the vision of a person, or group of people in a community, of how that nursery’s efforts today will affect the local landscape and community in 10, 50, or 100 or more years. This vision guides all work and must include an understanding of the nursery’s vital role in protecting ecosystem health, species diversity, and genetic diversity, all of which will be of enormous benefit to the environment, the community, and future generations of plants, animals, and people.

Determine Community Needs Visions are translated into practical objectives through interactions with the local community and the local environment. Some questions to explore, as suggested by Wilkinson and Landis (2014), are: • • • •

What is needed and wanted in our community? Who are the potential clients of our plant materials now? Who might be potential clients in the future? What are the needs and priorities of the potential clients?

Formal and informal discussions with members of trade groups and guilds, elders with traditional knowledge, and instructors that work with plant products are often tremendous sources of information to answer those questions, as well as define the challenges people in the community face when outplanting native species on diverse sites (Fig. 1). This assessment is key to helping the nursery determine not only what species to grow but also the nursery-dependent characteristics of the plant (stocktype) clients might prefer. Although the nursery must listen to the needs and wants of its community, it can also share its visions and goals and increase awareness of native or traditional species that can be planted for specific purposes and benefits. A good understanding of local ecology, environmental issues, history, soil types, people’s needs, and site characteristics is essential.

Define Target Plants for the Nursery Nursery planning is critical for matching the plant materials produced with the conditions of the client’s outplanting sites to ensure successful establishment.

Planning and Managing a Tropical Nursery

1047

Fig. 1 Nursery managers must understand the challenges that plants will face on their client’s outplanting sites in order to grow the best plants for project needs. Here, the client, a farmer, needs seedlings that will produce forage; the plants must be capable of establishing on a dry site in East Timor (Photo by J.B. Friday)

This ensures that clients get what they expect: vigorous plants providing medicine, wood, food, habitat, and so on. Site conditions dictate the plant’s target size, genetic source, stocktype, and management in the nursery. Some topics to consider when deciding what to grow include: • Species. What is the nursery capable of growing? • Client needs. What does the client want from the plants? Medicine, food, timber, habitat, soil stabilization, windbreak, commercial products? • Environments. What types of sites need plants? • Stocktype. What size and age is preferred by clients? • Planting season. When is the best time to outplant? • Quantities. How many of each species will need to be grown? • Distance. How far are people willing to travel to obtain plants? Another consideration is what species are most needed from an ecological standpoint. For example, in some tropical forests, rarer or later successional species needed to restore diversity may not return without human help, and these species may be a high priority for nursery work (Holl and Aide 2011). The target plant requirements will differ among species and outplanting sites and will influence all aspects of nursery design: location, structures, container types, scheduling, management practices, propagule collection, and so forth. Further details for defining target plants are provided in the next section of this chapter.

1048

K.M. Wilkinson et al.

Start with a Small Pilot Nursery It is prudent to start with a small pilot nursery (Fig. 2). Such a nursery offers a way to better understand how to grow native plants with much less financial risk while simultaneously evaluating whether starting a nursery is a realistic and achievable undertaking. During the pilot phase, viable propagation strategies can be developed and the feasibility of growing particular plants assessed, particularly in terms of client needs and production costs. Production costs should include all phases of production, from collecting seeds to mixing growing media to delivering the plants and the costs associated with each (labor, materials, shipping, advertising, and so on). This information can be scaled up through different planning scenarios to determine potential market price and the scope and location of the necessary nursery facility. It can also be invaluable for making a more accurate feasibility assessment when the time comes to expand. The pilot nursery phase is also the time to consider financing. Based on small-scale production costs and potential market prices, it becomes clearer what finances will be required to start and operate the full-scale nursery. Some nurseries may be funded through grants or government programs while others are aligned with a particular project or organization. Private, for-profit nurseries must earn enough income from the sale of plants to at least pay for development, infrastructure, production costs, and staff time. Whatever the circumstances, finance is a key part of nursery planning. It determines: • • • • •

How much money can be invested in the nursery at the outset If staff can be hired In what time frame the nursery can start to produce plants for sale How many plants can be produced What price can be charged for the plant materials

Even if plants are to be distributed freely, it is still essential for planning and ensuring financial viability to know the production cost of each kind of plant. Fig. 2 Often a key part of developing a successful nursery is to begin with a small pilot nursery that can identify factors and reduce risks before embarking on a larger effort (Photo by Thomas D. Landis)

Planning and Managing a Tropical Nursery

1049

Nurseries based on a pilot phase should have a good grasp of the production costs, and these can be scaled up for full-scale nursery. After production in the full-scale nursery is underway and actual costs are known, it is imperative to revisit the price structure of the plants to ensure that they are in line with actual costs.

Select a Nursery Site Nursery site selection involves working with, rather than against, nature for the most effective, efficient, and economical design. The less the natural environment has to be modified to produce high-quality plants, the less expensive it will be to create optimal crop conditions. Careful observation of site conditions and an assessment of past and present climatic records are important. Even in cases where the site is already chosen, the process of site observation and inventory is helpful to understand the strengths and challenges of that site. Critical nursery site selection factors include the following: • Access to good quality, affordable, and abundant water • Unobstructed solar access • An easily accessible flat area for the delivery or processing of bulky, heavy material such as soil, sand, mulch, and fertilizer • Easy access and close proximity to staff • Adequate land area • Reliable energy supply (if water pumps, fans, or lights are used) Water quality always needs to be tested when a nursery site is being evaluated; sites with poor water quality should be avoided. For irrigation purposes, water quality is determined by the quantities of salts and pests in it. Details on testing and assessing water quality are provided in chapter “▶ Tropical Nursery Concepts and Practices” and in Landis and Wilkinson (2014b). Ecopolitical site-selection factors, notably land-use zoning, and concerns about neighboring land uses that may involve herbicide, pesticide, or potential groundwater contamination are also important factors for determining suitable sites. The amount of land must be large enough for the production areas and any support buildings, allow for the efficient movement of equipment and materials, and be evaluated on the basis of available space for possible expansion. An ideal site is sheltered from extremes such as high winds, storms, and severe temperature fluctuations. Gentle topography favors nursery construction and management much more than hilly or steep terrain. Access to the nursery by staff and clients is useful for economical nursery production. A backup water supply ensures crop survival through periods of drought or uncertainty. Firebreaks or a site selected to minimize fire risks can preclude disaster. Each potential nursery site will have good and bad attributes. A decision matrix can be constructed by listing the potential nursery sites across the top and the significant site selection criteria down the side (Table 1). The next step is to assign

1050

K.M. Wilkinson et al.

Table 1 Potential nursery sites can be evaluated and compared using a decision matrix. In this example, Site A received the highest score and is therefore considered the best choice for a nursery site (Landis et al. 1994) Site A Site selection criteria Critical factors Good solar access Water quality Water quantity Available energy Adequate land area Zoning restrictions Pollution concerns Secondary factors Microclimate Topography Labor supply Accessibility Shipping distances Total Site suitability

Site B Weighted score

Site C

Weight value

Rating

10 9 8 8 7 7 6

9 9 10 9 8 10 9

90 81 80 72 56 70 54

7 7 8 9 8 6 7

70 63 64 72 56 42 42

9 4 9 10 10 8 9

90 36 72 80 70 56 54

6 5 4 4 3

9 10 9 8 9

54 50 36 32 27 702 #1

8 9 8 6 7

48 45 32 24 21 579 #3

9 10 10 8 10

54 50 40 32 30 664 #2

Rating

Weighted score

Rating

Weighted score

each site selection criterion an importance value on a scale from 1 to 10, with the most critical factors receiving the highest values and the less important ones receiving lower values. Next, the suitability of each potential nursery location is evaluated and rated, again on a scale of 1–10, based on the information that has been gathered. The score for each cell in the matrix is then calculated by multiplying the weights for each site selection factor by the rating for each site. Finally, the weighted scores are totaled for each site, and, if the weights and rankings have been objectively assigned, then the potential nursery site with the highest total ranking should be the best choice. Consider the nursery’s local effect, too. For example, consider not only the source of good quality irrigation water but also where wastewater from the nursery will go. It may contain fertilizers and be a potential source of pollution, possibly creating problems and legal issues for the nursery. With proper planning, that wastewater may be used as a resource to produce other crops or maintain a landscape.

Defining the Target Plant People have diverse goals and face a multitude of challenges on their outplanting sites. The Target Plant Concept is used to define what plant materials to grow in the nursery to meet the needs on the outplanting sites and achieve project objectives (Landis and Wilkinson 2014a). The definition of the target plant depends on how it

Planning and Managing a Tropical Nursery

1051

Fig. 3 The Target Plant Concept starts with a novel approach to plant propagation: the client and nursery manager forge a partnership that focuses on putting the best plant materials on specific project sites. The manager and client then answer eight questions about the project; their answers define the target plant material necessary to meet project objectives. The nursery produces the plants. The client and manager subsequently reassess successes and failures and use that information to improve the next crop (Illustration by Jim Marin)

will be used – its “fitness for purpose” (Sutton 1980). The Target Plant Concept has two equally important components (Fig. 3). The first component incorporates three simple approaches that provide the broad, fundamental basis necessary to successfully complete outplanting projects. These three approaches, when considered together, guide the broad approach for defining and selecting the target plant materials for a specific site. These are: • Start at the outplanting site – Clients specify the type of plant material best suited for their site conditions. Understanding these needs, the nursery grows plant materials that are locally adapted, genetically appropriate, and the optimal size, age, and so on to survive and thrive on the outplanting site. • Forge a nursery-client partnership – The client and nursery manager work together to define the ideal type of plant for the project, the nursery grows the plants, and they are outplanted. Based on performance after outplanting, the client and nursery manager work together to revise target plant characteristics as needed to improve survival and growth of future crops. Good communication

1052

K.M. Wilkinson et al.

between clients and nursery managers builds partnerships and ensures the best possible plants for the project. • Emphasize seedling quality, not appearance – Plant quality is not determined by how good a plant looks but by its outplanting performance. A beautiful crop of plants in the nursery may perform miserably if the plants are inappropriate for conditions on the outplanting site. Using the Target Plant Concept, plants are grown in the nursery based on their fitness to thrive on the outplanting site and their ability to fulfill the project objectives. The second component of the Target Plant Concept is the process of defining the target plant materials. The nursery manager and client use the characteristics of the outplanting site to systematically answer eight interrelated questions (Fig. 3) to ultimately define the target plant material. Each question is described in the following sections. 1. What Are the Outplanting Objectives? Native and traditional plants are grown for a variety of project objectives, such as reforestation; enriching sustainable agriculture with windbreaks or erosion control using native species; providing shade in pastures; producing timber or wood; rekindling traditional agroforestry practices; ensuring local supplies of cultural or medicinal plants; restoring degraded land; controlling invasive species; protecting and producing traditional foods; educating young people; and developing small businesses. These objectives influence the characteristics of the target plant. Project objectives can sometimes include finding reference sites – natural or recovered areas that serve as models for desirable recovery of native plant communities. The comparison of the soils, climate, vegetation, and other characteristics of reference sites to the project site provides guidance about what species may be established and is essential for setting attainable, site-appropriate goals for restoration (Steinfeld et al. 2007). Often several reference sites of different ages and recovery stages are used for one project. Reference sites also show succession, which is how plant communities may change over time. Some projects formally translate their objectives into measurable goals within specified time frames. For example, a reforestation project might set a goal of 800 living trees per hectare 2 years after outplanting. A native plant project may have a goal of 75 % vegetative ground cover of which 90 % will be composed of perennial native species 1 year after outplanting. Production of food, craft materials, or timber might be stated in yields. Habitat restoration will define desired fauna moving into the planting area. 2. What Are the Conditions of the Outplanting Site? A good site evaluation, at a minimum, includes mapping the basic land features of the site and basic information about soils, vegetation, climate, and site history. Topographic maps, aerial photos, soil surveys, and special maps can all provide valuable information about waterways and water features, soils, slopes, aspects,

Planning and Managing a Tropical Nursery

1053

Table 2 Comparison of different plant establishment methods (Landis et al. 1993)

Characteristics Efficient use of seeds and cuttings Cost of establishment Ability to establish difficult species Option of using specific genotypes Precise scheduling of plant establishment Control of stand composition and density Matching stocktypes to site conditions Depletion of adjacent plant stands

On-site methods Transplanting Outplanting wildlings nonrooted cuttings N/A No

Direct seeding No

Outplanting nursery stock Yes

High Yes

Moderate No

Low No

Moderate Yes

No

No

Yes

Yes

Yes

Yes

No

Yes

Yes

Yes

No

Yes

No

No

No

Yes

Yes

Yes

No

No

vegetation, and human-made features including access roads and structures. A soil test can provide information on pH and nutrients. These, combined with the client’s personal observations, can identify challenges to outplanting success, such as issues with drainage, compaction, erosion, prevailing winds, intense solar radiation, incorrect pH, depth to water table, and so on. Climate information can be obtained from national or regional weather services. At a minimum, clients need to understand the average rainfall and temperatures on their site and the annual fluctuations and patterns of these conditions. Trends of when the rainy season normally starts and ends and the intensity of rain events are important. Microclimates should be considered, too. Two sites situated only a few kilometers apart can experience dramatically different rainfall, temperatures, and wind effects, particularly prevailing winds that are anticipated, consistent, and bend or “flag” vegetation in one direction (Mollison and Slay 1991). Water and the way it moves through a site are important factors for microclimate. Salt spray, fog, and mist can be observed. Water features such as ponds, lakes, springs, and streams also create microclimates that affect vegetation. 3. What Factors on the Project Site Could Limit Success? The information gathered answering question 2 is used to identify the environmental factors that are most limiting to plant survival and growth on the site and to specify which plant species and stocktype would be most appropriate (Table 2). Water is often the most limiting factor on sites with a pronounced dry season or low annual rainfall. Where populations of grazing and browsing animals are high, animals may be the most limiting factor. High winds, salt spray, weeds, wildfire,

1054

K.M. Wilkinson et al.

Fig. 4 Overcoming limiting factors to restore disturbed sites involves a combination of nursery and client efforts. On a site with tall grass, the client would take steps to control the grass competition, and the nursery would grow larger seedlings able to overtop the grass quickly (Photo by Douglass F. Jacobs)

soil fertility, seasonal flooding, insect pests, and land use challenges, such as people driving over restoration areas, are further examples of limiting factors. Awareness of native plant needs and ecology (based on reference sites) is important to determine if the factors are truly limiting based on project objectives. For example, a site may not receive much rainfall, but if the goal is dryland ecosystem restoration and the dryland native species are adapted to that amount of rainfall, these plants will be able to thrive on the site. 4. How Will Limiting Factors Be Mitigated? Mitigating measures are the steps the client will take to overcome any limiting factors on the site. Some mitigating measures can be defined by the client and included in their target plant requirements, while others must be completed by the client on their site before outplanting. Nursery efforts to produce target plants are often key to overcoming limiting factors on the site. For example, an absence of beneficial microorganisms on the site can be mitigated by inoculating plants with beneficial microorganisms in the nursery. Sometimes mitigating measures will be a combination of nursery efforts to produce target plant materials and client efforts on the site (Fig. 4). For sites with browsing or grazing animals, species selection for less palatable species may be important, but the client must exclude these animals from newly planted areas or at least provide shelter for the plants as they establish. Limiting factors that cannot be mitigated through reasonable target plants and land manager efforts in the field require the client to revise project objectives to be appropriate and achievable on their project site. 5. What Species and Genetic Sources Will Meet Project Objectives? Clients decide what species they will plant based on project objectives, reference sites, project site conditions, and limiting factors as described previously. The

Planning and Managing a Tropical Nursery

1055

reference sites provide a natural model for determining appropriate species and plant community composition (what percent of which species, spaced how far apart) for the recovering site. In some cases, mitigation of limiting factors alone is sufficient to allow for natural regeneration of certain native and desirable plant species. In these cases, the nursery work may focus on those species that do not regenerate naturally. On many disturbed sites, however, no amount of mitigation will lead to adequate natural regeneration of native or desired species, and in these cases, the nursery will be involved in propagating these species for reintroduction by planting. In addition to proper species selection, genetic factors must be considered when collecting plant materials. Plants are genetically adapted to local environmental conditions, and for that reason, plant materials (seed, cuttings, and so on) should always be collected from the same geographic area, environmental conditions, and elevation in which the nursery stock is to be outplanted. Local adaptation can affect outplanting survival and growth and may also be essential for long-term viability and habitat value of restoration plantings. For example, local pollinators are often adapted to the flower sizes and shapes of their locally adapted food plants (Bawa 1990). To preserve biodiversity, target plant materials should attempt to represent all the genetic and sexual diversity present on the reference sites, but what that means in terms of the number of donor plants remains elusive (Mijnsbrugge et al. 2010). Guinon (1993) suggests collecting from at least 50 to 100 donor plants. When collecting cuttings from dioecious species, male and female plants should be equally represented (Landis et al. 2003). In special cases, seeds must be collected from parents that possess desired attributes. For example, if a native or traditional species is being grown for timber or craft use, seeds need to be collected from trees showing the desired form or wood characteristics. Plants grown for traditional medicinal purposes must meet the end user’s exacting requirements for quality and potency of the source plants. When applicable, these other source considerations are in addition to, not instead of, local adaptedness, genetic diversity, and sexual diversity. 6. What Types of Plant Materials (Stocktypes) Are Best Suited to the Project Site and Objectives? The characteristics of a particular species will help determine if direct seeding, nursery propagation, or other methods are the more appropriate strategy (Steinfeld et al. 2007). Many options can be considered (Table 2). Nursery stocktypes include container seedlings, bareroot seedlings, rooted cuttings, wildings, rootstock, nonrooted cuttings, and seeds (Table 3). • Seeds – Direct seeding is most successful for grasses, forbs, and some woody shrubs. Seeding with native grass species after wildfires is often used to stabilize soils and prevent erosion. Sowing seeds directly in the field can result in poor germination and survival. Therefore, direct seeding is recommended only for species in which efficient use of seeds is not necessary.

1056

K.M. Wilkinson et al.

Table 3 Examples of different types of plant materials that can be provided by nurseries Plant materials Seeds

Examples Grasses, forbs

Bareroot plants

Coconut (Cocos sp.), mahogany (Swietenia sp.), Cedrela sp.

Container plants

All species

Root stock

Yams (Dioscorea sp.), kava (Piper methysticum), breadfruit (Artocarpus sp.), bananas (Musa sp.) Erythrina sp., Gliricidia sp., gumbo-limbo (Bursera simaruba), Guazuma ulmifolia

Nonrooted cuttings

Layer cuttings

Citrus sp., breadfruit (Artocarpus sp.)

Rooted cuttings

Many species

Advantages • Small and easy to outplant • Seeds of some native plants can be stored for long periods • Plants develop natural root structure • Maintain genetic diversity • Less expensive to produce in the nursery than container plants • Easier to transport • Roots have not been restricted by containers • Well-established root systems means less transplant shock • Available in a variety of sizes • Can be planted all year long • Large stocktypes provide “instant” plants on site • Easy to store and transport • Excellent survival after outplanting • Ideal for live stakes • Can be efficiently and economically produced in nursery stooling beds • Do not have to rely on seed crops • Ideal for maintaining same genotype • Do not have to rely on seed crops • Ideal for maintaining same genotype • Stooling blocks can be developed for large multiyear projects

Disadvantages • Some species do not produce seeds regularly • Many tropical seeds do not store well • Direct seeding more inefficient use of seeds than nursery plants • Take longer to produce Roots dry out easily • Often lower survival and slower growth especially on drier sites • More expensive to propagate • More difficult to transport, especially larger stocktypes

• Only works with tap-rooted species

• Only work with species that root easily • Best for mesic environments • Require healthy mother plant • Only work with species that root easily • Must be handled carefully during transportation and outplanting

Planning and Managing a Tropical Nursery

1057

Fig. 5 Nursery stock can be grown in many sizes and shapes. Consider the advantages and drawbacks of different options when defining the target plant materials for a project (Illustration by Jim Marin adapted from Steinfeld et al. 2007)

• Bareroot plants – Bare root is suitable only for plants with root systems that can tolerate more disturbance and handling than species that need their root systems kept in contact with soil during harvest, shipping, and outplanting. • Container plants – Container plants are the stocktype of choice for many tropical nurseries. Because containers come in many sizes and shapes (Fig. 5), they can be matched to the project objectives and site conditions. When clients order container plants, age, stem diameter, height, root size, and depth are usually specified, in addition to species and seed source. • Wildings – Plants salvaged from areas, such as development sites or roadsides, can be an important component of protecting native plant diversity. Sometimes salvaged plants are simply relocated quickly from one area to another. At other times, plants may be transplanted into a nursery, cared for, and outplanted at a later time (Steinfeld et al. 2007). • Nonrooted cuttings – Long “pole” cuttings are a common type of nonrooted cutting used extensively in the tropics as live fence posts. Sometimes called

1058

K.M. Wilkinson et al.

“quick sticks,” these nonrooted stakes of easy-to-root species, such as Gliricidia sepium, Erythrina species, or Bursera simaruba, are cut from the major branches or stems of trees and inserted into the ground. • Rooted cuttings – Rooted cuttings are often used in tropical plant propagation. A stem section can be used to propagate plants from cuttings but it needs to have a healthy bud near the top (Dumroese et al. 2003). • Rootstock – Specialized roots, such as bulbs and corms, and modified underground stems, such as rhizomes and tubers, can be used for the vegetative propagation of certain grasses, wetland plants, food plants, and even some trees. 7. What Are the Best Outplanting Tools and Techniques? Each outplanting site has different climatic and soil conditions, and because no single tool or technique will work well under all site conditions, outplanting tools and techniques must be matched accordingly. Hand tools such as shovels, pick mattocks, planting hoes (“hoedads”), and planting bars are popular for outplanting tropical plants. Nursery managers must know in advance which planting tools will be used so they can develop proper plant material specifications, especially root length and volume or cutting length and diameter. 8. What Is the Best Time for Outplanting? Each site has an optimal time, the “outplanting window,” when chances for plant survival and growth are greatest. The outplanting window is usually defined by looking at the climate and historical information from question 2 and the limiting factors described in question 3. For example, in some tropical areas, soil moisture is the main limiting factor. In these cases, the outplanting window is at the onset of the rainy season when soil moisture is increasing. Areas without a dry season may have other limiting factors, such as heavy rainfall and flooding, which define outplanting windows differently. The specific dates of outplanting windows will change with latitude and elevation.

Case Study: Defining the Target Seedling The following is an example of how target plant requirements work with nursery crop planning. In March, a retired cattle rancher decides to start an Acacia koa plantation to be selectively harvested for fine wood products as an investment for her grandchildren. She wants to plant 500 Acacia koa trees on a former cattle pasture now fenced and free of cows. Her property is at 2,000-ft [610-m] elevation on the leeward side of the Big Island of Hawai’i. The area has been in pasture for more than a century. She had the soils analyzed and they are typical from a nutrient standpoint. She wants easy-to-plant trees. Although the rainy season starts in mid-November, she plans to plant with help from her family during their holiday, so her ideal delivery date is December 15. She plans to remove the grass from the planting areas and then plant the seedlings using a mattock and shovel.

Planning and Managing a Tropical Nursery

1059

With these details, the nursery manager and the client can work together to establish target plant requirements for the koa seedlings based on the eight target seedling questions: 1. Objective – Reforestation/plantation establishment with locally adapted native Acacia koa. 2. Site evaluation – Client’s soil analysis is unremarkable; client realizes control of existing vegetation (pasture grasses, weeds) is needed; client examined climate data and knows when the rainy season starts and ends. 3. Limiting factors on outplanting site – Likely shortage of viable population of appropriate beneficial local microorganisms (Rhizobium bacteria and mycorrhizal fungi), competing grass and weeds, and minor risk of cow or horse browsing. 4. Mitigating measures for limiting factors – Inoculate seedlings in the nursery with Rhizobium bacteria and mycorrhizal fungi. Before planting, client will remove competing vegetation and install fencing to prevent browsing and will be advised to mulch trees and to diligently control weeds. 5. Genetic considerations – Seeds sourced according to transfer guidelines from forestry department for locally adapted, genetically appropriate koa. Seeds will be collected (by permission) from a nearby koa forest at a similar elevation: collected from trees of good form and from a minimum 50 parent trees to ensure genetic diversity. 6. Type of plant materials – 500 Acacia koa seedlings from genetic sources listed previously. Containers 15-cm high, 3.8-cm diameter, 107 ml3; stem diameter 3.5 mm; roots, firm and nodulating with Rhizobium, inoculated with mycorrhizal fungi. Seedlings will be watered thoroughly while still in their containers immediately before outplanting. 7. Outplanting tool or technique – Outplant seedlings with a mattock and shovel. After outplanting, mulch seedlings with a biodegradable weed barrier topped with mulch that avoids contact with seedling stems. Flag trees with a bamboo stake and bright ribbon for ease of monitoring and maintenance. 8. Timing of installation/outplanting window – Although the rainy season usually begins in November, labor is not available until December, so the target date for seedling delivery is December 15.

Field Testing the Target Plant As described above, at the start of any planting project, the client and the nursery manager need to agree on certain morphological and physiological specifications based on answers to the eight questions that define a target plant. This prototype target plant is grown in the nursery, and its suitability is then verified by outplanting trials that monitor survival and growth. Problems with seedling quality, poor planting, or exposure to drought conditions result in plants gradually losing vigor and perhaps dying. Therefore, plots must be monitored during the first month or two after outplanting and again at the end of the first year for initial survival. Subsequent

1060

K.M. Wilkinson et al.

checks after 3- and 5-year periods will give a good indication of plant growth rates. This performance information is then used to give valuable feedback to the nursery manager, who can work with the client to refine the target plant specifications for the next crop (Fig. 3).

Nursery Management Tropical nursery management includes all aspects of growing plants, from propagule collection through all their growth phases and finally to outplanting (see Landis and Wilkinson 2014c). Management involves an understanding of practical, scientific, technical, interpersonal, and economic aspects of the nursery. Nursery management includes ordering materials and supplies, maintaining facilities, scheduling activities, keeping horticultural and financial records, solving problems, and more. In this section, we briefly describe key nursery topics where management is particularly important. Good nursery managers are adept at organization. One good starting point toward organizing crop production is to develop a checklist. This checklist can help provide an overview of the interrelated tasks involved in managing a nursery. The example checklist in Table 4 can be modified and customized to describe the daily, weekly, monthly, and seasonal tasks of a nursery. Only a few required tasks must happen each day; these tasks are the essential activities that keep the crops alive and healthy and the nursery functioning, such as watering, daily record keeping, and crop monitoring. Other important tasks need to be done less frequently. Good planning and oversight based on a checklist will ensure that all tasks are prioritized and scheduled accordingly. Schedule an overview and planning session on a weekly basis to assess immediate needs, periodic tasks, and long-term goals. This assessment provides an opportunity to prioritize tasks for the coming week and month. The observation skills of the nursery manager and staff are the greatest assets to effective planning. Keeping the nursery’s vision and objectives as a focus during meetings can help maintain a broader view for nursery activities. Taking time each week to review the daily log, plant development record, and other observations of the manager and staff will help prioritize necessary work and assign roles, tasks, and deadlines to the appropriate staff members. The needs of the plants, environmental conditions, and many other factors require flexibility and responsiveness in management style. Crops usually do not respond well to a rigid schedule and may perform differently in different years, which is why weekly assessment and planning is so important. Attempts to make rigid schedules (such as “weed every Tuesday”) are often far less effective than regular planning to tailor tasks to the observed needs and conditions of the crop. Some proactive planning should also occur, focusing beyond what is most urgent. Planning should include activities that are important to the nursery’s larger mission beyond the day-to-day details, such as public relations or educational activities. Updating plant protocols and working to improve plant quality with

Planning and Managing a Tropical Nursery

1061

Table 4 Example checklist for nursery management Planning and scheduling (weekly, monthly, yearly) Make a list of what needs to be done based on daily observations, daily logs, and crop development records Establish propagation protocols Create and update crop-growing schedules and facilities schedules Prioritize and delegate tasks Follow up to ensure tasks were done Refine nursery vision and objectives annually; anticipate new crops to grow, changes in production, infrastructure improvements or expansion, and other planning for the future Routine tasks (daily) Irrigate Crop culturing (e.g., weed or pest control, fertilizing) Monitor and observe the crops Record keeping (daily, weekly, or at the completion of growing each crop) Record observations and actions in daily journal or log (daily) Make notes in the plant development records for each crop (daily or weekly) Update and revise plant protocols (at end of each crop) Conduct crop inventory assessment (ongoing) Crop-phase production tasks (as needed for crops) Establishment tasks (e.g., making growing media, sowing seeds, inoculating with microsymbionts) Rapid growth phase tasks (e.g., fertilizing, monitoring) Hardening phase tasks (e.g., changing fertilization and light regimes) Update clients about crop development Harvesting, packing, and shipping tasks Seasonal cleanup (seasonal or between crops) Purge or transplant holdover stock Clean floors, tables, tools, equipment, and so on Clean and sterilize containers Check and repair equipment, tools, and infrastructure such as irrigation lines Financial management Determine expenses, including labor, time, and supplies needed to produce crops, and overhead costs (e.g., utilities) Determine estimated income Create and manage an annual budget based on anticipated income and expenses Administer contracts Inventory supplies for production and maintenance (e.g., growing media, fertilizers, containers and trays, irrigation parts) and order as needed Estimate future costs and income and adjust budget accordingly Problem-solving and troubleshooting Identify and analyze problems as they arise Know whom to call for help (e.g., another grower, a soil scientist, a pest expert, an irrigation specialist) and contact them as needed Develop and test hypotheses to solve the problem Implement a solution (continued)

1062

K.M. Wilkinson et al.

Table 4 (continued) Cultivating good relationships with staff, clients, and the community Provide staff education and training Connect staff with nursery vision and objectives Give and receive feedback and input (observations and improvement suggestions) Plan meetings, safety awareness, and so on Develop target plant specifications with clients Educate clients about key issues for handling, outplanting, and care Visit outplanting sites and clients to check up on survival and growth; follow up with clients to discuss field performance of plants, revisions of target plant criteria, and future needs Offer public education and outreach Learning and sharing Attend training events and conferences Learn from other nurseries; host and attend field days and visits Read published literature Explore ways to improve crop production and plant quality; reflect, plan, and adapt for the future

some simple experiments are also valuable activities that should be included in the nursery’s schedule. To ensure that plants germinate, grow, remain healthy, and become hardy to survive outside the nursery, all the plant’s environmental and nutritional requirements must be met while they are in the nursery. These requirements change as the plants develop. Nursery management coordinates time, resources, labor, and space to produce a healthy crop of plants on time (Wescom 1999). Developing propagation protocols and growing schedules that covers all phases of crop production and the time necessary to complete each step is an important part of nursery management (Wilkinson and Jacobs 2014).

Propagation Protocols A propagation protocol describes all the steps necessary to grow a species. It is meant to be a reliable, repeatable guide to producing and scheduling a crop of that species. It will also help you coordinate the production of all crops being grown simultaneously in the nursery. Protocols with the most detailed information make it easier to plan and schedule the next crop. After protocols are developed, they are refined with each crop, leading to improvements in nursery efficiency and effectiveness from season to season. Propagation protocols serve as an essential guide for planning and scheduling each crop. The protocol is developed using firsthand experiences and outside sources of information. If little to nothing is known about the species, the process of drafting a protocol can help you organize what is known and take an educated guess at how to proceed with growing a particular species. Start by systematically searching

Planning and Managing a Tropical Nursery

1063

relevant published literature. If specific information on the desired species is unavailable, try to find another species within the same genera or even a related species grown in similar climatic zones, to see if any information may be applicable. Next, gather information from observations of how the plant grows in nature. This information may be observed firsthand in the field and by local people who are familiar with the plant. Finally, seek advice from other nursery managers. Based on the information gathered, a protocol of what will be required to grow a species can be drafted. The plants themselves will prove the protocol right or wrong as they grow. Refine the propagation protocol after the production of each crop. Do not be discouraged if the first protocol drafted from research or another nursery’s experience does not produce the desired results; the goal is to adjust the protocol to reflect your nursery conditions. Year-to-year variations in weather or unforeseen operational changes often keep crops from growing exactly as projected. Allow room for flexibility and make adjustments to growing media ingredients, seed germination methods, irrigation practices, and so on, based on observed factors. Keeping records is key. Sometimes new information and discoveries will significantly improve production methods and will be added to the protocol. As the protocol is updated, the nursery develops an increasingly accurate and helpful guide for how to grow each species.

Propagule Collection, Processing, and Treatment For some tropical species, usually commercially important trees, recommendations for appropriate sources or seed zones may be available. For most other tropical species, the nursery will have to research, define, and obtain locally adapted and genetically appropriate propagules for their client’s site. For some species, it may be possible to collect throughout the year. For other species, a narrow window may be all that is available. Seeds of some species may store well for many years, enabling the nursery to develop collections for future use, whereas for some species, production may have to wait until seeds that cannot be stored become available. Techniques for collecting each species are developed based on published literature, local knowledge, and experience. To provide plants when the client needs them, the nursery manager has to account for the time needed to obtain the propagules and grow the plants. Propagules need to be processed immediately after collection. For most seeds, processing helps ensure that the seeds stay dormant until they are needed. Depending on the species, seeds may need to be processed in one or more of the following ways: extraction from pods or fruits, washing, winnowing, sorting, grading, and drying. Some seeds do not store and must be sown when they are fresh. Cuttings may need to be dried, soaked, placed in rooting beds, or other treatments (see chapter “▶ Collecting, Processing, and Treating Propagules for Seed and Vegetative Propagation in Nurseries”). Seeds of native species vary widely in dormancy, so seed treatments need to be scheduled properly. Processing

1064

K.M. Wilkinson et al.

and treatment requirements and the time and space needed to complete them need to be considered as part of the schedule.

Crop Growth Phases For planning purposes, it is essential to understand the growth phases of crops. The development of most nursery crops can be divided into three phases: establishment, rapid growth, and hardening. Plants in each phase have distinct requirements for light, water, nursery space, and the types of attention and labor necessary to keep them healthy (Table 5): Establishment – For plants grown from seeds, the establishment phase is defined as the time from sowing seeds until the growth of the first true leaves or primary needles (Fig. 6a). For plants grown from cuttings, the establishment phase extends from placing cuttings into containers through the initial development of roots and shoots. The establishment phase typically lasts 2–8 weeks, although species slow to germinate from seeds or root from cuttings may take 1 year or more. The main goal is to have uniform plant establishment and maximum survival. Rapid growth – During the rapid growth phase, plants, particularly their shoots, increase dramatically in size, often approaching target size (Fig. 6b, c). Plants need to be somewhat protected during this phase to minimize stress and encourage rapid (but not excessive) shoot growth. Hardening – During hardening, energy is diverted from shoot growth to root growth (Fig. 6d, e). It is important to develop a plant with stems that are firm, often brown in color, and lignified, supporting a well-developed crown with leaves extending more than three-fourths of the length of the stem. Leaves should be vigorous, healthy, and leathery, not succulent. The root system should be well balanced and not root bound. Hardened plants are conditioned to endure the stresses of shipping, handling, and outplanting and fortified so that they have the energy and nutritional reserves to survive and grow after outplanting. If the hardening phase is too short, plants will fail to reach the appropriate physiological condition and survival and growth after outplanting are compromised, even if plants have the correct physical size. Therefore, good crop planning ensures adequate hardening before on-time delivery to the client (Jacobs et al. 2014). Before shipment to the outplanting site, plants can be graded for size and quality according to established standards, outplanting objectives, or specifications agreed upon with the client, typically shoot height, stem diameter at the root collar, root plug integrity, physical injury, or disease. Plants meeting quality standards are considered shippable and this inventory can be shared with the client; others are culled and composted. For threatened or endangered species where every plant is valuable, undersized but otherwise healthy plants can be held for additional growth or transplanted into larger containers for future outplanting dates.

Planning and Managing a Tropical Nursery

1065

Table 5 The three general phases of crop development for seedlings (Landis et al. 1998). After the three phases of crop development are understood for a species, the growing and facilities schedules can be developed to meet crop needs during each phase Phase Definition

Establishment From germination through emergence and formation of true leaves

Duration

Typically 14–21 days for germination; 4–8 weeks for early growth • Maximize uniform germination • Fill containers efficiently • Maximize survival • Minimize damping-off • Protect from weather and pests • Keep temperature warm • Irrigate to keep soil “moist but not wet” • No or low fertilizer

Objectives

Special needs

Labor

• Monitor germination • Introduce beneficial microorganisms • Thin • Resow and transplant if necessary • Scout for pests and diseases

Rapid growth From emergence of true leaves to when seedling approaches target height. Rapid increase in size, particularly in terminal shoot Varies widely, typically about 10–20 weeks

Hardening Energy diverted from shoot to root growth; seedling reaches target height and stem diameter; seedling is conditioned to endure stress Varies widely by species, from 4 to 12 weeks

• Minimize stress • Encourage shoot growth • Maintain environmental factors near optimum levels • Monitor as seedling approaches target size and roots fully occupy container

• Slow shoot growth • Encourage root and stem diameter growth • Acclimate to outplanting environment • Condition to endure stress • Fortify for survival after outplanting

• Protect from stress • Monitor sun exposure • Irrigate appropriately • Fertilize properly

• Induce moderate moisture stress • Progressively expose to sun equivalent to outplanting conditions (full sun or partial shade) • Expose to ambient temperatures and humidity • Provide good airflow/ wind • Reduce fertilization rates and change mineral nutrient ratios • Manage water, fertilizer, and environmental conditions to induce hardening • Scout for pests and diseases • Communicate and plan to deliver crops to clients in a timely fashion to avoid problems with holdover stock

• Monitor environment • Modify crop density to encourage good development • Adjust culture to avoid excessive shoot size • Scout for pests and diseases

1066

K.M. Wilkinson et al.

Fig. 6 Shown is koa (Acacia koa) at different growth phases: germination and establishment (a), establishment phase 2 weeks after germination (b), early rapid growth phase (c), late rapid growth phase (d), hardening phase in outdoor area (e), and target plant ready to ship to client (f) (Photos by Douglass F. Jacobs)

Planning and Managing a Tropical Nursery

1067

Monitoring Crops The manager or a designated “crop monitor” must observe the crop every day, understand what “normal” is for that crop and for the nursery environment, and be sensitive to any deviations from that norm. Periodic measurements of plant growth (e.g., height) and environment conditions (e.g., daily high temperature) help define those deviations. Experience and daily observations can identify potential problems long before they become emergencies. Observations may include the following: • Appearance – Inspect the crop for appropriate plant size for the stage of growth, for nutrient deficiencies, for problems with insect or diseases, and for the development of beneficial microsymbionts. • Smells – Notice unusual smells, such as overheating motors, broken fans, composting troubles, and other potential problems. Some foliar diseases may be discernible to an experienced grower by odor. • Noises – Listen for unusual sounds, such as an engine running unsteadily or water running when it should not be. • Feel – Touch the growing media to determine if it is at the proper moisture level and sense whether the temperature and humidity in the nursery are appropriate. Although one designated crop monitor is responsible for this task, all staff need to understand that keen observation is key to precluding problems. The manager should welcome and encourage staff to share their observations and contribute to the daily log; this practice builds observation skills and greater crop awareness.

Problems with Holdover Stock The failure of clients to pick up plants on schedule is a common problem that can be avoided by good scheduling practices and communicating often with clients, especially by providing periodic updates to advise them when plants will be ready. In some cases, having penalties, such as storage fees, in the contract for late pickups may encourage clients to pick up their plants in a timely fashion. When communicating with clients, emphasize that prompt outplanting serves everyone’s best interest, not only for the nursery and the health of the plants but also for the success of the client’s project. When plants are held too long in the nursery, the root system becomes woody and loses its ability to take up water and nutrients. Structural problems may also occur; roots may spiral and, instead of expanding outward and downward into the soil after outplanting, these roots may girdle the plant or cause it to be unstable in high winds. Shoot growth may resume and negatively affect the root to shoot ratio and the plant loses its resistance to stress. Making a growing schedule as shown in Table 6 and sharing it with clients is a helpful way to provide plants on time and have them outplanted promptly.

None

600 viable seeds plus extras to store for future orders

Field collection of pods. Extract, dry, and clean seeds at nursery

Propagation environment Fertilization

Irrigation

Target size at end of phase

Actions

Seed collection and processing 2 days

Scout sites July 15 and collect if ready, if not, collect Aug 1 and process Aug 2 (Field-collection sites) None

Activity Length of phase Dates

Mechanical scarification (break seed coat with nail clipper); soak overnight in clean water

600 germinants

None

None

Aug 3 to scarify and Aug 4 to sow

Seed treatments 2 days

Make growing media; sow seeds; inoculate seedlings with rhizobia at 2 weeks of age

Germinant area protected from predators and harsh weather In growing media: triple super phosphate, dolomite lime, gypsum, and mycorrhizal fungi. Inoculate at 2 weeks with Rhizobium bacteria Daily gentle hand watering to keep moist Not applicable (they will usually be about 2.5–5 cm tall, but no target is set)

Aug 4–Aug 25

Establishment phase 3 weeks

Weed and pest management; check nodulation with rhizobia; space double in trays as soon as prudent; cull poor performers

Once daily by hand to saturation Approx. 10–12-cm height, 2.5–3-mm root-collar diameter

Main outdoor growth area and full sun None

Aug 26–Nov 3

Rapid growth phase 10 weeks

15-cm tall, stem diameter 3.5 mm; roots, firm and nodulating with rhizobia If to be outplanted in windy areas, “brush” daily to simulate wind and improve stem strength

Gradual reduction

Main outdoor growth area and full sun None

Nov 3–Dec 14

Hardening phase 6 weeks

Table 6 This example crop schedule indicates some necessary steps in each crop development phase and the time required to complete each. This schedule should be posted in the nursery so that staff can track the crop’s development and understand what cultural practices are required. If appropriate, the schedule can also be shown to clients so they fully understand the time required to produce their crop and when they will need to outplant

1068 K.M. Wilkinson et al.

Planning and Managing a Tropical Nursery

1069

Facilities Schedules The space, labor, equipment, and supplies required for each crop during the different stages of propagation must be included in crop planning; this will vary by crops and nurseries. Crops are often moved from one structure to another as they progress through the three development phases. Although the example in Table 7 does not go into such detail, the facilities schedule should include calculations for how much space each crop will use, how many hours of labor will be needed, and the quantities of materials (such as growing media) required during crop production. Facilities scheduling is indispensable in determining how resources within a nursery can be best distributed to maximize production and minimize conflicts associated with overlapping needs. The facilities schedule (Table 7) may be combined with or posted side by side with the crop schedule (Table 6), and the staff needs to have easy reference to it.

Scheduling Multiple Crops Most native plant nurseries must meet client needs by delivering a suite of diverse species on a single shipping date. Therefore, the growing and facilities schedules are essential to coordinate production of multiple crops of different species. The schedules must reflect the growth rates (time required to grow the crop) of each species. For example, without proper scheduling, the faster-growing species may become overgrown before the delivery date, or the slower-growing species may not be ready on time. Therefore, slow-growing species need to be sown earlier than fast-growing species so they will all be ready to plant at the correct time. In the hypothetical example in Fig. 7, the nursery is growing six species that need to be ready for the same outplanting date in December. To meet the target outplanting, Acacia koa needs to be sown in August, but if a client also wanted native sandalwood (Santalum sp.) trees in 4-L (1 gal) pots at the same time as the koa, nursery work would need to begin 1 year in advance to bring this slowergrowing species through all the phases of development.

Harvesting Plants are ready for harvest and delivery to clients after they have reached target specifications and have been properly hardened to withstand the stresses of handling and outplanting (Jacobs et al. 2014; Haase et al. 2014). Harvest timing needs to be coordinated with the client’s outplanting schedule and needs to coincide with optimum conditions on the outplanting site. Clear, frequent communication is essential to determine exactly when plants are needed for outplanting. It is crucial for the nursery staff to educate the client about the importance of timing and the consequences of hastening or delaying the harvesting window.

Materials needed

Facility/ space needed

Labor

Activity Length Dates

Seed collection and processing 2 days Scope July 15; collect Aug 1; process Aug 2 Two staff members to scout out collection site in July (1/2 day). Two staff to go to collection site on Aug 1 and collect (one full day per person). Two staff to dry, process, and clean seeds on Aug 2 (a few hours intermittently throughout day as pods dry) Permissions and permits to collect at collection site(s). A sunny, level area at the nursery to lay out pods for drying For collection: vehicle, pole pruner, pruning ladder, collection bags, maps, and written collection permissions For processing: a tarp and seed storage containers

Seeds, plus nail clippers, a clean container, and clean water for soaking the seeds

Indoors (home or nursery office)

Seed treatments 2 days Aug 3 to scarify; Aug 4 to sow One staff member to hand scarify seeds late in day on Aug 3 and put them in water to soak overnight

Table 7 An example of a facilities schedule for Acacia koa

Scarified seeds, containers and trays, potting media and amendments, mycorrhizal inoculant, fine-headed sprayer and hose for irrigation, and a blender and nodules to make rhizobia inoculant (on Aug 18)

Benches in germinant area (protected from seedeating birds/rodents and heavy rains)

Mix growing medium, fill containers, sow seeds, hand water daily, monitor germination, and protect from slugs, birds, and rodents. Collect rhizobia nodules and inoculate seedlings on Aug 18. Update client on progress

Establishment phase 3 weeks Aug 4–Aug 25

Extra trays for doublespacing seedlings and irrigation supplies

Move to outdoor growth area, irrigate, monitor growth, double-space in trays as seedlings grow larger, and manage weeds and pests. Update client as end of phase nears. If overstock, seedlings will be available, offer them to the client or look for another home for them Benches in main outdoor growth area

Rapid growth phase 10 weeks Aug 26–Nov 3

PVC pipe or bamboo pole for brushing, irrigation supplies, and boxes or buckets for transport – unless they will be transported in trays

Benches in main outdoor growth area

Monitor growth, monitor and gradually reduce irrigation, and brush to encourage stem diameter. Keep in regular touch with client. Schedule a pickup date and time

Hardening phase 6 weeks Nov 3–Dec 14

1070 K.M. Wilkinson et al.

Planning and Managing a Tropical Nursery

1071

Fig. 7 In this hypothetical schedule, six species must be ready for outplanting at the same time. Different stocktypes and species may require more or less time to grow in the nursery depending on many factors (Illustration by Jim Marin)

Plants destined for understory plantings or partially shaded sites can be held in a shadehouse until they are shipped. Plants that have been hardened for full-sun conditions can be held in an open compound because they may lose their conditioning if stored too long in the shade. Both structures are typically equipped with a reliable water source so irrigation and fertigation are possible.

Packing and Shipping Tropical container seedlings are typically shipped in their containers. Seedlings grown in trays in racks can be shipped as is while seedlings in individual polybag containers or plastic pots can be placed in open-topped boxes or crates to minimize toppling and protect against mechanical injury (Fig. 8). Nurseries may wish to charge a deposit or develop some other method to ensure containers are returned to the nursery for reuse. Plants need to be packed for shipping in a manner that encourages air exchange and allows for possible irrigation on the outplanting site. Restricted airflow can trap the heat generated by plant respiration and result in damaging stresses. After the plants are graded and packed, the final step before shipping is to clearly mark each

1072

K.M. Wilkinson et al.

Fig. 8 Plants need to be packed to minimize toppling and protect against mechanical injury during shipping (Photo (a) by Ronald Overton and photo (b) by J.B. Friday)

group of plants with the species, seedlot, number of plants, and other important information. Most nursery stock is delivered by truck. Containers in racks, pots, or polybags are usually placed carefully on the floor or stacked on metal or wooden shelves inside the delivery vehicle. Nursery plants can be subjected to severe mechanical shocks during transport (McKay et al. 1993), especially on gravel or dirt roads, and reducing speed will minimize potential injury (Stjernberg 1997). If shipped by boat, protect plants from sea spray, which can severely damage foliage and roots. Regardless of the vehicle used for shipping, air circulation created by spacers (such as wooden boards or foam blocks) between racks, pots, and boxes can be used to reduce heat buildup and to prevent the load from shifting and crushing.

Handling Tropical plants are difficult to store because they do not achieve a deep dormancy condition and can be vulnerable to handling stresses. Plants always need to be handled carefully to minimize effects of moisture, temperature, or physical stresses. It is critical that plants be transported quickly and outplanted as soon as possible after leaving the nursery. Properly handling nursery stock ensures that it has the best chance for survival and growth after outplanting. During packing and shipping, plants may be exposed to damaging stresses including desiccation, extreme temperatures, or mechanical injuries. This period holds the greatest financial risk because plants have reached their maximum value right before shipping (Paterson et al. 2001). Everyone handling the nursery stock must realize the plants are alive and perishable and must always be treated with utmost care.

Planning and Managing a Tropical Nursery

1073

Fig. 9 Nursery plants are subjected to a series of stresses from the time they are harvested to when they are outplanted. Each stage in the process represents a link in a chain, and overall plant quality is only as good as the weakest link (Illustration by Steve Morrison)

Desiccation is the most common stress encountered during handling, shipping, transport, and outplanting and can have a profound effect on survival and growth. As air temperature and wind speeds increase, so does moisture loss from seedlings, with significant losses in as little as 5 min. Roots are especially vulnerable to desiccation because, unlike leaves and needles, they have no waxy coating or stomata to protect them from water loss. Fine root tips have greater moisture content than woody roots; if fine roots appear dry, then they are probably already damaged or dead. Although roots of container plants are protected somewhat by the growing medium, which serves as a reservoir of water and nutrients, desiccation can still be severe when the plug gets too dry. Desiccation of container plants can be avoided by ensuring plugs are kept moist (but not saturated) throughout their journey from nursery to outplanting. Temperature extremes can quickly reduce plant quality, too. Plants are alive and respiring. When exposed to warm temperatures, their respiration adds heat to their environment; this condition is particularly serious when air circulation is inadequate. Physical damage can occur from dropping, crushing, vibrating, or simply rough handling. It is useful to think of plant quality as a chain in which each link represents one of the sequences of events from harvesting at the nursery until outplanting (Fig. 9). Because all types of abuse or exposure are cumulative, nursery plant quality can be thought of as a checking account. Plants at the nursery are at 100 % of quality; all subsequent stresses are withdrawals from the account. Note that it is impossible to make a deposit; nothing can be done to increase plant quality after a plant leaves the nursery. Therefore, care must be taken during all the harvesting and shipping processes to help ensure outplanting success.

Record Keeping Propagation protocols, growing schedules, and facilities schedules can be improved by keeping good records. Propagation protocols should be revised on a seasonal or

1074

K.M. Wilkinson et al.

annual basis, becoming more accurate each time, which then improves the usefulness of the crop and facilities schedules. The best way to improve the accuracy of protocols is through written records. A daily log is simply a journal that notes nursery conditions, activities, and management practices on a daily basis. In addition, plant development records describe the development of each crop over time in response to management practices. Keeping a simple plant development record for each crop is a great way to build a foundation for accurate, site- and species-specific protocols. Some growers choose to record a large amount of detail in a daily log and then later summarize information about each crop into a plant development record. Others jot notes into a plant development record for each crop on a regular basis, weekly, monthly, or when the crop is entering a new stage in its development. One way to make it easier to keep track of this valuable information is to use a form such as the example in Table 8. Keep a separate plant development record for each crop grown, even if it is just a small trial of a few plants or even if the species has been grown before. Any intended improvements, trials, or experiments done with a species can also be recorded in a plant development record. When these records are reviewed, the information enables nursery managers to determine if intended improvements actually had a positive effect on plant health and growth compared with what was normally done.

Problem Solving Good management, staff training, monitoring, and planning will generally preclude emergency situations in the nursery. Even the best manager, however, cannot avoid problems entirely. Some problems, such as difficulties with the irrigation system, appear suddenly and must be handled instantly. Others require a longer-term approach. With experience, troubleshooting problems may become easier. Do not be reluctant to reach out to colleagues, other nursery managers, or other professionals. Everyone has problems once in a while, and we can all help each other learn more about plants as we share our experiences. This five-step systematic approach can be helpful when approaching long-term challenges (Landis 1984): 1. Identify the problem – Is it really a problem? What seems to be wrong? 2. Analyze the problem – What happened exactly? When did it start? 3. Generate ideas – Identify potential sources of the problem. Consult the literature, other nurseries, staff members, or outside sources of help such as extension agents or specialists. 4. Develop and test hypotheses – At some point, a conclusion about the source of the problem must be decided and acted upon. 5. Implement a solution – Decide on a way to solve the problem. Observe the results. If the problem is not solved, start again with step 2.

Planning and Managing a Tropical Nursery

1075

Table 8 An example of a plant development record for an Acacia koa crop (Adapted from Wilkinson and Elevitch (2003, 2004)) Plant development record Species name Propagule source

Date(s) of propagule collection Establishment Type and length of propagule treatment (e.g., scarified, stratified) Date of propagule establishment Growing media and tray or container type used

Germination notes (including date begins and ends, percent germination) Cutting notes (e.g., special conditions, hormone treatments) Date transplanted (if not direct sown) Container type and potting media for transplanting Microorganisms used? Misting/irrigation (type and frequency) Fertilization type, rate, and frequency, if any Environmental conditions for crop (light, temperature, humidity) Horticultural treatments (e.g., cultivation practices) Date establishment phase completed Notes (resowing or thinning activities, problems or challenges) Rapid growth Time after sowing and sticking to enter rapid growth phase Plant size at start of phase (height) Container type and potting medium Irrigation (type and frequency, e.g., daily, every other day, and so on) Fertilization type, rate, and frequency Environmental conditions for crop (light, temperature, and humidity)

Crop name: Koa crop for Waimea Ranch due Dec 15 Acacia koa Collection site ABDA, at 2,000-ft (610-m) elevation the leeward side of the Island of Hawai’i from 55 parent trees of good form July 15 and Aug 1 Hand scarified then soaked in water overnight Aug 2–3 Direct sown in containers (15-cm deep, 3.8-cm diameter, 107 ml3); growing media 50 % coir, 25 % perlite, and 25 % vermiculite with amendments and mycorrhizal fungi Fairly uniform germination of about 92 % from Aug 5 to 12 N/A N/A N/A Mycorrhizal fungi inoculant in media; inoculated with Rhizobium on Aug 20, 2014 Daily hand watering N/A Under plastic cover of screened in greenhouse during establishment, so birds and rodents do not eat the seeds N/A Aug 26 N/A

3 weeks (seeds scarified Aug 1 and moved to rapid growth area Aug 25) About 5-cm tall Same Daily hand watering N/A Moved to full sun Aug 27 (continued)

1076

K.M. Wilkinson et al.

Table 8 (continued) Plant development record Horticultural treatments (e.g., spacing, cultivation practices) Date rapid growth phase completed Notes (development, vigor and health, challenges or problems) Hardening Plant size at start of phase (height and stem diameter) Irrigation (type and frequency, e.g., daily, every other day, and so on) Fertilization type, rate, and frequency, if any Environmental conditions for crop (light, temperature, and humidity) Horticultural treatments (e.g., spacing, cultivation practices) Plant size at end of phase (height and root-collar diameter) Date hardening phase completed Date plants delivered Notes (vigor and health, challenges or problems) Other notes Notes on performance of crop after outplanting

Crop name: Koa crop for Waimea Ranch due Dec 15 Spaced out 50 % on Sep 9 Nov 3, 2014 Culled about 5 % that were not growing quickly before beginning hardening phase on Nov 3 Average about 10–12-cm tall, 2.5–3.0-mm stem diameter on Nov 3 Approximately every other day hand watering or as needed (very hot days will water daily) N/A Full sun, ambient temperature Some weeding. Weather has been windy; no need to brush 15-cm tall; 3.5-mm stem diameter Dec 10 – ready to go Dec 15

Client will be using a mattock and shovel to plant; follow-up phone call scheduled for Jan 5

N/A not applicable Acknowledgments This chapter draws heavily on Wilkinson et al. (2014), and we thank Brian F. Daley, Douglass F. Jacobs, and Tara Luna for their contributions.

References Bawa KS (1990) Plant-pollinator interactions in tropical rain forests. Ann Rev Ecol Syst 21:399–422 Davis AS, Jacobs DF, Dumroese RK (2012) Chapter 15: Challenging a paradigm: toward integrating indigenous species into tropical plantation forestry. In: Stanturf J, Lamb D, Madsen P (eds) Forest landscape restoration: integrating natural and social sciences. Springer, Dordrecht, pp 293–308 Dumroese RK, Wenny DL, Morrison SL (2003) A technique for using small cuttings to grow poplars and willows in containers. Native Plants Journal 4:137–139 Guinon M (1993) Promoting gene conservation through seed and plant procurement. In: Landis TD (tech coord) Proceedings, Western Forest Nursery Association, US Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station, General Technical Report RM-211, Ft. Collins, CO, pp 38–46

Planning and Managing a Tropical Nursery

1077

Haase DL, Landis TD, Luna T (2014) Harvesting and shipping. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 303–311 Holl KD, Aide TM (2011) When and where to actively restore ecosystems? For Ecol Manage 261:1558–1563 Jacobs DF, Landis TD, Wilkinson KM (2014) Hardening. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 293–301 Landis TD (1984) Problem solving in forest-tree nursery with emphasis on site problems. In: Duryea MS, Landis TD (eds) Forest nursery manual production of bareroot seedlings. Martinus Nijhoff/Dr. W. Junk, The Hague, pp 307–314 Landis TD, Wilkinson KM (2014a) Defining the target plant. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 45–65 Landis TD, Wilkinson KM (2014b) Water quality and irrigation. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 207–229 Landis TD, Wilkinson KM (2014c) Nursery management. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 351–359 Landis TD, Lippitt LA, Evans JM (1993) Biodiversity and ecosystem management: the role of forest and conservation nurseries. In: Landis TD (tech coord) Proceedings, Western Forest Nursery Association, US Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station, General Technical Report RM-221, Fort Collins, CO, pp 1–17 Landis TD, Tinus RW, McDonald SE, Barnett JP (1994) The container tree manual. Volume 1, nursery planning, development, and management. US Department of Agriculture, Forest Service, Agriculture Handbook 674, Washington, DC, 188 p Landis TD, Tinus RW, Barnett JP (1998) The container tree manual. Volume 6, seedling propagation. US Department of Agriculture, Forest Service, and management, Agriculture Handbook 674, Washington, DC, 166 p Landis TD, Dreesen DR, Dumroese RK (2003) Sex and the single Salix: considerations for riparian restoration. Native Plants Journal 4:110–117 McKay HM, Gardiner BA, Mason WL, Nelson DG, Hollingsworth MK (1993) The gravitational forces generated by dropping plants and the response of Sitka spruce seedlings to dropping. Can J Forest Res 23:2443–2451 Mexal JG, Cuevas Rangel RA, Negreros-Castillo P, Paraguirre Lezama C (2002) Nursery production practices affect survival and growth of tropical hardwoods in Quintana Roo, Mexico. For Ecol Manage 168:125–133 Mijnsbrugge KV, Bischoff A, Smith B (2010) A question of origin: where and how to collect seed for ecological restoration. Basic Appl Ecol 11:285–299 Mollison B, Slay RM (1991) Introduction to permaculture. Tagari, Tyalgum, 198 p Paterson J, DeYoe D, Millson S, Galloway R (2001) The handling and planting of seedlings. In: Wagner RG, Colombo SJ (eds) Regenerating the Canadian forest: principles and practice for Ontario. Ontario Ministry of Natural Resources and Fitzhenry & Whiteside Ltd, Markham, pp 325–341 Stape JL, Gonҫalves JLM, Gonҫalves AN (2001) Relationships between nursery practices and field performance for Eucalyptus plantations in Brazil. New For 22:19–41

1078

K.M. Wilkinson et al.

Steinfeld DE, Riley SA, Wilkinson KM, Landis TD, Riley LE (2007) Roadside revegetation: an integrated approach to establishing native plants. Western Federal Lands Highway Division, Vancouver, WA, 413 p Stjernberg EI (1997) Mechanical shock during transportation: effects on seedling performance. New For 13:401–420 Sutton R (1980) Evaluation of stock after planting. N Z J For Sci 10:297–299 Wescom RW (1999) Nursery production scheduling. In: Wescom RW (1999) Nursery manual for atoll environments. SPC/UNDP/AusAID/FAO Pacific Islands Forests and Trees Support Programme, RAS/97/330. Working paper 9, pp 12–14 Wilkinson KM, Daley BF (2014) Why start a tropical nursery for native and traditional plants? In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 17–28 Wilkinson KM, Elevitch CR (2003) Growing koa: a Hawaiian legacy tree. Permanent Agriculture Resources, Holualoa, HI, 104 p Wilkinson KM, Elevitch CR (2004) Propagation protocol for production of container Acacia koa Gray plants: permanent agriculture resources, Holualoa, Hawai’i. In: Native plant network. University of Idaho, College of Natural Resources, Forest Research Nursery, Moscow. http:// www.nativeplantnetwork.org. Accessed Sept 2009 Wilkinson KM, Jacobs DF (2014) Crop planning: propagation protocols, schedules, and records. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 67–84 Wilkinson KM, Landis TD (2014) Planning a tropical nursery. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 31–43 Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) (2014) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, 376 p Williams MI, Dumroese RK (2013) Preparing for climate change: forestry and assisted migration. J For 111:287–297

Collecting, Processing, and Treating Propagules for Seed and Vegetative Propagation in Nurseries R. Kasten Dumroese, Diane L. Haase, Kim M. Wilkinson, and Thomas D. Landis

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Propagule Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dormancy in Orthodox Seeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Treatments to Overcome Seed Dormancy and Enhance Germination . . . . . . . . . . . . . . . . . . . . . . . Cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scarification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Soaking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Germination Stimulators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Moist Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Environmental Factors Influencing Germination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sowing Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Direct Sowing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sowing Germinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transplanting Emergents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seed Coverings (Mulch) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vegetative Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shoot or Stem Cuttings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1081 1081 1082 1082 1084 1085 1085 1086 1087 1088 1088 1089 1090 1091 1092 1093 1094 1096 1097

R.K. Dumroese (*) USDA Forest Service, Rocky Mountain Research Station, Moscow, ID, USA e-mail: [email protected] D.L. Haase State and Private Forestry, USDA Forest Service, Portland, OR, USA e-mail: [email protected] K.M. Wilkinson Gibsons, BC, Canada e-mail: [email protected] T.D. Landis Native Plant Nursery Consulting, Medford, OR, USA e-mail: [email protected] # Springer-Verlag Berlin Heidelberg (outside the USA) 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_93

1079

1080

R.K. Dumroese et al.

Collecting, Transporting, and Storing Cuttings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Rooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cutting Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rooting Hormones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Striking, Monitoring, and Transplanting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Vegetative Propagation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Root Cuttings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Layering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Offsets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Grafting and Budding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1099 1100 1101 1102 1103 1106 1106 1106 1108 1108 1109 1110

Abstract

Tropical plants can be regenerated either sexually (from seeds) or asexually (from vegetative portions of a donor plant). These sources are collectively called “propagules” and link the evolutionary processes of the past with the potential for future adaptation. Thus, the way propagules are collected, processed, and treated can have strong repercussions on the future health and diversity of tropical ecosystems. Matching propagule source with outplanting site and collecting a genetically and, when necessary, sexually diverse group of propagules ensure the future adaptability of the plants and success of reforestation and restoration projects. Tropical seeds can be classified based on dormancy and potential for storage. Viviparous seeds germinate before dispersal from the mother plant. Recalcitrant seeds can germinate immediately once dispersed from the mother plant but lose viability when dried. Similarly, intermediate seeds can germinate immediately upon dispersal, but they also retain viability after some drying. Orthodox seeds can be dried without losing viability because they are “dormant.” A variety of methods exist to treat seeds to ensure germination. Methods for planting properly treated seeds include direct sowing of seeds, sowing germinants (sprouts), or transplanting emergents. Whichever technique yields the most efficient use of nursery resources depends on the species and the abundance and quality of the seeds. Vegetative propagation uses stems, leaves, roots, or other portions of a single mother (donor) plant to produce genetically identical daughter plants and is commonly employed if seeds are unavailable or difficult to germinate or if some special characteristic of the donor plant needs to be exploited. Vegetative propagation is usually more labor intensive than seed propagation and therefore more expensive, especially if special propagation structures are required. The best portion of the mother plant to use, as well the best timing for collection and proper propagule treatment, varies by species and requires experience to discern. While some species readily grow from stem cuttings, others grow better from root cuttings, divisions, or layering. Grafting and micropropagation are options for rare and endangered species or others that are difficult to propagate by other methods.

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1081

Keywords

Seed dormancy • Seed storage • Asexual reproduction • Cuttings • Germination • Genetics

Introduction In nurseries, plants can either be propagated sexually (from seeds) or asexually (from vegetative portions of a mother [donor] plant). Each method has its advantages and disadvantages. Often, the propagation method is most dependent on the natural reproductive traits of the desired plant. If the desired species produces abundant seeds on a regular basis and those seeds do not have a complicated dormancy, then growing the plants from seeds is advised. This method is usually less expensive than vegetative propagation and yields more genetic diversity in the resulting crop, especially when seeds are collective from a large number of individuals. If the seeds can be stored, that would be another advantage. If, however, the species regularly produces a scant amount of seeds and only produces seeds irregularly and those seeds cannot be stored, it has a very complicated seed dormancy mechanism, and/or if specific traits are desired (e.g., stem quality, fruit production, biomass), vegetative propagation may be best. On one hand, genetic diversity can be improved by harvesting propagules from many mother plants. If the species is dioecious (meaning individual plants are either male or female), collecting from a large population should ensure sexual diversity too. On the other hand, vegetative propagation can ensure desired traits in a single mother plant that are passed along to the daughter plants, as the offspring will be genetically identical. Thus, some forestry applications, such as developing seed orchards or plantations (e.g., palm oil production), depend on vegetative propagation to ensure inclusion of desired parents and their traits. The topic of propagule collection is large. As evidence, multi-chapter books are devoted to seed collection and processing (e.g., Schmidt 2007), and many other books discuss how this topic interrelates and is essential to overall nursery production (e.g., Wilkinson et al. 2014). Therefore, in this chapter our objective is to provide a broad overview of important facets of this topic and remind the reader that specific information about growing the resulting plants can be found in chapters “▶ Planning and Managing a Tropical Nursery” and “▶ Tropical Nursery Concepts and Practices,” as well as in Landis et al. (1998), Dumroese et al. (2008, 2012), and Wilkinson et al. (2014).

Propagule Collection Because forests in the tropics are diverse (e.g., humid, dry, lowland, mountain), the associated forestry activities are too. Forestry objectives range from reforestation of abandoned pastures (e.g., Weber et al. 2008) to establishment and management of

1082

R.K. Dumroese et al.

monocultures of palm oils developed through breeding (e.g., Basiron 2007), to interplanting plantations of nonnative species with native species (e.g., Davis et al. 2012), to sustainable silvicultural practices in natural stands (e.g., Putz et al. 2012), and to restoration within and adjacent to bioreserves (e.g., Luna 2003). This continuum of forest restoration activities from completely degraded to nearly “natural,” driven by a dynamic and complex set of human values, is diverse (Stanturf et al. 2014). Thus, a diversity of plant materials is needed as well. In this chapter, we maintain a conservative approach to propagule collection in that retention of biodiversity is paramount. While recognizing that traditional forest paradigms concerning tree improvement work have an important role, preserving species diversity (and genetic diversity within those species) through proper propagule collection techniques has an equally important role in fostering the preservation of diverse, strong, well-adapted populations regardless of the ultimate forestry objective. Seeds and other propagules link the evolutionary processes of the past with the potential for future adaptation (Flores 2002). Given the dramatic loss of biodiversity in the “Anthropocene” and recognition that maintaining biodiversity of species, and genetic diversity within species, is an important strategy for meeting the challenges of a changing climate (e.g., Mawdsley et al. 2009), we encourage propagule collection practices that reverse trends of genetic degradation and species loss and protect genetic diversity at collection sites and in locations where nursery stock is outplanted (Luna and Wilkinson 2014). Foresters have, for more than a century, recognized that “locally adapted” sources generally perform the best (e.g., Johnson et al. 2010). Thus, as much as possible, propagules should be collected from a nearby habitat similar in elevation, aspect, and soils to that of the outplanting site to ensure local genetic adaptation (e.g., Landis 2008). Working with locally adapted plant sources is important not only for the survival and health of the plants but also for the native birds, insects, and animals that depend on the plants. For situations where forest restoration and ecosystem restoration intersect, ensure that the source population is of wild origin (e.g., USDI BLM SOS 2011). Because reproductive strategies vary by species, no standard collecting procedure exists that will ensure genetic integrity for all species, but guidelines presented in Table 1 are a good starting point. Schmidt (2007) and Luna and Wilkinson (2014) provide more details on a variety of seed harvesting techniques, and Luna and Haase (2014) discuss techniques for collecting vegetative propagules.

Seed Propagation Seed Characteristics Seeds of many species are challenging to germinate. One way to reduce the challenge is to learn as much as possible about the life history, ecology, and habitat of the species being grown because understanding the processes seeds go through in nature can provide sound clues to improve germination in the nursery. Tropical

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1083

Table 1 Collection techniques to safeguard genetic diversity of native species. Although specifically addressing seed propagation, several guidelines presented here apply to vegetative propagation as well (USDI BLM SOS 2011) Seed collection method Assess the target population and confirm that a sufficient number of individual plants have seeds at natural dispersal stage Monitor seed maturation and carefully examine a small, representative sample of seeds using a cut test and, for smaller seeds, a hand lens for insect damage and empty seeds Collect seeds equally from more than 50 random individuals across the population For trees, collect seeds or fruits equally from all parts of the crown – top, sides, and bottom. Gather from individual trees at least 50 m distant from each other (Dawson and Were 1997) Return to the site to collect seeds throughout the dispersal season. Repeated collections taken from the same population during a single collecting season may be combined but do not mix collections between locations or years. Note multiple dates of collections on the seed label Collect no more than 20 % of the viable seeds available at the time of collection Clearly label all bags (inside and out) with the species, date collected, location of collection, and name of collector

Rationale Ensure that adequate genetic diversity can be sampled from the population and that the seeds are likely to be at maximum possible viability and longevity Estimate the frequency of empty or damaged seeds and confirm that most seeds are mature and fully formed and collection is worthwhile. (Seed development can vary within and among populations of the same species) Capture the widest possible genetic diversity from the plant population and to avoid only a few genotypes being propagated Ensure genetic diversity, as different flowers within the crown may have been pollinated at different times by different pollinators. Distance recommendation is to avoid collecting from closely related individuals (Dawson and Were 1997) Maximize genetic diversity in the collection, capturing early, middle, and late bloomers. (These different dispersal times may also interact with different pollinators)

Ensure that the sampled population has sufficient seeds needed for natural recruitment of new plants and to sustain local fauna Ensure that each collection is properly identified so seeds can be used appropriately in restoration efforts

seeds can be divided into four categories related to their level of dormancy and their ability to be stored (Hong and Ellis 2002; Kettle et al. 2011). Viviparous seeds germinate before they disperse from the mother plant. The most common examples are some genera of mangroves, such as Avicennia and Rhizophora, and some tropical legumes. Recalcitrant seeds can germinate immediately once dispersed from the mother plant but lose viability when dried. Most species from the wet, humid tropics have recalcitrant seeds because conditions in these environments are consistently favorable for germination and seedling establishment. Examples of common species with recalcitrant seeds include cacao (Theobroma cacao), mango (Mangifera indica), longan (Dimocarpus longan), and jackfruit (Artocarpus heterophyllus). Similar to recalcitrant seeds, intermediate seeds can germinate immediately once dispersed from the mother plant, but these seeds can experience

1084

R.K. Dumroese et al.

Fig. 1 Key to dormancy types. Knowing the type of seed dormancy is essential to successful seed propagation (Illustration by Jim Marin)

partial drying without losing viability. For example, papaya seeds (Carica papaya) dried to 10 % moisture content have been stored 6 years at 50 % relative humidity without affecting viability (Bass 1975). Species with intermediate seeds include neem (Azadirachta indica), cinnamon (Cinnamomum spp.), citrus (Citrus spp.), mahogany (Swietenia spp.), and coffee (Coffea spp.). Orthodox seeds can be dried without losing viability and are considered “dormant,” usually requiring specific treatments to encourage germination. Examples of species with dormant seeds include koa (Acacia koa) and Caribbean pine (Pinus caribaea).

Dormancy in Orthodox Seeds Tropical species inhabiting areas with a strong wet-dry seasonal cycle, arid or semiarid climates, or at high elevations subjected to cold temperatures often have dormant seeds (Fig. 1). Dormancy is an adaptation that ensures seeds will germinate only when environmental conditions are favorable for survival. The conditions

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1085

necessary to allow seeds to break dormancy and germinate can be highly variable among species, within a species, or among seed sources of the same species. This degree of variability is advantageous because seeds will germinate at different times during a period of days, weeks, months, or even years, ensuring that some offspring will be exposed to favorable environmental conditions for survival. External seed dormancy may be physical, chemical, mechanical, or physicalphysiological (Baskin and Baskin 1998, 2004). Physical dormancy, the most common dormancy type in the tropics, occurs when hard, thick seedcoats physically prevent water and oxygen movement into seeds. Species with physical dormancy include many of the legumes (Fabaceae), mallows (Malvaceae), fireadapted plants, and plants that inhabit arid to semiarid island habitats or areas with pronounced wet-dry seasonal cycles. Chemical dormancy describes fruits that contain high concentrations of germination inhibitors that prevent spontaneous germination of seeds. Mechanical dormancy describes tough, woody fruit walls that restrict seed germination and is best exemplified by the husks that surround coconuts (Cocos nucifera). Seeds with morphological dormancy have an underdeveloped embryo when dispersed and usually require a period of warm and moist conditions (afterripening) to allow the embryo to fully mature before germination is possible. Physiological dormancy is seen in arid and semiarid tropical environments. Seeds are permeable to water, but certain environmental conditions are necessary to modify the internal chemistry of the seed and promote germination.

Treatments to Overcome Seed Dormancy and Enhance Germination Many seed treatments, in response to the diversity of seeds grown in nurseries, have been developed. Pertinent information for particular species (or closely related species) may be available in the literature, on the Internet, and/or from other professionals. When information is lacking, personal observations made on the species in its habitat may provide useful clues on the best seed treatment. In general, nondormant seeds are planted immediately after collection and cleaning; intermediate seeds may be stored for several weeks or months in suitable conditions and then cleaned, fully rehydrated, and sown; and dormant (orthodox) seeds must be treated with one or more of the methods described below before seeds can be rehydrated, enabling germination.

Cleaning Seed cleaning helps prevent diseases in the nursery, especially for species that easily mold or require a long time to germinate. Often, molding is related to the most common nursery disease, damping-off. One of the easiest and best cleaning methods is to soak seeds in running water for 24–48 h, which flushes bacterial and fungal spores from the seeds (James and Genz 1981). This treatment can also help

1086

R.K. Dumroese et al.

Fig. 2 Mechanical scarification works well with large, easy-to-handle seeds. Great care is needed, however, so that the embryo or cotyledons are not damaged. This photo shows koa (Acacia koa) seed hand scarification (Photo courtesy of Wilkinson and Elevitch (2003))

remove any chemical inhibitors present on or within the seeds, and using an aquarium pump can agitate the seeds to improve the cleaning effect and keep the water well aerated. Seeds can also be cleaned with chemicals, some of which can stimulate germination. Bleach (5.25 % sodium hypochlorite) is the most common, and depending on the species, seeds can be soaked in solutions ranging from one part bleach in eight parts water to two parts bleach in three parts water for 10 min or less. Species with very thin seedcoats should not be cleaned with bleach. Hydrogen peroxide (one part peroxide in three parts water) can also be efficacious, e.g., Narimanov (2000); for example, Albizia spp. and Camphor tree (Cinnamomum camphora) seeds benefit from this treatment (Vozzo 2002).

Scarification Scarification is any method of disrupting an impermeable seedcoat so that water and oxygen can enter the seeds. Common scarification methods include mechanical, heat, and chemical. Mechanical scarification can be accomplished several ways. Large seeds can have an opening carefully made (to avoid damaging the internal portions of the seed) by filing or nicking their seedcoats individually by hand; examples include Acacia, Cassia, and Sesbania (Fig. 2). Small seeds can be placed into a shallow wooden box and rubbed under a block of wood covered with sandpaper. Often, however, the degree of scarification achieved with sandpaper can be variable. For larger quantities of seeds where treating individual seeds would be impractical, hobby-size rock tumblers can be used (Fig. 3). Seeds are tumbled, depending on species, for several hours to several days with coarse carborundum grit (sold by rock tumbler dealers) and pea gravel in the tumbler, either with or without water. A benefit of wet tumbling is that seeds are soaked in well-aerated water and chemical inhibitors may be leached from the seeds (Luna et al. 2014).

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1087

Fig. 3 Hobby-size rock tumblers can be used to scarify seeds and avoid seed destruction that can occur with sulfuric acid or heat scarification (Photo by Tara Luna)

Heat scarification, either wet or dry, can stimulate germination in species from fireadapted and other ecosystems. Using wet heat, accomplished by immersing seeds in boiling water, is an effective method for many small-seeded species because it provides a rapid, uniform treatment that can be assessed within a few hours. Wet-heat treatments are effective for many species including Acacia, Cassia, Senna, Sesbania, and Tamarindus (Vozzo 2002). Duration of exposure will vary by species because of differences in seedcoat thickness, so it is prudent to test a small subsample before treating the crop. In general, seeds are added to boiling water for 5–60 s (depending on species) and then immediately transferred to a vat of cool water so that they cool quickly to prevent embryo damage and begin imbibing water. After soaking for 1 day, seeds are ready for sowing or for stratification. For dry-heat scarification, seeds are placed in an oven at 80–120  C until the seedcoat cracks open, which can take from a few minutes to an hour depending on the species and thus needs to be closely monitored. Chemical scarification relies on acids to disrupt species with very thick seedcoats or stony endocarps. Sulfuric acid is commonly used, and treatment duration varies with the species and often among seed sources. Seeds must be carefully monitored to avoid damage to internal tissues. Sulfuric acid is very dangerous; using it requires special equipment, training, personal protective gear, and proper disposal after use. Some species can be easily damaged by sulfuric acid, but acceptable results can be achieved with citric acid or sodium or calcium hypochlorite baths with longer treatment durations.

Soaking Once cleaned and, if necessary, scarified, seeds require exposure to water and oxygen before germination can occur. The standard procedure is to soak seeds in water for 1 to several days until they are fully hydrated (often this can be accomplished during the cleaning process). Hydration can be checked by taking a sample, allowing it to dry until the seedcoat is still wet but dull, not glossy, and weighing it. When the weight no longer increases substantially with additional soaking time,

1088

R.K. Dumroese et al.

the seeds have absorbed sufficient water. Scarified seeds will be more obvious; the seeds will enlarge dramatically during the soak. Non-dormant seeds and those only having physical dormancy can be planted immediately after soaking. As described above, running water can help remove chemical inhibitors to germination, and an aquarium pump can keep the water well aerated. For seeds not soaked with running water, change the water often (at least a couple of times each day).

Germination Stimulators Seeds with physiological dormancy, once fully hydrated, may have their germination increased by exposure to some chemicals. Chemicals that stimulate germination include: • Gibberellic Acid – It is the most important plant hormone for the regulation of internal seed dormancy, and gibberellic acid can be purchased from horticultural suppliers and is generally applied at concentrations from 500 to 1,000 ppm depending on species. • Ethylene – Naturally occurring in plants, ethylene gas is readily released from ethephon, a commercially available product. Ethephon, used alone or in combination with gibberellic acid or other treatments, has enhanced germination in doum palm (Hyphaene thebaica) (Mousa et al. 1998) and may be used for other species inhabiting arid to semiarid tropical and saline environments. • Smoke – Smoke stimulates germination in many fire-adapted species, especially those that have thin, permeable seedcoats that readily allow entry of smoke (Keeley and Fotheringham 1998). Seeds can be fumigated, a method where smoke is piped into a specially constructed tent containing seeds, or seeds can be soaked in smoke-infused water. • Potassium Hydroxide – The optimum concentration of potassium hydroxide varies from 5.3 to 7.6 M for 1–10 min depending on the species; longer soaks at higher concentrations were found to be detrimental (Gao et al. 1998).

Moist Treatments Exposure to warm (22–30  C) and moist conditions enhances afterripening of seeds with underdeveloped embryos. Seeds are usually placed in moist peat moss, sawdust, coir, or another moisture-holding substrate. Although warm, moist treatments are not commonly used on tropical species, and it should be considered for seeds with morphological or physiological seed dormancy. The exposure to cold (1–5  C) and moist conditions to alleviate dormancy in seeds from temperate areas or high-elevation habitats in tropical regions is called stratification. Some subtropical species may also benefit from stratification. In climates with four seasons, seeds sown in flats or containers in late summer or autumn and left outdoors during winter undergo “natural” stratification. Otherwise,

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1089

Fig. 4 Small seeds requiring only a few weeks of stratification can be stratified by moistening paper towels and holding by corner to let excess water drain away (a) or placing seeds onto moistened towels inserted into an unopened plastic-zippered bag (b) (Illustrations from Dumroese and others (1998))

seeds can have “artificial” stratification when refrigerated within a moist substrate for a species-specific duration (Fig. 4a, b). Artificial stratification has several advantages: (1) seeds can be routinely checked for moisture and mold, (2) a large number of seeds can be stratified in a small space, and (3) seeds or seedlots can be removed and planted as soon as germination is noted. For these reasons, artificial stratification is generally preferred unless natural stratification provides higher rates of germination (Luna et al. 2014).

Environmental Factors Influencing Germination Light, water, oxygen, and temperature can all influence germination. In nature, seeds of pioneer species require high light levels associated with a canopy gap for germination and establishment, whereas shade-tolerant species generally can germinate in very poor light or deep shade. Thus, pioneer species with very small dustlike seeds need full light for germination to occur; even shallow burial (2 mm) may prohibit germination (Drake 1993). Conversely, some species are conditioned to germinate only if they are buried in the soil. Excessive watering can reduce oxygen levels in the medium and promote disease; both reduce germination. Conversely, lack of adequate moisture can delay or prevent germination. Therefore, seeds need to be kept evenly moist during germination. Optimum temperature for germination varies by species. Tropical species from high elevations or subtropical species generally germinate best at temperatures below 25  C, although they will often germinate across a wider range of temperatures (5–30  C). Most tropical species

1090

R.K. Dumroese et al.

require temperatures above 21  C, and some species germinate to their highest percentages when daily temperatures fluctuate from 20–30  C.

Sowing Methods Seeds may be sown three ways (Table 2) as described in the following sections. The most efficient method to use depends on species, type of seed, seed quality, and nursery environment. Table 2 Methods for sowing seeds (Landis et al. 1998)

Propagation method Direct sowing: Seeds are sown into containers

Planting germinants: Seeds sprouting or germinating in trays or bags are sown into containers, while roots are just beginning to emerge

Transplanting emergents: Seeds are sown into flats or seedbeds for germination; once germinated and leaves appear, seedlings are transplanted to containers

Good method for seeds with the following characteristics • Have a known high-percentage germination • Are inexpensive • Are in abundant supply • Have uniform, smooth shapes

• Are of unknown viability • Are valuable or rare • Have unknown germination requirements • Germinate during an extended period of time or during stratification • Are being tested but will not be transplanted to produce a crop • Do not respond well to other sowing methods • Have long or unknown dormancy

Advantages • Fast and easy • Economical • Minimizes seed handling • Seeds are all sown at once

• Efficient use of seeds • Efficient use of nursery space • Can adjust for unknown seed quality or performance

• Good for trials to observe seed performance • Useful with fibrousrooted species • Efficient use of seeds Efficient use of nursery space • Can adjust for unknown seed quality or performance

Disadvantages • Less efficient use of space, seeds, and/or growing medium • Causes of poor germination are difficult to track • May require thinning and/or consolidation and associated labor costs • Not good for large or irregularly shaped seeds • Labor intensive • May result in nonuniform crop development • Root deformation possible • Requires frequent, skilled monitoring

• Not recommended for woody and/or taprooted species because of problems with transplant shock and/or root deformation • Requires skilled labor

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1091

Fig. 5 Direct sowing works well for seeds that have little or no dormancy (or have been treated to overcome dormancy), are easy to handle, and are in abundant supply (a). Simple tools like a small plastic container (b) or a folded envelope (c) can be used to accurately sow small seeds of native plants (Photo a by Douglass F. Jacobs and photos b and c by Dawn Thomas)

Direct Sowing For direct sowing to be efficient, seeds must be easy to handle, abundant in supply, have simple dormancy treatments, and germinate at a high rate (Fig. 5). If direct sowing will be mechanized, seeds must also be uniform in size and shape. A germination test is the necessary first step in direct sowing. In general, seedlots having less than about 50 % germination are poor candidates for this method. Percentage germination observed in the test is used to determine the number of seeds to put in each container (typically 2–5) to ensure that most containers (90 % or more) will have a viable seedling (see Table 3). At some point, adding more seeds per container does not appreciably increase the number of containers with plants but does drastically increase the number of containers with too many plants and the amount of seed wasted (see Dumroese et al. 1998; Luna et al. 2008). Sometimes, particularly when seeds are scarce, expensive, or are expected to achieve nearly 100 % germination, single seeds are directly sown into containers. This practice ensures that every seed has the potential to become a plant and no thinning will be necessary.

1092

R.K. Dumroese et al.

Table 3 For a given seed germination, increasing the number of seeds sown per container increases the number of filled containers. In general, a target of 90–95 % filled containers is reasonable (Dumroese et al. 1998) Seed germination percentage 90 + 80–89 70–79 60–69 50–59 40–49

Seeds to sow per container 1–2 2 2 3 4 5

Percentage of containers with at least one seedling 90–100 96–99 91–96 94–97 94–97 92–97

Follow these steps for successful direct sowing: 1. Test the seedlot for germination. 2. Based on test results, determine the number of seeds to sow per container to meet production targets. 3. Cleanse and treat seeds as necessary to break dormancy. 4. Sow seeds, ideally centering seeds in each container and, if necessary, in the correct orientation for optimal growth and development. 5. Either apply mulch to the correct depth or avoid mulch depending on the light requirements of the species. 6. Gently water the seeds to settle them into the growing media. If multiple seeds are sown, containers with more than one germinant will need to be thinned as soon as possible after emergence in order to reduce seedling competition for resources. Extra seedlings can be clipped or pulled from the container. This labor-intensive practice can damage remaining seedlings if done improperly. Retaining the strongest seedling closest to the center should be done as soon as possible because it becomes more difficult as root systems develop. Discard culled plants into compost or waste and ensure that the remaining seedling has the best environment possible. If done soon after germination, thinned seedlings may be transplanted to empty containers as described below for transplanting emergents.

Sowing Germinants When done properly, germinant sowing, the practice of sowing sprouting seeds into containers, ensures that one viable seed is placed in each container thereby making efficient use of space and seeds (Fig. 6). This technique works best for seeds that are from a rare or valuable seedlot, have a low or unknown germination percentage, are large or irregularly shaped, germinate during stratification, germinate during a long period of time, or rapidly produce a long root after germination. Germinant sowing is a relatively simple process. Seeds are treated as necessary. Smaller seeds are placed between layers of moist paper towels or moist cardboard.

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1093

Fig. 6 Germinants are best sown as soon as the radicle emerges from the seedcoat (Photo by Tara Luna)

Fig. 7 When planting germinants, seeds must be sown as soon as the radicle is visible and must be oriented correctly when planted. Incorrect orientation leads to severe root deformation in woody species (Photo (a) by Thomas D. Landis and photo (b) by R. Kasten Dumroese)

Larger seeds are sometimes placed in plastic bags filled with a moist medium, such as Sphagnum peat moss or coir. Seeds are closely spaced, but far enough apart so that mold does not spread if it forms, and are checked every few days. As soon as the root (radicle) begins emerging from the seedcoat, the seed is carefully planted into the growing medium in the containers. Larger seeds can be planted by hand; smaller seeds are often sown using tweezers. If the radicle becomes too long, it may be difficult to plant without causing root deformation (Fig. 7). Once planted, the medium should be gently firmed around the root and the seed covered with mulch.

Transplanting Emergents Transplanting emergents is a practice for germinating many seeds in a small area. Seeds are sown in shallow trays that are usually filled with about 5 cm of growing medium. Soon after the seeds germinate, emergents are carefully removed from the tray by gently loosening the medium around them (Fig. 8). A small hole is made in the medium of the container and the germinant is carefully transplanted, ensuring

1094

R.K. Dumroese et al.

Fig. 8 Transplanting emergents works well for fibrous-rooted species. Great care must be taken to lift the emergent from the pricking-out tray without damaging the roots and to carefully and properly transplant it into the new container filled with a moistened growing medium (Photos by Tara Luna)

proper root orientation (Fig. 8). The potting medium is then firmed around the root and stem (Fig. 8). Transplanting emergents works best when: • Species have a fibrous root system that recovers well from transplanting (herbaceous forbs without a taproot, grasses, sedges, and rushes). • Seeds are too small or fragile to be sown by any other method. • Seeds have very complex dormancy or germinate during an extended period of time. • Limited nursery growing space makes direct seeding uneconomical. • Timing to transplant emergents is scheduled promptly. Although the exact size or age to transplant the germinating seedlings varies by species, it is usually done at the primary leaf stage (after cotyledons emerge) and well before root systems reach the bottom of the seed tray. This technique is not generally recommended for woody plants and other taprooted species, because if timed incorrectly or done improperly, transplanting emergents can produce a “J-root” or kink in the seedling stem or root (Fig. 9). These malformations can cause mechanical weakness, poor growth, or even mortality.

Seed Coverings (Mulch) Regardless of the seed sowing method, a seed cover or “mulch” is recommended to create an optimal environment for germinating seeds, unless the species requires light to germinate. Mulch is usually a light-colored, nonorganic material spread

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1095

Fig. 9 This palm seedling was permitted to get too large in the tray, causing the root to grow at a 90 angle. This root malformation will go on to cause poor growth and performance later in the plant’s life. Transplant emergents early to avoid root malformation (Photo by Brian F. Daley)

thinly over the seeds. Examples of mulches include very small rocks or grit (such as poultry grit), pumice, perlite, coarse sand, or vermiculite. When properly applied, mulches: • Create an ideal “moist but not saturated” environment around germinating seeds by making a break in the texture of the growing medium (water will not move from the medium into the mulch). • Keep seeds in place. This practice improves contact with the medium and minimizes the number of seeds washed out of the containers by irrigation or rainfall. • Reduce the development of moss, algae, and other weeds. The appropriate mulch depth varies by species; a general rule is to cover the seed twice as deep as the seed is wide. If mulch is too shallow, seeds may float away in the irrigation water. If the mulch is too deep, small plants may not be able to emerge above it (Fig. 10). Seeds requiring light need to be left uncovered. Tiny seeds should be left uncovered or barely covered with a fine-textured material such as fine-grade perlite or milled peat moss, coir, or compost. Uncovered and barely covered seeds must be misted frequently to prevent them from drying out. After lightrequiring and light-sensitive species have emerged and are well established, mulch can be applied to prevent moss, algae, and weed growth and to conserve moisture.

1096

R.K. Dumroese et al.

Fig. 10 A general rule of thumb for covering seeds with mulch is to cover the seed twice as deep as the seed is wide. Species requiring light for germination should never be covered with mulch, although mulch can be added after germination to reduce the growth of moss and liverworts (Illustration by Jim Marin)

Vegetative Propagation Many desirable and ecologically important tropical plant species are difficult or time-consuming to propagate by seeds but can be propagated vegetatively (i.e., asexually, without seeds). Vegetative propagation is commonly used with species that have short seed life, low seed viability, or complex seed dormancy. All new daughter plants that arise from vegetative propagation are genetically identical to the mother (donor) plant, and these resulting individuals are known as “clones” (Fig. 11); this makes vegetative propagation an attractive option when maintaining a desired characteristic, such as wood quality or fruit production, is an objective. Selection of mother plants must be done carefully as described in “Propagule Collection” above; specifically, origin of the source material must be adapted to the outplanting environment, collections should promote biodiversity within the resulting nursery crop, and material should be collected ethically to ensure sustainability of the source. Many tropical species are dioecious, meaning that plants are either male or female, so collecting both sexes is essential (Landis et al. 2003). Once an original collection has been made, it can be advantageous to maintain mother plants in a convenient location at the nursery to serve as a continual source of cutting material. The advantage is that it is more efficient to collect cuttings from mother plants at the nursery than collecting from wild populations, especially if the same ecotypes will be used for long-term projects. The disadvantage is that mother plants require nursery space and must be regularly managed. Mother plants are usually planted in field beds often referred to as “stooling beds” or “hedgerows.” Sometimes mother plants are kept in large containers. Regardless, mother plants must be clearly labeled as to species and origin. Mother plants should be pruned on an annual basis to maintain wood juvenility, discourage thick shoots or dominant leaders, and encourage production of numerous straight shoots to use as cutting

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1097

Fig. 11 Propagation of Plumeria spp. by cuttings produces several genetically identical plants (Photo by Tara Luna)

material. When pruning, it is important to leave sufficient foliage to sustain the root system. Application of water and fertilizer will keep mother plants vigorous. Stooling beds should be kept weed free.

Shoot or Stem Cuttings A cutting is the portion of a plant that is collected, treated, and planted to develop into a new intact plant complete with stems, leaves, and roots. Striking is the process of placing the cutting into soil or a rooting substrate. Shoot cuttings, also referred to as stem cuttings, are the most common type of cuttings and can be broadly placed into three categories depending on the stage of growth they are in when collected. Hardwood cuttings are made from the previous year’s mature wood and are usually collected during the dry season, when the leaves of deciduous species have dropped. Deciduous hardwood stem cuttings are the easiest, least expensive type of cuttings because they are easy to prepare, can be stored in coolers or shipped if necessary, and require little or no special equipment during rooting. Hardwood cuttings typically vary in length from 10 cm to 2 m, although shorter cuttings may

1098

R.K. Dumroese et al.

Fig. 12 Left to right: straight, heel, and mallet cuttings. Straight cuttings are used on easy-to-root species, while mallet and heel cuttings are used on species that are more difficult to root (Photo by Tara Luna)

work too (e.g., Dumroese et al. 2003). Shorter cuttings are used in nursery production. Longer, thick cuttings can be struck directly on the outplanting site for reforestation or plantation establishment or used, for example, to establish living fences that can help initiate natural forest restoration in abandoned fields (Zahawi 2005). When struck directly on the outplanting site, hardwood cuttings can be live stakes (30–40-cm long), poles (3–5-m long), or branched cuttings (0.5–2-m long). These cuttings are generally outplanted when the soil at the outplanting site is conducive for survival. Cuttings need to be planted deep enough to reach moisture in the soil profile. In nurseries, finished cuttings should have at least two nodes, with the basal cut just below a node and the top cut just above a node. These cuttings can be straight, heel, or mallet (Fig. 12). Straight cuttings, made from straight stems, are the most common type for easy-to-root species. Heel cuttings are made from 2-year-old side shoots. To make a heel cutting, pull the side shoot away from the tip so that a section of older wood remains at the base of the cutting. Mallet cuttings include a cross section of older stem at the base of the side shoot. All hardwood stem cuttings have an inherent polarity and will not root if struck upside down. Because the absence of leaves can make it difficult to discern the top from the

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1099

Fig. 13 Rooted softwood stem cutting of Hibiscus spp. Softwood stem-cutting material has some degree of flexibility but is mature enough to break when bent sharply (Photo by Tara Luna)

bottom, it is useful to cut the bottoms diagonally and the tops straight across. In addition to assisting with proper striking, the diagonal cut maximizes the water uptake area and the straight cut minimizes water loss. Semihardwood cuttings are made from newer shoots of leafy broad-leaved evergreen plants and leafy deciduous species. Cuttings are taken just before the onset of the dry season, toward the end of the active growth season when stem tissues have hardened, or just after a flush of growth when the wood is partially matured. Often, a terminal bud has formed. Cuttings are best rooted in special rooting environments after being wounded or treated with rooting hormone (described below). Cuttings are usually 10–20-cm long, leaves are removed from the lower half, green tips and side shoots are removed, and the large leaves of broad-leaved evergreen plants are usually cut in half to reduce water loss during rooting. Softwood cuttings, collected when stems and leaves are actively growing, generally root easier than other types of cuttings but require a special rooting environment and more attention to prevent desiccation. The best cutting material has some degree of flexibility but is mature enough to break when bent sharply (Fig. 13). Softwood stem cuttings are usually straight, 7–15-cm long with two or more nodes, and root best when collected from terminal shoots. Extremely fastgrowing, tender shoots are not desirable.

Collecting, Transporting, and Storing Cuttings Some basic equipment and supplies are necessary to efficiently collect cuttings and ensure their health until they are struck. The following items are recommended:

1100

R.K. Dumroese et al.

• High-quality, sharp pruning shears and pruning poles for collecting trees • Spray bottles filled with disinfectant (1 part bleach [5.25 % sodium hypochlorite] in 10 parts water) to disinfect pruning shears • Permanent labels and marking pens for noting the origin of collection • Large, white plastic bags with ties for bulk collections • Spray bottles filled with water to keep cuttings moist in the plastic bags after collection • Portable, insulated coolers for transport back to the nursery • Newspaper, moss, coir, or other materials to moisten and wrap around cuttings When collecting and handling cuttings, it is important to: • • • • •

Collect only from healthy donor plants. Keep cuttings cool to avoid wilting and desiccation. Handle cuttings carefully so that tissues are not bruised. Make sure that some buds or leaves are present on stem cuttings. Collect from nonflowering shoots because, in general, cuttings root better before or after flowering. • Place cuttings in the same direction when bundling to avoid confusion with polarity. Cuttings should be carefully collected on cloudy, cool days and/or during the early morning and kept cool, moist, and shaded during collection and transport back to the nursery to avoid water loss and physical damage. Cuttings should be clearly labeled with origin information and the date. Cuttings should be properly made to facilitate healing of the mother plant; take the cutting just above a node to minimize wood above where new growth will occur. Then, the base of the cutting can be trimmed to just below the node where rooting is more likely to occur. Between collection sites, use the bleach solution to disinfect pruning shears to avoid spreading disease. Deciduous hardwood cuttings can be stored in a shaded, dry environment for several days with periodic moisture to prevent desiccation. Inspect stored cuttings frequently to make certain that tissues are slightly moist and free from disease. Hardwood and softwood evergreen cuttings, deciduous softwood cuttings, and semihardwood cuttings should be struck in propagation beds the same day of collection.

Types of Rooting The development of new roots on a cutting is known as “adventitious root formation.” Many tropical species have buds that give rise to “preformed” or “latent” roots. These species root easily without a special rooting environment so their cuttings are usually struck directly into containers, making this the easiest and most economical vegetative propagation method because no transplanting is needed. Species without these buds produce “wound-induced” roots that form

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1101

Fig. 14 Semihardwood cuttings in a shadehouse at the University of the Virgin Islands Agroforestry facility are struck in trays of 100 % perlite and misted every hour to ensure that no standing water exists, but the cuttings are always moist. The shadecloth around the PVC frame provides a shady, low wind, high-moisture atmosphere for the cuttings while they produce roots. Rooted cuttings are later transferred to small pots in the shadehouse for 2–3 weeks and later transferred to full sun (Photo by Brian F. Daley)

only in response to the wound caused by preparing the cutting, and the ability to form roots can vary considerably among species. Callus tissue forms at the wound (usually the base of the cutting), primarily from the vascular tissue, and is the source of the adventitious roots. In general, all species with wound-induced roots must be rooted in a special propagation environment with tightly controlled air and medium temperatures, high relative humidity, reduced light levels, and a special rooting medium that maintains “moist but not wet” conditions (Fig. 14). After roots form, cuttings are transplanted into containers to continue their growth.

Cutting Preparation Cuttings are best prepared in clean work areas using sharp, regularly disinfest, wellmaintained shears and knives to make clean cuts and reduce potential for spreading disease. Cuttings that will require hormone treatment to encourage rooting should have one-third to one-half of the leaves and buds removed to reduce transpiration. Flower buds should be removed. Hand stripping small lower branches from cuttings creates wounded areas along the basal portion of the cutting; this can also be done by scraping the base of the stem with a small, sharp knife or potato peeler (Fig. 15) or slicing one or two long, shallow slivers (2–3-cm long) of tissue from the base of the stem, making sure to penetrate the cambium layer. Wounding increases rooting percentages and improves the quantity and quality of roots produced on difficult-to-

1102

R.K. Dumroese et al.

Fig. 15 Wounding the lower end of a stem cutting increases rooting with difficult-to-root species (Photo by Tara Luna)

root species because it exposes more cells to rooting hormone, encourages callus formation, and, in some cases, removes thick woody tissue that can be a barrier to root formation.

Rooting Hormones Auxins, natural plant hormones that encourage root formation, are available from natural and synthetic sources. Most cuttings are treated with synthetic hormones that are available in powder and liquid form, and some preparations may contain chemical fungicides. Synthetic hormones, purchased from suppliers, either come ready to use or can be mixed to specific concentrations. All rooting hormones have a limited shelf life of 18–24 months. Indole-3-butyric acid (IBA) and naphthaleneacetic acid (NAA) are the most widely used synthetic auxins and often are more effective when combined than when either is used alone. The effect of rooting hormones varies widely between species and, in some cases, among genotypes. In general, powder forms of synthetic hormones are preferred because varying strengths are available, they are easy to use, and large quantities of cuttings can be treated quickly. Powder must be applied uniformly to all cuttings. The following precautions and special techniques are necessary: • Wear gloves during application. • Transfer enough hormone to a smaller container from the main stock container for use; never transfer unused hormone back to the main stock container. • Ensure that the base of the cutting is moist so that the powder adheres; pressing cuttings lightly onto a moist sponge is a useful technique. • Apply the hormone uniformly to a depth of 5–10 mm and that cut surfaces and other wounds are also covered with rooting hormone. • Remove excess powder by lightly tapping the cuttings on the side of the dish.

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1103

Liquid products are formulated with alcohols and often must be diluted with great care to create the desired strength. The quick-dip method is the most common procedure for treating cuttings with liquid products – the base of the cutting is dipped into the solution for 3–10 s. Whole bundles of cuttings can be treated at once. An alternate method is to soak cuttings for a longer time in a more dilute hormone solution. When using liquid rooting hormones, it is important to: • • • •

Wear gloves during mixing, preparation, and application. Ensure the concentration is correctly and precisely made. Place the solution in a clean jar. Ensure that the treatment time is constant for a uniform application rate and to avoid damaging the plant tissue. • Dip the basal ends of cuttings to a uniform depth, especially if bundles of cuttings are treated. • Wait a minute or two to allow the alcohol to evaporate from the cutting before striking. • Properly discard any remaining solution. The optimum auxin rate for cuttings varies by plant species. A good starting rate is a 0.25–0.5 % (2,500–5,000 ppm) that can be adjusted through trial and error until optimum rooting occurs. Longman (1993) reports these IBA rates: 0.2 % for Triplochiton scleroxylon, Vochysia hondurensis, and several other tropical tree species; 0.4 % for Cordia alliodora; 1 % for Khaya ivorensis; and a range of 0.05–0.4 % for Albizia guachapele.

Striking, Monitoring, and Transplanting For easy-to-root species, cuttings can be struck directly into the same containers that will be used to grow the finished plants. Difficult-to-root species typically need a special rooting medium and propagation environment. Regardless of method, some methods are common to both and we encourage workers to (1) wear gloves if cuttings were treated with hormones, (2) strike the cutting with the correct end up (a dibble with the same diameter as the cutting is a useful tool for making openings in the medium), (3) ensure that at least two nodes are below the surface of the substrate and the cutting is firmly inserted, (4) strike all cuttings within a day or two, and (5) label cuttings. Once struck, maintain a clean rooting environment and routinely inspect cuttings for proper temperature, humidity, and moisture. Regularly inspect for, and remove, dead leaves or cuttings that could be a source of disease. Direct striking into containers is more efficient and therefore more economical than striking into a special rooting environment because the cuttings are handled only once and expensive transplanting is avoided. Easy-to-root hardwood cuttings, such as many Erythrina species, Gliricidia sepium (known locally as “quick stick” in much of the world), and mangrove (Avicennia and Rhizophora spp.) propagules,

1104

R.K. Dumroese et al.

Fig. 16 Easy-to-root hardwood cuttings (a) or mangrove propagules (b) can be directly stuck into containers for rooting and are the most economical way of producing cuttings (Photo (a) by Tara Luna and photo (b) by Thomas D. Landis)

should always be directly struck (Fig. 16). Containers are filled with regular growing medium (see chapter “▶ Tropical Nursery Concepts and Practices”), and once shoots appear, the cuttings can be grown with the same methods used for seedpropagated plants (see Wilkinson et al. 2014; Landis et al. 1998; Dumroese et al. 2008, 2012). Even directly struck cuttings may, however, benefit from modified light and humidity levels discussed below. For difficult-to-root species, start by striking cuttings into a sterilized rooting medium. A good rooting medium provides aeration and moisture, physically supports the cuttings, and promotes the development of fibrous root systems that retain rooting medium during transplanting. A pH of 5.5–6.5 is optimum for most plants. Rooting media generally include two or more of the following: large-grade perlite, pumice, Sphagnum peat moss, sawdust, sand, coir, grit, and fine bark chips. The ideal rooting medium drains freely and does not become waterlogged from misting. Until new roots form, cuttings require an environment with high relative humidity to slow the rate of water loss from the cutting (Fig. 14). Automatic misting or fogging systems can be used, but a simple frame covered with clear or white polyethylene sheeting with a reserve of water below a moist rooting medium works well too (Longman 1993). Cuttings in this frame should be misted with water in late afternoon and early morning, especially when the weather is hot and dry. Achieving optimum humidity and medium moisture can be one of the most challenging aspects of successful propagation with cuttings. Daily monitoring is important. Because too much sunlight can elevate temperatures, shadecloths providing 30–50 % can effectively reduce air temperature while still providing sufficient light to the plants to allow photosynthesis to occur. The optimum air temperature for rooting cuttings is 20–28  C, and optimum temperature of the rooting medium is about 3  C cooler than air temperatures.

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1105

Some difficult-to-root cuttings may remain in a special rooting environment for a long time, during which nutrients can leach from the leaves because of exposure to overhead misting, resulting in yellowing leaves or leaf and needle drop. Depending on the species, the application of a dilute, complete foliar fertilizer can be applied to improve cutting vigor and aid in rooting. Conduct preliminary trials before treating all the cuttings. The addition of nutrients can encourage unwanted growth of mosses and algae on the medium surface. Regular inspections will reveal when cuttings have developed adequate root systems and are ready to be hardened for transplanting. As discussed in chapter “▶ Tropical Nursery Concepts and Practices,” hardening conditions plants for the rigors of a harsher environment. Hardening may include gradually reducing the misting frequency during a 3–4-week period, increasing the frequency and duration of ventilation, and increasing the amount of sunlight. The medium, however, cannot dry out completely. Once hardened, cuttings can be transplanted into containers and grown on with the same methods used for seed-propagated plants (see Wilkinson et al. 2014; Landis et al. 1998; Dumroese et al. 2008, 2012). During transplanting, cuttings should be handled carefully to avoid damaging the roots by following these methods: • Transplant only on cool, overcast days or during early morning hours to avoid transplant shock. • Transplant cuttings in an area protected from wind and sunlight. • Prepare containers, medium, labels, and transplanting tools before removing cuttings from the rooting medium. • Moisten the growing medium before transplanting to prevent tender roots from drying out. • Remove cuttings from the rooting medium carefully and remove only a few at a time so roots will not dry out. Examine each cutting to ensure it has a sufficient root system. Cuttings with sufficient roots can be wrapped loosely with moist paper towels until they are transplanted. Cuttings with only a few slender roots or very short roots should remain in the propagation bed for further root development. • Handle cuttings carefully by gently holding the stem; allow any medium attached to the root mass to remain. • Partially fill the container with moistened medium before inserting the cutting. Then add additional moistened medium and gently firm the medium with the fingers without breaking the roots. • Do not transplant the cuttings too deep or too shallow. Once transplanted, cuttings should be placed in a shadehouse or protected from full sun and wind for at least 2 weeks. When the cuttings appear to be well established, gradually increase the level of sunlight by moving them to a different area of the nursery or by exchanging the shadecloth for one with a more open weave. After a couple of weeks, move sun-requiring species into full sun.

1106

R.K. Dumroese et al.

Other Vegetative Propagation Methods Root Cuttings Root cuttings can be made by dividing roots into individual segments containing dormant shoot buds capable of developing into new plants (Fig. 17). Root cuttings are typically used on species that fail to root well from stem cutting, such as breadfruit (Artocarpus spp.) and noni (Morinda citrifolia). Collected at any time of the year, most root cuttings are then struck horizontally in fields or containers with the dormant leaf buds on the upper side. Some root cuttings are struck vertically so it is important to maintain the correct polarity, accomplished by cutting the upper end horizontally and the basal end diagonally. Root cuttings generally do not require a special rooting environment.

Layering Air layering is useful for producing a few plants of relatively large size. An advantage of air layering is that the rooted layer will be physiologically similar in age to the parent plant and will therefore flower and fruit sooner than a seedling or cutting, which could be useful for breeding programs. Air layering is used mostly on fruit trees and to propagate rare and endangered species. For optimum rooting, air layers are made on shoots produced during the previous season or during the mid- to late-active growing season on shoots from the current season’s

Fig. 17 Root cuttings, such as those sown here from breadfruit (Artocarpus spp.), can be used when stem cuttings do not root well (Photo by Thomas D. Landis)

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1107

Fig. 18 After wounding the stem, an air layer is created by wrapping the area with peat moss or coir (a) and then enclosing this stem area in plastic wrap (b) and sealing the ends (c) After the layer has rooted (d) it can be severed from the stem and potted (Photos by Thomas D. Landis)

growth (Fig. 18). Stems of pencil-size diameter or larger are best. An area directly below a node is chosen, normally about 30 cm from the tip. Leaves and twigs on the stem are removed 7–10 cm above and below this point. Air layering techniques differ slightly depending on whether the species is a monocot or a dicot. The following steps describe air layering of monocots: • Make an upward 1.0–1.5-in (2–4-cm) cut about one-third through the stem. • Hold the cut open with a toothpick or wooden matchstick. • Surround the wound with moist, unmilled Sphagnum moss (about a handful) or coir that has been soaked in water and squeezed to remove excess moisture. • Wrap the moss or coir with plastic and hold in place with twist ties or electrician’s tape. No moss should extend beyond the ends of the plastic. • Fasten each end of the plastic securely, to retain moisture and to prevent water from entering. If it will be exposed to the sun, the plastic needs to be shaded with a cover. • After the rooting medium is filled with roots, sever the stem and pot the layer. Keep it shaded with adequate moisture until the root system becomes more developed.

1108

R.K. Dumroese et al.

The following steps describe air layering of dicots: • With a sharp knife, make two parallel cuts about 1 in apart around the stem and through the bark and cambium layer. • Connect the two parallel cuts with one long cut. • Remove the ring of the bark, leaving the inner woody tissue exposed. • Scrape the newly bared ring to remove the cambial tissue to prevent a bridge of callus tissue from forming. • Application of a root-promoting substance to the exposed wound is sometimes beneficial. • Wrap, cover, and harvest using the same procedure as that described for monocots. Other types of layering, e.g., simple layering, French layering, and mound layering, can also be used to grow adventitious roots on a stem while it is still attached to the plant. These are mostly used on species that fail to root from stem or root cuttings and can be especially important techniques for producing non-tree species desired for forest restoration projects. These techniques are fully described in Luna and Haase (2014).

Offsets Plants, such as plantain and banana (Musa spp.), reproduce by forming new shoots, called offsets, at the base of the main stem or in the leaf axils (Fig. 19). Offsets can be cut close to the main stem of the plant with a sharp knife. If well rooted, an offset can be potted individually. If not well rooted, offsets may be removed, placed in a rooting medium, and grown like other cuttings.

Grafting and Budding Grafting is the art of connecting two pieces of living plant tissues, the scion and the rootstock, together so that they will unite and grow as one plant. Grafting is used primarily in the tropics for mango, citrus, other tropical fruits, and forest tree seed orchards. In Hawaii, grafting has been used to propagate the highly endangered species, koki‘o (Kokia cookei) (Fig. 20). It is also used to repair existing trees, to topwork existing trees to change varieties, and to produce new plants. The scion, a short piece of shoot including several buds, is united with the rootstock to form the upper portion of the plant. The rootstock develops into the root system. The rootstock may be a seedling, a rooted cutting, or an older tree. When the scion is only a single bud attached to a piece of bark that may or may not include a thin sliver of wood, the grafting is termed “budding.” Budding is the most commonly used technique for propagating new plants and is used to topwork

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1109

Fig. 19 Plantain and banana are often propagated with offsets (Photo by Ronald Overton)

existing fruit trees to form new varieties. Grafted surfaces must be held tightly in place using a budding rubber or grafting tape that must either breakdown by weathering (as budding rubbers do) or be removed 2–3 weeks after the union has healed; otherwise it will girdle the rootstock. Then, the portion of the rootstock above the graft must be removed to force the scion bud to grow, and any unwanted sprouts must be removed as soon as they appear. The rootstock and the scion must be compatible for successful grafting. Compatibility is never a problem when grafting within a clone. Grafting between clones within a species is usually successful. Grafting between species in a genus is sometimes successful and is most often seen in the genus Citrus. Grafting between genera within a plant family is rarely done and the chances of success are slim. Grafting between plant families is impossible for woody plants.

Other Methods Plants can also be regenerated by using divisions of tuberous roots, tubers, and rhizomes, rooting stolons and runners, and using micropropation techniques. Some

1110

R.K. Dumroese et al.

Fig. 20 Grafting has been used to propagate highly endangered species in Hawaii. Shown: Kokia cookei scion grafted onto the rootstock of Kokia drynarioides (Photo by Tara Luna)

of these techniques apply to forest trees, while some of them have greater utility for growing species desired for forest restoration activities. Luna and Haase (2014) provide a full description of these methods. Acknowledgments This chapter draws heavily on Wilkinson et al. (2014), and we thank Tara Luna and Brian F. Daley for their contributions.

References Basiron Y (2007) Palm oil production through sustainable plantations. Eur J Lipid Sci Technol 109:289–295 Baskin CC, Baskin JM (1998) Seeds: ecology, biogeography and evolution in dormancy and germination. Academic, San Diego, CA, 666 p Baskin CC, Baskin JM (2004) Determining dormancy-breaking and germination requirements from the fewest numbers of seeds. In: Guerrant EO Jr, Havens K, Maunder M (eds) Ex situ plant conservation: supporting species survival in the wild. Island Press, Washington, DC, pp 162–179 Bass LN (1975) Seed storage of Carica papaya L. HortScience 10:232 Davis AS, Jacobs DF, Dumroese RK (2012) Chapter 15: Challenging a paradigm: toward integrating indigenous species into tropical plantation forestry. In: Stanturf J, Lamb D, Madsen

Collecting, Processing, and Treating Propagules for Seed and Vegetative. . .

1111

P (eds) Forest landscape restoration: integrating natural and social sciences. Springer, Dordrecht, pp 293–308 Dawson I, Were J (1997) Collecting germplasm from trees—some guidelines. Agroforest Today 9(2):6–9 Drake DR (1993) Germination requirements of Metrosideros polymorpha, the dominant tree of Hawaiian lava flow and rainforests. Biotropica 25:461–467 Dumroese RK, Landis TD, Wenny DL (1998) Raising forest tree seedlings at home: simple methods for growing conifers of the Pacific Northwest from seeds. Contribution 860, Idaho Forest, Wildlife and Range Experiment Station. University of Idaho, Moscow, 56 p Dumroese RK, Wenny DL, Morrison SJ (2003) Propagation protocol for container willows and poplars using mini-cuttings. Native Plants Journal 4:137–139 Dumroese RK, Luna T, Landis TD (eds) (2008) Nursery manual for native plants: a guide for tribal nurseries. Volume 1: nursery management. US Department of Agriculture Forest Service, Agriculture Handbook 730, Washington, DC, 302 p Dumroese RK, Landis TD, Luna T (2012) Growing native plants in nurseries: basic concepts. RMRS-GTR-274, US Department of Agriculture Forest Service, Rocky Mountain Research Station, General Technical Report, Fort Collins, CO, 84 p Flores EM (2002) Chapter 1: Seed biology. In: Vozzo JA (ed) The tropical tree seed manual. US Department of Agriculture, Forest Service, Agriculture Handbook 721, Washington, DC, pp 13–118 Gao YP, Zheng GH, Gusta LV (1998) Potassium hydroxide improves seed germination and emergence in five native plant species. HortScience 33:274–276 Hong TD, Ellis H (2002) Chapter 3: Storage. In: Vozzo JA (ed) The tropical tree seed manual. US Department of Agriculture, Forest Service, Agriculture Handbook 721, Washington, DC, pp 125–136 James RL, Genz D (1981) Ponderosa pine seed treatments: effects on seed germination and disease incidence. US Department of Agriculture, Forest Service, Northern Region, Forest Pest Management Report 81–16, Missoula, MT, 13 p Johnson R, Stritch L, Olwell P, Lambert S, Horning ME, Cronn R (2010) What are the best sources for ecosystem restoration on BLM and USFS lands? Native Plants Journal 11:117–131 Keeley JE, Fotheringham CJ (1998) Smoke induced seed germination in California chaparral. Ecology 79:2320–2336 Kettle CJ, Burslem DFRP, Ghazoul J (2011) An unorthodox approach to forest restoration. Science 333:35 Landis TD (2008) Chapter 2: The target plant concept. In: Dumroese RK, Luna T, Landis TD (eds) Nursery manual for native plants: a guide for tribal nurseries. Volume 1: nursery management. US Department of Agriculture, Forest Service, Agriculture Handbook 730, Washington, DC, pp 14–31 Landis TD, Dreesen DR, Dumroese RK (2003) Sex and the single Salix: considerations for riparian restoration. Native Plants Journal 4:111–117 Landis TD, Tinus RW, McDonald SE, Barnett JP (1998) The container tree nursery manual. Volume 6: seedling propagation. US Department of Agriculture, Forest Service, Agriculture Handbook 730, Washington, DC, 167 p Longman KA (1993) Rooting cuttings of tropical trees: propagation and planting manuals, vol 1. Commonwealth Science Council, London, p 137 Luna T (2003) Native plant restoration on Hawai‘i. Native Plants Journal 4:22–29, 32–36 Luna T, Haase DL (2014) Vegetative propagation. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 185–205 Luna T, Wilkinson KM (2014) Collecting, processing, and storing seeds. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 141–161

1112

R.K. Dumroese et al.

Luna T, Wilkinson KM, Dumroese RK (2008) Seed germination and sowing options. In: Dumroese RK, Luna T, Landis TD (eds) Nursery manual for native plants: a guide for tribal nurseries. Volume 1: nursery management. US Department of Agriculture, Forest Service, Agriculture Handbook 730, Washington, DC, pp 132–151 Luna T, Wilkinson KM, Dumroese RK (2014) Seed germination and sowing options. In: Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, pp 163–183 Mawdsley JR, O’Malley RO, Ojima DS (2009) A review of climate-change adaptation strategies for wildlife management and biodiversity conservation. Conserv Biol 23:1080–1089 Mousa H, Margolis HA, Dubay PA, Odongo J (1998) Factors affecting germination of doum palm (Hyphaene thebaica Mart.) from the semi-arid zone of Niger, West Africa. For Ecol Manage 104:27–41 Narimanov AA (2000) Presowing treatment of seeds with hydrogen peroxide promotes germination and development in plants. Biologia 55:425–428 Putz FE, Zuidema PA, Synnott T, Pen˜a-Claros M, Pinard MA, Sheil D, Vanclay JK, Sist P, Gourlet-Fleury S, Griscom B, Palmer J, Zagt R (2012) Sustaining conservation values in selectively logged tropical forests: the attained and the attainable. Conserv Lett 5:296–303 Schmidt LH (2007) Tropical forest seed. Springer, Berlin/Heidelberg/New York, 409 p Stanturf JA, Palik BJ, Williams MI, Dumroese RK, Madsen P (2014) Forest restoration paradigms. J Sustain Forest 33:S161–S194 [USDI BLM SOS] US Department of the Interior, Bureau of Land Management, Seeds of Success (2011) Technical protocol for the collection, study, and conservation of seeds from native plant species for seeds of success. Native Plant Materials Development Program, pp 8–11. http:// www.nps.gov/plants/sos/training/index.htm. Accessed Oct 2011 Vozzo JA (ed) (2002) The tropical tree seed manual. US Department of Agriculture, Forest Service, Agriculture Handbook 721, Washington, DC, 899 p Weber M, G€unter S, Aguirre N, Stimm B, Mosandl R (2008) Reforestation of abandoned pastures: silvicultural means to accelerate forest recovery and biodiversity. In: Beck E, Bendix J, Kottke I, Makeschin F (eds) Gradients in a tropical mountain ecosystem of Ecuador, vol 198, Ecol Stud, pp 431–441 Wilkinson KM, Elevitch CR (2003) Growing koa: a Hawaiian legacy tree. Permanent Agriculture Resources, Holualoa, HI, 104 p Wilkinson KM, Landis TD, Haase DL, Daley BF, Dumroese RK (eds) (2014) Tropical nursery manual: a guide to starting and operating a nursery for native and traditional plants. US Department of Agriculture, Forest Service, Agriculture Handbook 732, Washington, DC, 376 p Zahawi RA (2005) Establishment and growth of living fence species: an overlooked tool for the restoration of degraded areas in the tropics. Restor Ecol 13:92–102

Plant Nutrition in Tropical Forestry Alfredo Alvarado

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nutrient Storage and Mineral Cycling in Tropical Natural Forest and Forest Plantations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plant Adaptation to Nutritional Constraints and Ecosystem Modification by Plants . . . . The Nutrient Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Silvicultural Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Visual Diagnoses of Mineral Disturbances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Causes and Occurrence of Mineral Deficiency and Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Symptoms of Mineral Deficiency and Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Deficiency Symptom Identification Guides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analytical Diagnoses by Soil and Foliar Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Possibilities and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plot Selection, Size, and Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Site and Soil Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Soil Sampling and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Number of Trees to Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Foliar Sampling and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpretation of Soil and Foliar Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpretation Without Reference Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evaluation of Soil and Foliar Data by Correlation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The “Critical Level” Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The DRIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interpretation of Soil Analysis to Choose Lands for Planting Forest Species . . . . . . . . . . . .

1114 1116 1116 1117 1133 1135 1136 1137 1140 1145 1145 1147 1149 1150 1152 1152 1156 1156 1157 1158 1160 1162

A. Alvarado (*) Centro Investigaciones Agrono´micas, Universidad de Costa Rica, San Pedro Montes de Oca, Costa Rica e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_105

1113

1114

A. Alvarado

Amelioration of Site and Tree Nutrient Status Without Mineral Fertilizer . . . . . . . . . . . . . . . . . . Fertilization of Forest Plantations and Natural Tropical Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experimental Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Time of Fertilizer Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Method of Fertilizer Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rates of Nutrient Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Type of Fertilizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Magnitude of Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fertilizer Recommendations for Plantations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fertilizer Recommendations for Natural Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Economics of Fertilizer Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1166 1168 1169 1169 1170 1171 1172 1173 1176 1179 1181 1182

Introduction The chapter summarizes available information on nutrition and fertilization of natural forest and tree plantations in tropical regions with emphasis on about 30 species of economic relevance in lowland wet tropical America (mainly Eucalyptus ssp., Tectona grandis, Acacia mangium, and Pinus spp.). Information on dry or seasonally dry tropical forests is scarce. Recent documents on tropical forest and plantation nutrition considered in this review include Fisher and Binkley (2000), Reddy (2002), Hartemink (2003), Rodríguez and Alvárez (2010), Paudyal (2012), and Alvarado and Raigosa (2012). Forest fertilization in tropical countries was initiated in the middle of the twentieth century (Mustanoja and Leaf 1965) with the expansion of forest plantations around 1950 (Evans 1992). However, the gain in knowledge in forest nutrition in tropical areas is very slow (Figueroa 1986) as reflected by the fact that studies on the subject were initiated in 1970 in Cuba and in 1980 in Brazil (Herrero 2001) and in Central America (Alvarado et al. 1997). For many years forest nutrition was based on the concept that trees should grow in soils other than those devoted to planting crops. Forest operations should be carried out in poor soils leaving good fertile irrigated soils to crops, a concept accepted by foresters that never considered the nutritional or other needs of the trees (Nwoboshi 1975). This hypothesis of trees growing without inputs lasted for years and is actually accepted in many parts of the tropical world. Today we know that trees can grow in that kind of soils but also that in order to maximize productivity of plantations, adapted species to restringing conditions should be employed and soil amendments should be implemented to overcome growth impediments (i.e., drainage, subsoiling, additions of fertilizers and lime, etc.). We also know that after solving the nutritional problems of a site, other factors like water stress and weed control should be taken into account to avoid nutrient and water competence and possibly allelopathic effects from the weeds (Ladrach 1992).

Plant Nutrition in Tropical Forestry

1115

During the past decades demands for forest products have drastically increased and natural forest resources mainly in the tropics have been steadily depleted. In consequence there have been increases in afforestation and replanting programs, which in the tropics have generally been on relatively infertile soils where nutritional problems have soon become apparent. Parallel to these developments, the need for increased productivity has resulted in the use of more intensive production techniques including the use of fertilizers, with increasing interest from both silviculturists and physiologists in the nutritional problems encountered in tropical conditions (Brunck 1987; Reis and Barros 1990). The chemical analysis of plants usually reveals the following elements of vital necessity: carbon (C), hydrogen (H), and oxygen (O), which constitute the organic matter and represent about 90–95 % of the dry weight. C and O, which are supplied for photosynthesis by the air, can hardly become limiting factors, while the supply of H depends on the availability of water. The other elements constitute on the average 5–10 % of the dried plant material and are the soil’s contribution to the growth of the plants: nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), and sulfur (S), which are considered as macronutrients, and manganese (Mn), iron (Fe), zinc (Zn), copper (Cu), boron (B), chlorine (Cl), and molybdenum (Mo), which are called micronutrients due to their lower quantities in the tissue in comparison with the macronutrients. The reader is referred to specific papers while looking for information on individual elements like P (Ballard 1980), S (Johnson 1984), Zn (Boardman and McGuire 1990), or B (Stone 1990; Lehto et al. 2010). Today, several other elements are known which will be useful for some physiological processes or could replace other nutrients in some functions [cobalt (Co), silicon (Si), sodium (Na), etc.]. These elements are, however, of minor interest in tropical forest nutrition management. The fundamental physiological relationships between mineral nutrition and growth are the same in trees as in other plants. The basic knowledge about the nutrients needed for plant growth, the nutrient uptake mechanisms, and the function of nutrients in plant physiology has been extensively examined and summarized (e.g., Bowen and Nambiar 1984; Marschner 1986; Mengel and Kirkby 1987; Wild 1988). On the other hand, there is little known, especially in the tropics, about the complex long-term relationship between soil nutrient status and nutrient availability on the one hand and tree nutrition as well as tree growth on the other. In this chapter, information about tropical tree nutrition and nutrient cycling is summarized as it is valuable to know for nutrient management in silviculture. In the following sections recommendations for diagnosis of mineral disturbances, using visual and analytical methods, discuss the problems associated with these approaches. The last paragraphs concern practical conclusions for nutrient management and summarize experiences on fertilization.

1116

A. Alvarado

Nutrient Storage and Mineral Cycling in Tropical Natural Forest and Forest Plantations Plant Adaptation to Nutritional Constraints and Ecosystem Modification by Plants An understanding of nutrient cycling is necessary for improving site fertility by nutrient management with or without use of fertilizers. There are several reviews and case studies on this subject (e.g., Switzer and Nelson 1972; Singh 1982; Miller 1981, 1984; Turner and Lambert 1983; Jordan 1985; Heal et al. 1997; Reddy 2002; Alvarado and Raigosa 2012; Chapin III et al. 2012). To improve upon sustainable management of natural forest and short-rotation plantations, especially when nutrient deficiencies are common, fertilization plans based on better knowledge on tree physiology, forest nutrition, and morphological properties of trees related to their reproduction, survival, and growth characteristics are required (Go´mez and Burley 1991; Grubb 1995; Medina 1995). Under natural forest conditions the majority of the nutritional problems are related to nutrient cycling, root growth habit of the species, its rotation length, and the large amount of biomass of the trees (Ballard 1980). To overcome nutritional problems plants might have to adapt to environmental conditions or the sites should be amended to correct for such problems. Various adaptation mechanisms of trees to nutritional stress conditions had been described, but no one species manifests all of them (polygenetic inheritance), and cases where two or more mechanisms act together are considered rare. Many articles are available to explain mechanisms of adaptation to soil physical constraints (Vitousek 1984; Reis and Barros 1990), low or high soil fertility (Vitousek and Standfor 1984; Bond 2010; van Breugel et al. 2011), nutrient constraints such as soil acidity (Wright 1976; Bouldin 1979; Robson 1989; Baligar and Duncan 1990; do Vale et al. 1996; Vitorello et al. 2005; Poschenrieder et al. 2008), phosphorus deficiencies (Rao et al. 1999), salinity and waterlogging conditions (Holdridge et al. 1971; Wright 1976), and light interception (Ortín 1997; Herbert and Fownes 1999; Soethe et al. 2008). How plants modify the ecosystems is also being studied by various authors (Duchaufour 1977; Nath et al. 1988; Blazer and Camacho 1991; Márquez et al. 1993; Binckley and Giardina 1998; Widmer 1999a, b; Tobo´n et al. 2010; Hafich et al. 2012; Samndi and Jibrin 2012); Eviner and Chapin III (2003) recommend the use of the functional matrix concept, which builds upon the functional group and single trait approaches to account for the ecosystem effects of multiple traits that vary independently among species. Chapin (1980) and Chapin et al. (1986) mention that the majority of species growing in infertile soils are tolerant to natural stress conditions through mechanisms like (i) synchronized low growth rates with a high photosynthetic capacity and low nutrient absorption, (ii) high biomass production of long-lasting roots commonly associated with mycorrhizae, (iii) low response to fluxes or additions of nutriments, (iv) low nutrient demands along the year to produce new tissue, and

Plant Nutrition in Tropical Forestry

1117

(v) low throughfall nutrient loss from old leaves. Poschenrieder et al. (2008) utilize soil and plant variables to estimate the risk of Al toxicity on trees; the author measured in the soil (i) the Al in soil solution, (ii) the percentage of base saturation, and (iii) the relationship exchangeable bases/exchangeable Al (values N > Ca > K > Mg > S > P), P. patula (C > Ca > N > K > S > Mg > P), and C. lusitanica 42-yearold plantations, Ramírez et al. (2007) and Leo´n et al. (2011) mention additions in the order of 7.5–7.9, 7.8–8.4, and 3.5–3.7 Mg ha1, respectively, with slow return of residues and nutrients of C. lusitanica due to the lower C/N ratio of its litter. Leo´n et al. (2011) measured residence time of the residues finding values of 3.3, 2.1, and 1.8 years for cypress, pines, and oaks, the last genera contributing more nutrients with time to the ecosystem biochemical cycle. Tree species from the tropical lowland wet forest have different nutrient concentrations (Ortín 1997); some absorb large quantities such as Croton smithianus, while others absorb lesser amounts (Calophyllum brasiliense) but all of them in the order N  K > Ca > Mg > P (Table 1). Normally, nutrient concentration on residues is (i) higher in Ca and N than in K, Mg, and P, (ii) higher on leaf rather than branch residues, and (iii) higher on residues from deciduous species (i.e., teak) than evergreen species (i.e., pines or eucalyptus). Sharma and Pande (1989) report that leaf residue nutrient concentration decreases as the amount of falling residues and the tree biomass increases. Under different environmental conditions and vegetation cover, Geigel (1977) found in Cuba that the concentration of nutrients under T. grandis, Hibiscus sp., S. macrophylla, and P. caribaea contained more Ca than N and Mg and much lower quantities of K, P, and Na. Ogbonna and Nzegbule (2009) also reported Ca, K, Na, and P contents to be higher in residues under G. arborea plantations than under P. caribaea plantations. While testing the nutritional value of primary tropical wet forest residues on the growth of secondary forests of the same area, Wood

Species Croton smithianus Apeiba membranacea Rollinia microsepala Simarouba amara Virola sebifera Laetia procera Pentaclethra macroloba Tapirira guianensis Qualea paraensis Vochysia ferruginea Calophyllum brasiliense Range

Samples No 5 10 7 18 20 10 20 19 15 20 15

Median (%) N 4.27 2.67 2.34 2.62 2.34 2.52 2.84 1.88 1.63 1.86 1.45 1.45–4.27 K 1.45 1.61 0.82 0.83 0.48 0.47 0.42 0.36 0.37 0.34 0.29 0.29–1.61

Ca 0.45 0.90 1.10 0.58 0.74 0.45 0.22 0.92 0.78 0.63 0.36 0.22–1.10

Mg 0.29 0.79 0.54 0.31 0.40 0.30 0.17 0.40 0.32 0.17 0.12 0.12–0.79

Table 1 Foliar nutrient content of 11 species of the lowland tropical wet forests of Costa Rica (Adapted from Ortín 1997) P 0.22 0.11 0.12 0.09 0.08 0.10 0.09 0.08 0.06 0.06 0.05 0.05–0.22

Total 6.68 6.08 4.92 4.43 4.04 3.84 3.74 3.64 3.16 3.06 2.27

Plant Nutrition in Tropical Forestry 1121

1122

A. Alvarado

et al. (2009) found that tree leaf foliage biomass increased but diameter did not nor soil fertility parameters measured. Effect of Ecological Succession Stages, Plantation Age, and Plantation Density: Bruijnzeel (1991) differentiates (i) primary and secondary forests in advanced stages of succession that mainly depend on nutrient cycling to sustain growth, N fixation, and other nutrient sources like dry and wet additions from (ii) forest plantations on early stages of succession with high photosynthesis rates and high demand of nutrients in short periods of time, at least during the first 5–10 years of growth, that dominate vast riverine areas where large amounts of residues are added to the system and less soluble forms of P and K are converted to more soluble forms (Salo et al. 1986; Szott et al. 1999). Some authors assume that nutrient requirements of trees diminish with plantation age since annual growth increments also decrease with plantation rotation age (Miller 1981, 1995). Davidson et al. (1998) add that trees in early succession stages of growth have a larger foliar and wood concentration of nutrients, a fact that correlates with high growth rates, a superficial root math, and high soil coverage that minimizes soil erosion and leaching. In terms of foliar nutrient concentration, Gonc¸alves et al. (2005) also mention that species of medium succession stages of development like Swietenia macrophylla are higher when shaded, while species of late succession stage of development like Dipteryx odorata are not. Other authors (i.e., Segura et al. 2006b) also report that growth pattern and nutrient content of aerial biomass components vary with plantation age since wood production increases with time; similar results were found by Jiménez and Arias (2004) while studying root biomass and nutrient content for various tropical forest species. Residues returned to the environment are generally plantation density dependent. Akinnifesi et al. (2002) found that residue turnover increased in plantation density of T. grandis that was not affected in plantations of Shorea robusta and decreased in plantation density of E. camaldulensis and Pinus roxburghii. Effect of Understory Vegetation: A less studied source of residue production is that related to understory vegetation in natural forest and plantations. These residues might be abundant or not depending on the main species of the ecosystem and plantation management. The nutrients associated with this biomass and residues are relevant not only for their amount and quality but also because the volume of soils they explore is different to that explored by the principal species recycling nutrients that otherwise could be leached away of the system. Frequently weeds are considered also a part of this type of vegetation even though they might play an important role in preserving nutrients in the ecosystem like absorbing N from the mineralization of residues left on the ground after harvest. As a way of weed control, some small wood producers raise cattle in the plantations “improving nutrient dynamics” but also negatively affecting the quality of the wood produced. Effect of Elevation and Associated Variables: Residue production in mountain tropical forests is lower than the one found in tropical lowland forests due to a lower rate of photosynthesis and the effect of strong winds that reduce tree growth. Heaney and Proctor (1989) measured residues added to Andisols at 100, 1,000, 2,000, and 2,600 masl in Costa Rica to be in the order of 9.0, 6.6, 5.8, and

Plant Nutrition in Tropical Forestry

1123

5.3 Mg ha1 year1. Soethe et al. (2008) did not find the uniform effect of altitude on similar soils of Ecuador at 1,900, 2,400, and 3,000 masl, but they found that total amounts of N, S, and Mg were higher over 3,000 masl and that contents of P, K, and Ca were not affected by altitude. Szott et al. (1999) mentions that biomass accumulation in wet tropical secondary forests is essentially linear during the first 10 years of growth and in the order of 4–15 t ha1 year1, while in dryer areas biomass accumulation accounts for 1–8 t ha1 year1 due to water availability limitations. Nambiar (1998) and Montagnini (2002) report annual increments of wood biomass production from broad-leaved plantation species in the order of 1–28 t ha1 year1, while short- and medium-term plantations might yield 1–2 to 25–30 m3 ha1 year1, respectively. Fast-growing species like Gmelina arborea and Eucalyptus saligna yield 10–20 and 8–28 t ha1 year1 and other species with lower growth rates like Swietenia spp. and Tectona grandis 1–4 and 3–12 t ha1 year1, respectively. Under African conditions Australian acacias produce larger amounts of residues than Terminalia, teak, and eucalyptus and even more than cypress and pines.

Soil Litter Deposition and Turnover Tropical forests and plantations benefit from nutrient release after mineralization of residues through nutrient cycling. Some elements, remarkably K, can be washed from foliage by rainwater and large amounts of N can be retranslocated to more woody tissues before senescence of the tissues contributing less to the nutrient cycle (Aerts 1996). Nutrient cycling relevance was summarized by Nye and Greenland (1960) and later reassumed by Sánchez (1985) and Jordan (1985) who distinguished the management of residues in the “slash-and-mulch” system that mimics the effects of slow fertilizer release from the “slash-and-burn” system which causes effects similar to the starting fertilizer (lime) effects of commercial fertilizer additions. Nutrient cycling is of particular relevance for secondary forests in dystrophic environments where low nutrient availability partially defines succession step length. Apart from the nutrient effects under the slash-and-mulch pathway of nutrient cycling, it is worth to mention that residues left on the ground are focuses for disease and plagues and also fuel sources for forest fires. Amounts of Residues and Nutrients Deposited: In a tropical forest plantation, the residues added to the system range from 5 to 20 Mg ha1 year1, and nutrients associated with the residues represent 50 % of total nutriments associated with aerial biomass contributing with 40–45 % of the N, 54 % of the K, 56 % of the Mg, and 28 % of the P (Montagnini 2002). Geigel (1977) and Akinnifesi et al. (2002) indicate that the amount of residues and associated nutrients increases linearly with plantation age mainly due to woody tissue production until they reach 8–10 years of age and attain its maximum growth. The effect of sites and species on residue addition is demonstrated by Akinnifesi et al. (2002) at Ibadan, Nigeria, in 10-yearold plantations of C. alliodora, G. arborea, Irvingia gabonensis, Leucaena leucocephala, and Pterocarpus soyauxii that deposited residues in the range of 4.6–9.7 Mg ha1 and those of 7–10-year-old plantations of P. caribaea that ranged 5.9 Mg ha1. Samra and Raizada (2002) found that the amounts of residues

1124

A. Alvarado

deposited by five eucalyptus plantations in Nilgiris, India, ranged between 2.1 and 6.8 Mg ha1 year1 with nutrient quantities of Ca > N > K > Mg > P, indifferently of the species compared (E. acmenoides, E. eugenioides, E. paniculata, E. pilularis, and E. propinqua). Deposition of residues on fast-growing plantations of E. terricormis, E. globulus, and P. patula is normally more than 2 Mg ha1 year1, mainly during the early stages of growth. Exceptionally high residue additions are mentioned for Casuarina equisetifolia (8.7–29.6 Mg ha1 year1) growing on good soils and more than 2,000 mm of rain per year in different regions of India. Amounts of Residues Deposited Along the Year: Residue deposition is not constant throughout the year. In tropical deciduous forests or plantations, the major addition of foliage and root residues occurs at the beginning of the dry season, like what happens after pruning or thinning of the plantations. Most of the litter deposited starts to mineralize at the beginning of the wet season creating a nutrient boom that enhances new root growth and the formation of new roots (fertilizer if needed should be added later considering this as natural addition of nutrients). The dilution effect of nutrients in the canopy created after regrowth can reduce growth rates of trees unless foliar deficiencies are corrected through fertilizer additions at the pick of the rainy season. Apart from residues added after selective extraction of wood from the forest or pruning, thinning, and harvest of the trees in plantations, other natural additions of residues are the result of falling leaves and twigs that depends mainly on drastic climatic changes and also on natural soil fertility, age of the trees, season of the year, species composition, and occurrence of catastrophic events (i.e., hurricanes, acid rains, earthquakes, etc.). While comparing tree phenology in the wet and the seasonally dry tropical forests of Costa Rica, Frankie et al. (1974) found that in the wetter environments species diversity is higher (185 species) than in the dryer environments (113 species) and that a larger amount of deposited leaves occurred at different times of the year (17 % of trees defoliate in the less humid months of the wet areas, while 75 % of the trees defoliated during the clear dry season in the seasonally dry forest).

Nutrients in the Biomass Nutrients in the biomass depend among others on age and floristic composition of the forest (Nye and Greenland 1960; Cuevas and Medina 1986), the bioclimatic characteristics or the sites (Bruijnzeel 1991; Soethe et al. 2008; Arias et al. 2011), and the natural fertility of the ecosystems (Schlesinger 1997; Szott et al. 1999; Hartemink 2003). Nutrient content of E. globules, G. arborea, and T. grandis is considered larger than nutrient concentration found in many other tree species (Totey 1992), particularly for N, Ca, and K concentrations that tend to accumulate in woody tissues (Grimm and Fassbender 1981a, b; Totey 1992; Stanley and Montagnini 1999). Nutrient Use Efficiency: To explain the relationship of nutriments and forest ecological processes, various indices had been developed, among them the nutrient use efficiency (NUE), the nutrient use efficiency of residues (NUER),and the plantation stability index (PSI). The ability of a species to produce biomass with

Plant Nutrition in Tropical Forestry

1125

small amounts of nutrients is a property used to screen genetic material for recuperation of degraded lands (Montagnini 2002). In mature forests NUE values are high maybe limiting net primary productivity, but in less efficient ecosystems the nutrient availability is adequate and NUE values are low (Vitousek 1982); however, Vitousek (1984) found that tropical forest ecosystems yield systematically more dry biomass residues per N unit than other ecosystems such as temperate, conifer, Mediterranean, and fertilized plantations with less N circulation in aerial biomass. To improve the NUE of a species, different mechanisms can be applied including (i) genetic improvement to obtain plants that are photosynthetically more active and (ii) through plantation arrangement that maximizes nutrient absorption and biomass production (Ewell and Hiremath 1998). While comparing species planted alone or combined, Hiremath et al. (2002) and Hiremath and Ewell (2001) found that NUE is not constant and relevant deviations from the mean depend on variable environmental conditions that determine among all nutrient availability. The NUE decreases in the following order Hieronyma > Cedrela > Cordia while planted as monocrops, but when helicoids and palms are included, NUE values follow the order Cedrela > Cordia > Hieronyma. In the majority of forestry ecosystems, NUE values for N, K, Ca, and Mg are alike. High values of NUE are common when tropical forests with naturally available P are low with species adapted to mycorrhizae associated with their root math. However, species like Bombax macrophyllum and Plathymenia foliolosa show extremely high NUE values for most elements and are considered as ideal for restoring degraded ecosystems (Montagnini and Jordan 2002). In teak plantations nutrients are bound to the biomass very efficiently returning little amounts to the ecosystem: in evergreen species such as pines and eucalypts, nutrients (particularly N and P) are recycled in large quantities. In the case of eucalyptus, the fast recycling of nutrients and the large amounts of nutrients exported in the wood harvest tend to mine soil fertility, particularly of sandy soils (Samra and Raizada 2002). Nutrient Translocation (Resorption): Some authors (Karmacharya and Singh 1992; Konsaeng et al. 2005; Tully et al. 2013) agree that nutrients move via xylem to growing tips where they move via phloem (translocate) to other tissues like roots, branches, and leaf cells; however, mobility of nutrients varies notably in the phloem. Nutrients kept in tissues and organs can later be translocated to new growing tips or reproductive organs. Nambiar (1984) mentions that this mechanism is highly developed in trees where they are active as long as they live; trees have large amounts of nutrients in the leaves and the stem, and they possess the ability to mobilize them to maintain their short- and long-term growth. Translocation of nutrients is a mechanism used by trees from temperate zones before senescence at autumn to translocate nutrients to wood tissues and similarly by senescent species in tropical regions responding to changes in photoperiodism, small changes in air temperature, or lag of humidity in regions with seasonal rainfall distribution (Reich and Borchert 1984; Borchert 1994; Borchert and Rivera 2001). Nutrients like N are translocated to woody tissues before senescing in deciduous species, but not in less efficient evergreen species; in deciduous species nutrients need to be translocated to the foliage to form new leaves every year and to

1126

A. Alvarado

replace nutrients removed by rain as leaching, avoiding the costly energy requirement of absorbing them from the ground (Samra and Raizada 2002). Monthly leaf variations of nutrients like N, P, and K are larger than those of Ca and Mg due to the rainwashing effect and the translocation of the first elements mentioned. The amount of nutrients translocated in wild species is considered to vary between 0 % and 83 % (Chapin 1980). Gordon and Jackson (2000) describe the translocation of nutrients in the root math of conifers and broad-leaved species as dependent on the C/N and C/N/P ratios, and Aerts (1996) found differences in resorption of N in deciduous species (54 %) against evergreen species (47 %); P resorption in the two groups of species did not vary significantly (50–51 %). Inagaki et al. (2011) found larger amounts of litter N in residues of Acacia mangium than in residues of Swietenia macrophylla and Araucaria cunninghamii, but values of P are 12–22 % lower than those found in the latter two species: these results indicate the lower translocation efficiency of P by A. mangium in contrast to its high N cycling mechanism under field conditions.

Nutrient Additions Through Rainfall and Tissue Leaching In spite of the difficulties encountered while measuring nutrient addition from rain (Bruijnzeel 1991), some values from various tropical forest environments are available (Vitousek and Standfor 1984; Fo¨lster and Khanna 1997; Hendry et al. 1984; Eklund et al. 1997; Montagnini and Jordan 2002). The following values represent ranges of most likely amounts added to the ecosystems in kg ha1 year1: 5.0–21.0 N, 0.2–1.1 P, 2.5–24.0 K, 1.4–34.0 Ca, and 1.1–26.0 Mg. The quantities might seem not too large, but through the 20–25 years of length to harvest, the nutrients added this way might have some impact on growth. Park and Cameron’s (2008) summary on water interception by trees concludes that major pathways for rainfall water to reach the ground includes throughfall (and leaf drip) and stemflow. Through the second pathway, nutrients can be washed away from the canopy by rain, particularly P, K, Ca, and Mg, whose foliar concentration diminishes with precipitation augments (Santiago et al. 2005) and is related to the form of the canopy of species (Hiremath et al. 1997). Quality of rainfall water is also affected by the properties of tissues of the trees of the wet tropical regions (McColl 1970; Grimm and Fassbender 1981a, b; Hiremath et al. 1997; Wilcke et al. 2001; Cavalier and Vargas 2002; Cleveland et al. 2004; Jiménez et al. 2006). In areas near the ocean, small amounts of nutrients like Na, B, Mg, Cl, and SO4 can be dissolved and suspended in rainfall water affecting vegetation growth in land (Forti et al. 2000). In areas around deserts rainfall water can be enriched with Ca and near industrial centers or active volcanoes with SO4 and NO3 (Hendry et al. 1984; Rodríguez et al. 1995; Eklund et al. 1997; Schlesinger 1997; Montagnini and Jordan 2002; Fabian et al. 2010). The effect of burning the Amazonian forest on the acidification of rainfall water that affects vegetation growth in the Ecuadorian Andes is being described by Boy et al. (2008).

Plant Nutrition in Tropical Forestry

1127

Effect of Microorganisms on the Nutrient Cycle of Tropical Forest and Forest Plantations Dommergues (1997) mentions that N fixation of species like Casuarina sp. and Alnus sp. is high, but real amounts of fixed N are low since the potential for fixing the element is limited by unfavorable environmental conditions, wrong silvicultural practices (which can be improved), and the low fixation capacity of actinomycetes (which can also be optimized). In greenhouse conditions the growth of Frankia on the root math of Alnus glutinosa is severely diminished at pH values lower than 5.5 (Griffiths and McCormick 1984) as also happens with the one on Alnus incana due to B deficiencies (Lehto et al. 2010). In a natural humid tropical forest, Moreira and Arnáez (2009) and Eaton et al. (2012) found large amounts of Frankia spores and a large amount of nodules attached to the fine roots of H. alchorneoides, a mechanism proposed as a strategy of trees to survive in soils of low fertility. The need to apply mycorrhizae in native species plantations is minimal since the amount of spores in the soil is large even after long periods of time (Rojas 1992; Fisher et al. 1994; Johnson and Wedin 1997). However, it is imperative in highly deteriorated soils (Alvarado et al. 2004) and when exotic species are introduced like what happens with the introduction of P. caribaea in Costa Rica (Vega 1964). The response to the application of mycorrhizae is common under low-fertility soil conditions, particularly low P availability (Davey 1995), where it is attributed to the fungus capacity to absorb nutrients directly from decomposing residue surfaces and absorbing the ones released after mineralization and making them available after death, reducing chances of P retention in acid soils (Jordan 1985). Inoculation with microorganisms is a common practice in forest nurseries for species like Acacia mangium (Pérez et al. 1998; Martin-Laurent et al. 1999; Schiavo and Martins 2003), Alnus acuminata (Echandi 1994; Gardner et al. 1984; Michelsen and Rosendahl 1990; Rondo´n and Hernández 1995; Russo 1995; Budowski and Russo 1997), pines (Costa et al. 2002; Carlson and Dawson 1985), eucalyptus (Trappe 1977; Xianheng et al. 1998; Adjoud and Halli 2000), and teak (Verma and Jamaluddin 1995; Raman et al. 1997; Kelly et al. 2004; Durga and Gupta 1995; Ramírez et al. 2011; Aditya et al. 2009; Zhou et al. 2012). However, results are commonly found to be erratic or negative since the introduced microorganisms do not efficiently compete with native and more adapted microorganisms that end feeding on them and inoculating seedlings (Marx et al. 2002). Soil microorganisms and root distribution in plantation ecosystems are closely associated (except for the free-living organisms) topics recently reviewed by Vogt et al. (2011). In general root distribution in the soil profiles decreases with soil depth and the distance from the tree trunk and so does population of soil organisms like nematodes, bacteria, and fungus (Srivastava et al. 1986; Behling 2009). Root biomass is also affected by vegetation species composition as reported by Jiménez and Arias (2004) who found 441 g m2 root biomass under a 24-year-old secondary forest of the humid tropical area of Costa Rica but only 75 g m2 under the nearby grassland.

1128

A. Alvarado

Freiberg (1998) mentions that free-living organisms (Scytonema, Cyanobacteria) can fix up to 2–5 kg N ha1 year1 on the phyllosphere of the premontane forest trees of Costa Rica. This N fixation mechanism might explain why many forest ecosystems do not show N deficiencies (Hedin et al. 2009) as well as the high foliar concentration of N described in other species like teak (Nwoboshi 1984) and Gmelina (Rodríguez 2006).

Contribution of Nutrients by Soil Weathering Nutrients supplied by the soil to forest growth by weathering can be small or large depending on factors of soil formation, mainly the composition of the rocks where they form. In Suriname where soils develop from highly siliceous materials, they show sandy texture, and during the rainy season nutrient contribution to the trees is meaningless (de Graff 1982). In the tropical humid forest soils of Costa Rica developed from basaltic to andesitic rocks, nutrients are more abundant, and lateral underground nutrient additions occur due to the presence of lavas that impede deep leaching (Pringle et al. 1986; Pringle et al. 1990; Pringle et al. 1993; Generoux and Pringle 1997; Generoux et al. 2002; Jordan 2003). Aboveground Residue (Litter) Accumulation Accumulation of large quantities of residues in forest plantations is not frequent. According to Reddy (2002) in broad-leaved plantations like teak, residue accumulation is not common, since undergrowth vegetation is limited and falling leaves decompose rapidly. In the case of Eucalyptus, the open crown and the small size of the leaves allow weeds to grow and a rapid decomposition of residues. In the case of leguminous species that fix N and produce large quantities of residues, microorganisms mineralize them rapidly due to their favorable C/N ratio. In the tropical humid lowlands, recycling is fast and happens in very short periods of time (Montagnini 2002). In the dry tropical lowlands, residue accumulation depends on the length of the dry season and their fast mineralization at the beginning of the rainy season (Bernhard-Reversat and Loumeto 2002). In fast-growing plantations the amount of residues produced is larger than the one added under natural forest conditions. In the same type of plantations, residues accumulate more under middle altitude conditions where productivity is large and mineralization rate is moderate. Under natural conditions the rate of mineralization is larger than in exotic species plantations since decomposer organisms are more adapted to the climatic conditions. Residue Mineralization Effect of Residue Quality on Mineralization: Concentration of minerals in the tissues, their organic chemical composition, and soil physical properties associated define its quality in leaves or in the total of the residues. Some properties of residues are little known (i.e., leaf thickness), while others like N content are abundant, and to a lesser degree the presence of soluble organic substances, secondary metabolites like phenols and tannins, fiber, cellulose, and lignin contents is relatively unknown. In very rare cases the content of tannins is mentioned due to its allelopathic effects

Plant Nutrition in Tropical Forestry

1129

(Bernhard-Reversat and Loumeto 2002). Verhoef and Gunadi (2002) attribute the pine needles’ relatively slow decomposition rate (50 % after 7 years) to its high tannin content, the low pH, and nutrient contents of the organic layers of the soil, common in mountainous tropical regions. Hard-to-decompose residues are considered as a good attribute for trees to have when coverage is needed in site restoration. Montagnini (2002) mentions the case of large amounts of hard residues added by V. ferruginea against lower amounts of soft residues added by H. alchorneoides in relation to reduce soil erosion. Rate of Mineralization: Fresh forest residue mineralization rate lasts approximately 15–20 months (Fassbender 1987; Babbar and Ewell 1989; Byard et al. 1996; Horn and Montagnini 1999), except for woody residues which might last up to 10 years to decompose under tropical humid forest conditions and even more in drier and cooler ecosystems (Poels 1994). Nutrients associated with the residues are rapidly released at first, changing to a moderate rate and to a very slow rate at the end in a synchronized mechanism similar to the one trees use to absorb nutrients after planting. During the first 4 months, more than 50 % of the residues decompose, and after 14 months 80 % of the residues are gone (Horn and Montagnini 1999; Boniche et al. 2008). Potassium is released during the first 3 months (Waterloo 1994), while Mg, P, N, and Ca are released more slowly leaving a small amount of materials that accumulate in the soils associated with the organic matter fraction. In mountain tropical forests 15–70 % of residues decompose per year, while in tropical humid lowland forests, more than 80 % are mineralized per year (Fassbender 1987) and usually all residues mineralize in less than 15 months (Montagnini and Jordan 2002). The slow residue mineralization rates in mountain ecosystems is attributed to (i) the low soil temperature, (ii) the low relative humidity related to relief conditions, (iii) the low N concentration of the residues (Heaney and Proctor 1989) that also induces N limitations for vegetation growth (Tanner et al. 1998), (iv) the increase of phenols of residues with elevation (Montagnini and Jordan 2002), (v) the low population densities of arthropods at high elevation (Bruhl et al. 1999), and (vi) the large amounts of soil organic matter that form humic-allophanic complexes resistant to decomposition at middle elevation Andisols. Coefficient of Decomposition (K ): Some authors use the ratio “deposited residues/remnant residues” (K ) to measure the rate of residence of residues in the ecosystem. The (K ) values do not relate to the age of a tree or plantation, but it relates to species genera, origin of the plant material, and climatic conditions. Results show that lower values are found in pine and eucalyptus residues, while species like Terminalia rank in the high range. If addition of residues is continuous, then the decomposition rate follows a linear model with time and (K) values in plantation 12–24 months old in India that vary in the order of 0.69–0.91 for A. nepalensis, 0.39–0.93 for C. equisetifolia, and 0.46–0.78 for P. kesiya (Akinnifesi et al. 2002); for various Eucalyptus spp., they varied in the range of 0.30–0.94 (Lima 1996) and in tropical natural forests and plantations in a range of 0.2–5.0 (Bernhard-Reversat and Loumeto 2002). Decomposition rates of fine roots vary among tropical humid forest species being very slow (K = 0.29  0.15 year1)

1130

A. Alvarado

for V. koschnyi and seven times faster for V. guatemalensis (K = 2.00  0.13 year1), while for H. alchorneoides and P. macroloba, they were 1.36 and 1.28 year1, values related to the lignin/N ratio of the roots (Raich et al. 2009). Effect of Decomposer Population: Decomposers are very abundant in forests under natural conditions where they are a key factor in biochemical degradation but also in comminuting residues and then augmenting the surface area exposed to mineralization. Wilde (1958) considered that the smaller the organisms, the more abundant they are and the more specific is their function in the residue cycling. The complexity and specificity of microorganisms in the soil are considered important to assure ecosystem sustainability and soil health (Kibblewhite et al. 2008). According to Garretson et al. (1998) leaf-cutter ants (Atta cephalotes) affect the diversity of plants in tropical humid lowland ecosystems. Heneghan et al. (1998) estimate the effect of micro-arthropods on residue decomposition from various tropical forest ecosystems using decomposition bags treated and untreated with naphthalene finding that decomposition rates varied with N residue content and the N released through decomposition was significantly larger when micro-arthropods were more abundant. Microbial activity is moisture dependent and varies with rainfall distribution. During the rainy season residue decomposition is enhanced due to a larger population of micro- and macroorganisms (particularly that of earthworms, ants, and termites) and the proliferation of fine roots. Microorganisms like fungus and bacteria play an important role in decomposing residues along the year but are more relevant during the dry season when macro-arthropod activity reaches it minimum (Costa et al. 2002). Mesofauna of arthropods is normally larger (at least 1.5 times larger) on residues of exotic plantation of eucalyptus where they favor from the microclimate created by the shade of trees rather than the harsh climatic condition of the adjacent open savanna (Reddy 2002). However, this effect is temporal since with time the quality of residues under the plantations changes negatively affecting their populations (Folgarait 1998; Bernhard-Reversat and Loumeto 2002). Other studies conducted to better understand diversity and abundance of collembolans (Guillén et al. 2006a, b) and scarabide (Hall 2003) under primary and secondary forests and coffee-cocoa plantations in Costa Rica document reductions in diversity and abundance while changing land use and the effects of soil compaction, organic matter (litter) content, and low fertility and pH on the variables. Studies carried on in Java on P. merkusii plantations (Reddy 2002) and India’s eucalyptus plantations (Verhoef and Gunadi 2002) show that among arthropods collembolans and acari are the most abundant dominating populations in the upper soil layers of plantations and adjacent savannas. Samra and Raizada (2002) also report that collembolans and acari dominate the population of arthropods over earthworms that do not tolerate high amounts of needle litter under conifer species. The effects of lignin and phenols on the wood of various species on population dynamics and residue decomposition of termites and ants are being documented by Samra and Raizada (2002) and Bernhard-Reversat and Loumeto (2002).

Plant Nutrition in Tropical Forestry

1131

Nutrient Fate After Residue Mineralization The quantity of nutrients in residues returned annually to the soil varies with forest/ plantation ecosystem differences and is species-age dependent. Under many different circumstances, various authors have demonstrated that nutrient concentration in aerial biomass of species varies little with age but nutrient absorption increases exponentially with tree growth (Nwoboshi 1984; Waterloo 1994; Samra and Raizada 2002; Segura et al. 2005). Nutrients released from P. patula plantations 6–14 years old and various eucalyptus species 7 years old (kg ha1 year1) are 16–89 N, 1–4 P, 3–20 K, 17–120 Ca, and 2–15 Mg (Samra and Raizada 2002). Wood et al. (2009) estimated additions of P as residues in the secondary humid tropical forest of more than 15 years of age to be 5.2–10.6 kg ha1 year1. When residues mineralize their C/N ratio decreases rapidly in the residues and the tissue decomposer bacteria and fungus. It is in order to reach a new nutrient balance that some of the elements get lost (leached) but others are used to synthesize proteins and microorganism’s biomass, a process known as “nutrient immobilization.” Later on, organic residues limit energy for decompositions and microorganisms die releasing the nutrients immobilized in the biomass; the term then is “partial immobilization.” Nutrients can also be tied very strongly to clay and oxyhydroxide surfaces, almost permanently, a process also known as immobilization, retention, or fixation. Rodin and Bazilevich (1967) cited by Gosz (1984) use the term immobilization of nutrients referring to the fact that conifers and deciduous species of temperate zones keep (“immobilize”) nutrients in their residues in the order N > K > Ca = P = S = Al = Si > Mg > Fe = Mn = Na. Burning Vegetation and Residues in Tropical Forest Ecosystems Burning the land is the most common mechanism to clear areas for cropping or raising cattle in tropical regions. Fires also are used to cheaply (not sustainable) convert residues into nutrients and lime (slash and burn and slash and mulch), remove weeds and pests in grasslands, reduce expansion of diseases and pests, and pasture regeneration. From the forestry point of view, it is worth to mention that planting pyrophilic species such as pines, eucalyptus, and bamboos might increase the quantity of flammable material in tropical environments. More ecological implications of fires in tropical regions can be looked in papers written by various authors (Nye and Greenland 1960; Ewell et al. 1981; Otsamo et al. 1995; Richter et al. 1982; Lal et al. 1986; Balagopalan 1987; Cochrane 2009; Rodríguez et al. 2011). Naturally forest fires might start under climatologically predetermined conditions (i.e., long dry periods) and of management (residue accumulation after pruning during the dry season) releasing instantly all nutrients associated. Their effects as related to the additions and losses of nutrients are observed in all altitudinal belts, in any soil depth, and far away from the fire occurrence (Forti et al. 2000; Titiz and Sandford 2007; Boy et al. 2008). Fires can be divided considering their intensity, duration, and effect on the soil properties as “partial fires” when they occur naturally or are provoked by humans by chance and “total

1132

A. Alvarado

fires” when the forest is clear and burned intentionally. Most fires in the world concentrate in tropical regions (Cochrane 2009) particularly in monzonic areas where they concentrate during long dry seasons associated with ENSO fluctuations as it happens in Nicaragua, Honduras, and Guatemala in Central America (Observatorio del Desarrollo 2002; Asociacio´n Conservacio´n de la Naturaleza 2005). Fire effect on soil properties starts by diminishing organic matter content and Al saturation at the soil surface with an increase of soil base saturation (this last effect not as clear under pine litter burning). Waterloo (1994) reported that the decomposition and burning of residues of P. caribaea in Fiji reduced their amount from 37 to 6 t ha1, increasing some nutrient availability and losing large amounts of P, N, and K. Forest fires also might accelerate erosion of bare soils at the beginning of the rainy season and the volatilization of nutrients like C (31 %), N (22 %), and S (49 %) (Ewell et al. 1981). In terms of homogeneity soil fertility properties are highly deteriorated since ash deposited on the ground concentrates around large pieces of wood or where residues are accumulated before burning. Frequently soil sterilization is mentioned as the major damage due to the high temperature developed by the fire over the soil, the situation being enhanced when soils are compacted (Costa 1990); however, the soil humidity, the length of time required to burn large amounts of residues, and the high porosity and low thermal conductivity of the soils impede soil temperature to increase but only in the atmosphere-soil interface. Ewell et al. (1981) report burning temperatures on the soil surface over 200  C that decrease to 100  C at 1 cm and lower than 38  C at 3 cm depth and conclude that soil sterilization should not be a problem ever. A similar conclusion was found by Ramsay and Oxley (1996) after finding temperature oscillation between 400 and more than 500  C on the burning herbaceous vegetation of the Ecuadorian Paramus but lower than 65  C at 2 cm soil depth. Forestry-wise it is relevant to mention that clearing the land by burning increases the germination of the Gramineae family making it more difficult (at least more expensive) to establish plantations (Horn and Sandford 1992). Grasses compete with trees for light, water, and nutrients and might produce allelopathic substances and increase chances for soil compaction and fire development (Otsamo et al. 1995). Hartemink (2003) and Costa (1990) also mention that after fires soil water infiltration capacity is reduced due to superficial pore sealing and also by increasing soil hydrophobicity. Fires also provoke a reduction of fine roots attached to organic particles, and in some cases enhancing shoot production also correlated with a process of blocking water and nutrient absorption; high occurrence of nutrient deficiencies in the shoots is also associated with dystrophic environmental conditions. The loss of soil coverage and nutrients by erosion retarded the initial growth of teak plantations (Ramnarine 2001). The amount of ashes deposited after prescribed fires is relatively small (0.58–1.14 t ha1), but the burning of residues that might be considered as fuel reduces the possibilities for natural fires to occur and also helps in preparing the land for establishing new plantations (Richter et al. 1982; Zelaya and Guillén 2002; Shlisky et al. 2009). Larger quantities of ash are deposited after clearing and (total)

Plant Nutrition in Tropical Forestry

1133

burning of the land (1.2–12.1 t ha1) depending on the type and age of the forest cut and the natural fertility status of the soils where the forest grew (Nye and Greenland 1960; Ewell et al. 1981; Richter et al. 1982; Lal et al. 1986; Balagopalan 1987). In general elements like N, Ca, Mg, and K are the most deposited followed by those of Fe, S, P, and Mn and very much less by those of Zn, Cu, and B. Drechsel and Zech (1994) consider that the N added this way plus the one that forms rainfall might supply over 70 % of the N required for a turn of teak. Szott et al. (1991) and Fo¨lster and Khanna (1997) established that large quantities of the nutrients deposited as ash will not be available for trees to grow but will be lost through leaching, erosion, seepage, and volatilization. Nutrients like P and in some cases K can also be retained on clay-sized particles (Sánchez 1985). Cravo and Smyth (1997) proved that nutrient losses after burning the residues vary within elements; just after burning Ca, Mg, K, and Zn, pH increases, but exchangeable Al is reduced; however, Ca, Mg, and K contents decreased by 50 % after 23, 15, and 5 months later, respectively. Costa (1990) adds that in hilly lands nutrient losses are high due to erosion and wind removal. Ewell et al. (1981) report volatilization losses after burning a humid tropical secondary forest in the order of 1.600 g C m2, 49 g N m2, and 13 g S m2, associated with aeolian erosion and leaching losses of g N m2, 20 g K m2, 1 g P m2, 39 g Ca m2, and 7 g Mg m2. Sarmiento (1984) considers that S losses via volatilization are not limiting the establishment or growth of new plantations in the same ecosystem since aerial biomass might have 5 kg ha1 (assuming nutrient concentration to be 0.1 % and a biomass of 500 g m2), quantity being similar to deposition of S through rainfall and aerosols in suspension. Even though volatilization losses of N are large, its soil concentration increases after burning the land (in the same manner N content is increased in drinking water in areas with high volcanism activity) suggesting that this practice favors symbiotic and nonsymbiotic N fixation.

Silvicultural Implications According to Miller (1984), there are two or three distinct phases of plantation development and nutrient requirements, showing that nutrient deficiencies are essentially problems of youth and old age: 1. In the early stages of stand development prior to canopy closure, the annual rate of nutrient accumulation increases rapidly and tree growth is very dependent on current nutrient uptake. Mineral deficiencies are common during this stage. Turner (1986) shows that Eucalyptus grandis on poorer quality sites reaches its peak of biomass accumulation some years earlier than on the better quality sites. On the other hand, Herbert and Scho¨nau (1991) found that fertilization will increase CAI in Pinus taeda but not the time where CAI peaked. 2. Once the canopy has closed, the reduction in the rate of nutrient accumulation is associated with attaining maximum foliage biomass, high internal retranslocation of mobile nutrients, as well as increasing amounts of nutrients

1134

A. Alvarado

in litterfall and by capture from the atmosphere. This will decrease the nutrient contribution by soil reserves to the amount incorporated in the wood or less. Therefore, fertilizer response will be unlikely during this second stage (Miller 1984), unless thinning will not return the stand to the first stage (prior to canopy closure). The magnitude of nutrient requirement met by retranslocation may be the dominant influence on the longevity of response to early fertilization (Switzer and Nelson 1972). On the other hand, it is supposed that on marginal sites, early fertilization (enhancing root growth) may become eroded later in the rotation as internal competition within the stand sets in. With decreasing stand density by thinning, the nutrient demand and content of the stand decrease but of the single tree increase due to enhanced biomass production, especially for renewed crown development. The magnitude of crown diameter increase depends on species and age. In this stage fertilization may accelerate recovery, but also weed growth and water depletion. Studies on teak and Pinus caribaea in Nigeria showed that after recovery, the amount of litter has reached the same or higher levels in comparison with unthinned control (Egunjobi and Onweluzo 1979; Nwoboshi 1980). Since on the other hand the nutrient demand of the thinned teak stand was reduced, the nitrogen balance of this plantation became more harmonized by thinning (Nwoboshi 1980). A major role of nutrient management is to find the best compromise between ecological and silvicultural rotation in view of perpetuity. Due to increased nutrient use efficiency in the stand with time, lengthening the rotation age of fast-growing species will reduce net nutrient uptake (from soil reserves) (Switzer and Nelson 1972). “Nutritional costs” decrease especially after heartwood commences to form due to P and K withdrawal. In many eucalypts this will be at about age 5–7 years, but in P. radiata about 10 years later. Therefore, early thinnings or short rotations (5 years) will cause high nutrient removal due to the high nutrient contents in the sapwood and increasingly greater removals by pine than eucalyptus harvesting in rotation periods longer than about 7 years. For Eucalyptus delegatensis, the P export by harvesting wood and bark increased from 51 g P per ton of wood (57-year rotation) to 87 g P per ton of wood (18-year rotation). For Pinus radiata, the comparable increase was from 169 g per ton of wood (40-year pulpwood and log rotation) to 258 g P per ton of wood (18-year pulpwood rotation). Therefore, eucalypts on longer rotations seem to be more efficient (mainly in their phosphorus requirements) than the pines under study (Crane et al. 1981; Baker and Attiwill 1985). 3. There may be a third phase (Miller 1984) in older stands or in successive rotations of short-rotation forestry, where growth rate decreases due to disturbed nutrient cycling and soil nutrient depletion. The reasons may be, e.g., N or P deficiency due to nutrient “lockup” in undecomposed litter or continuous N loss by fire (see section “The Nutrient Cycle”). In the example of Fig. 1 (Drechsel and Zech 1993), stem and branch harvesting at the age of 25 years will result in a loss of 872 kg N ha1. The input by precipitation (and N2 fixation) over 25 years will be in the same order. This agrees with Stewart and Kellman (1982), Hase

Plant Nutrition in Tropical Forestry

1135

and Foelster (1983), Chijioke (1980), and Baker and Attiwill (1981) that undisturbed older forest plantations are nearly in N balance. On the other hand, the demands of K and P usually exceed atmospheric inputs. Mean pantropical inputs (kg per hectare and year) of water-soluble nutrients present in bulk precipitation are 32.5  25.3 kg N, 0.41  0.31 kg P, 7.6  6.2 kg K, 12.2  9.9 kg Ca, and 6.4  3.5 kg Mg (Stewart and Kellman 1982). Besides precipitation, inputs due to stemflow and throughfall are of significance, e.g., for Ca (Yadav and Mishra 1982). In view of successive rotations, it is strongly recommended for, e.g., teak and smooth-barked eucalyptus that bark and foliage be left in the plantation during harvest (Turner and Lambert 1983; Hase and Foelster 1983; Ferreira et al. 1984). Large amounts of Ca (and P) are stored in the bark of these species and only smaller amounts in the bark-free bole. Although fertilizing may correct intermediate-age deficiencies, research is also necessary into the management of organic matter, species rotation, and the use of cover crops and other practices which improve soil fertility. The correct management of N2-fixing species either as groundcover or understory (e.g., Lupinus spp.) or accompanying species (e.g., Leucaena spp.) may be a critical factor in tropical forestry. Once the perennial tree legume Lupinus arboreus had become established in New Zealand pine plantations, natural seeding ensures a lupin understory between rotations and whenever thinnings reduced tree canopy (Nambiar et al. 1984). Maggs (1985) pointed out that a second P addition could also have a positive effect on the N economy of (pine) plantations by increasing nonsymbiotic N2 fixation in the upper forest floor (and litter breakdown) due to higher P concentrations in the litter. The nutrient dynamics over several rotations are under discussion, since there are only few long-term projects (Sanchez et al. 1985). An attempt to predict the expected reduction of total nutrient stores during transformation by Drechsel and W. Zech of exploited rain forest and during successive generations of forest plantations is given by, e.g., Fo¨lster and Ruhiiyat (1991).

Visual Diagnoses of Mineral Disturbances A tree or stand may suffer a “nutritional disturbance” if the supply of one or more nutrients is either too low or too high for optimum growth. The lack or excess of one or more nutriments produces visible symptoms that, when observed, have already negatively affected growth. Major symptoms include chlorosis, death of tissues, and growth reduction and are specific for each nutriment. Nutritional disturbances which are caused by excess of nutrients are rare in natural forests or plantations with some exceptions mentioned below. The common case is insufficient nutrient supply per se or induced by other environmental factors.

1136

A. Alvarado

Causes and Occurrence of Mineral Deficiency and Toxicity Causes: Major soil constraints associated with nutrient deficiencies or sufficiencies include the following (after Evans 1992 and Webb et al. 2001): • Sandy and eroded soils or soils derived from large deposits of garbage, mine deposits, etc. • Low-fertility soils where residue decomposition allows litter accumulation and then nutrient immobilization in the topsoil. • Nutritional deficiencies induced by factors like: 1. Nutrient leaching in very rainy areas 2. Rain shortages inducing nutrient deficiencies during the dry season 3. Overliming inducing Fe and Mn deficiencies 4. Low pH effects like P precipitation under alkaline conditions or retention as P-Al and/or P-Fe in acid soils or P retention on short-range crystallization order clays 5. Soil physical properties that affect root growth like shallowness, water pounding, compaction, etc. • Interactions with other nutrients. Concentration of nutrients in the soils can affect requirements of others, notably the relations N-P and P-K. The addition of fertilizer P might induce K deficiency in soils without the deficiency under natural conditions. • Low or inexistent mycorrhizal associations or poor population of N-fixing microorganisms. • Strong weed competence. • Planting the trees in wrong sites, i.e., Leucaena leucocephala in soils with pH below 5. Occurrence: Nutrient deficiencies might be the result of (i) the stage of growth of the trees or (ii) interactions between soil, site, and species. In the first case it should be remembered how nutrient dynamics in a plantation changes as trees grow and pass from one stage of growth to the next. However, it is worth to mention that nutrient deficiencies, or the lag of, are more common during the first stage of growth and few weeks after establishing the plantation as follows: • After planting seedlings, do not show deficiencies until the residual effect of fertilizer applied in the nursery disappears. • Once established, for a period that lasts a few weeks up to a few months, seedling growth in height, branching, surface area, and root system growth are accelerated augmenting nutrient needs. • Nutrient cycling in soils is almost nil until the crown fully develops. • Root math development might not be capable to compete with weeds, especially when mycorrhizae are not present like what happens in acid soils or other adverse soil physical properties in teak plantations.

Plant Nutrition in Tropical Forestry

1137

Table 2 Nutrient deficiencies commonly found in Cuba by species, soil type, and region (Adapted from de Herrero 2001) Especies Pinus tropicalis

Pinus maestrensis Pinus cubensis Eucalyptus saligna Eucalyptus citriodora Tectona grandis Swietenia macrophylla

Soil type Ferralítico cuarcítico amarillo Rojizo lixiviado Ferralítico rojo lixiviado

Limiting elements P > Ca > B > K

Region Vin˜ales

Ferrítico pu´rpura hidratado Arenoso cuarcítico

N>P>K

Vin˜ales Los Nu´meros, Sierra Maestra Pinares de Mayarí

P>B>N>K

Las Taironas

Ferralítico cuarcítico amarillo Gley oscuro plástico Ferralítico rojo típico

P>B

Galindo

N > Mg > P P>B

Itabo, Villa Clara Artemisa

P > N > Mg

In the second case, many factors interact affecting tree growth and nutrient availability making it difficult to predict when problems are to be encountered. It is more the personal experience on the sites that will help in “expecting” problems like water deficits that will induce the deficiencies of N or the lower absorption of P by the tree. Waterlogging conditions affect nutrient availability and absorption as demonstrated by Hernández et al. (1993) in teak plantations where foliar concentrations of Mg, Fe, and Mn were higher in well-drained soils but K and P concentrations were lower. Natural geographical distribution of forest species (Table 2) helps planting commercial plantations in Cuba avoiding nutrient deficiencies (Herrero 2001).

Symptoms of Mineral Deficiency and Toxicity There are two intensities of deficiency: (i) latent deficiency without any visual foliar symptoms but with lower nutrient levels and some lower growth than under optimal conditions (the observation of foliar deficiency symptoms is a simple method to detect mineral disturbances) and (ii) visual deficiency with reduced or stunted growth and/or foliar discoloration and in severe cases shoot dieback and mortality. Latent as well as visual mineral deficiency could be best seen and verified by a positive response after some months to controlled fertilization experiments (diagnostic fertilization). Due to high requirements in time and costs, this method will only be practiced in commercial plantations. Great efforts have been made been made to find other diagnostic methods such as (i) visual symptoms in the trees, (ii) greenhouse missing element technique, (iii) field experiments to determine fertilizer and lime requirements, (iv) searches for

1138

A. Alvarado

indicator plants to identify nutrient deficiencies, (v) chemical and physical soil analysis, and (vi) soil mapping that helps in selecting best sites to plant. Symptom description is differently made for conifers and broad-leaved species (Zo¨ttl and Tschinkel 1971; Cannon 1983b) as well as for their rotation length (short, medium, long). Short-term rotation length species are Eucalyptus, Acacia, Gmelina, and Pinus; medium-term rotation species are Cedrela, Terminalia, Aucoumea, samba (Triplochiton scleroxylon), Araucaria, and Agathis; and long-term rotation species are teak, caoba (Khaya and Entandrophragma spp.), and Chlorophora excelsa (Brunck 1987). Two major characteristics of the intensity of foliar symptoms can be differentiated: (i) chlorosis (reversible yellowing or loss of green color) and (ii) necrosis (irreversible brown discoloration, dead tissue). Mobile elements in the tree (N, K, Mg, P) show symptoms in older tissues, while less mobile elements (Ca, B, Zn, Fe, Mn) show symptoms first in younger leaves and shots. The detailed description of the kind and location of these phenomena on the leaflets is important for identification and differentiation of the symptoms: there could be chlorosis or necrosis at the leaf margins, as spots on the leaf and as patches between the veins (intercostal or intervenal), at the tip of needles or leaves, etc. In several cases the margins are scorched or look “burnt,” the leaf surface could be wrinkled, the tips could be curled, etc. (Fig. 2). The age of the affected foliage gives hints as to the mobility of the deficient nutrient: for example, intercostal chlorosis on older leaves may be caused by retranslocation of the “mobile” Mg from older foliage toward younger shoots; the same symptom on young leaves could be caused by deficiency of the usually less mobile Mn. Although there are for some elements typical symptoms which do not greatly vary between the species, several nutrient deficiencies do not induce a specific symptom or the same symptom on different species, like deficiency. Therefore, the given keys for the identification of mineral deficiencies should be used with care. In addition, the more common case in nature is multiple deficiency of more than one nutrient due to marginal soils (e.g., low organic matter: N, P, S deficiency) or physiological interactions (N-P, N-Cu, P-Zn, etc.) resulting in more or less complex deficiency symptoms (Lambert 1984; Zech and Drechsel 1992; Webb et al. 2001). Symptoms caused by plagues, diseases, and agrochemicals can be confused with nutritional deficiencies, and some criteria are being developed to properly identify each of them. Problems of identification also arise from abnormalities or discolorations caused by water stress, leaf aging before leaf shed, and virus and similar diseases or combinations of nutrient deficiency followed by virus attack. In addition, deficiency symptoms may vary over the vegetation period due to changing nutrient availability in the soil or requirement in the plant. Toxicity Symptoms: These are usually less specific than most deficiency symptoms. Common toxicity problems are due to excess of Al or Mn in the humid tropics or excess of B, e.g., in the arid tropics. While Mn and B are essential for all plants, Al is essential for only a few plants like teak, but toxic to a wide range of crops and trees. In most cases Mn toxicity may be visible by dark brown spots on older leaves and veins, starting at the margins. Al toxicity may reduce growth by root damage,

Plant Nutrition in Tropical Forestry

1139

Fig. 2 Diagram showing position of new shots and nutritional deficiency symptoms in eucalyptus (Taken from Dell et al. 2001)

1140

A. Alvarado

inhibition of mycorrhiza, and uptake of several nutrients including Al itself. Therefore, there is no clearly identifiable symptom. On the other hand, several tropical trees like Pinus taeda and P. elliottii are well adapted to moderately acid soils. Their foliage could show Al concentrations of more than 1,000 ppm without physiological problems (Reissmann 1981; Zo¨ttl 1978). The same has been reported for well-growing Eucalyptus viminalis and E. grandis on acid soils in South Brazil (e.g., Bellote 1990; Neufeldt, personal communication) which passively take up high amounts of Al. According to these circumstances, the tissue which is best suited for determination of deficiency may not be the best for toxicities. For Al (Cu, Na) often fine root analysis (e.g., Ca/Al ratio) is recommended. Similar to Al, Mn tolerance varies greatly between the species. Manganophobic trees like cypress may suffer from Mn toxicity even at low Mn levels ( Nitrogen Spotting First, Yellowing Later (a) Numerous dark spots appear on the green leaf. These subsequently increase in size and the background develops an orange-yellow tint. Leaves of seedlings have bluish-purple blotches, and the surrounding green is darker than normal. Branching is restricted as in nitrogen deficiency. — > Phosphorus (b) Reddish spots appear on a pale green background. As the deficiency becomes more acute, the tissues in the reddish areas die (necrosis), and the leaves wither and fall. — > Calcium Yellowing Between the Main Lateral Veins Older leaves yellow along the midrib. Green color gradually changes to brown with dying of tissues. The affected regions are separated from the main lateral veins by areas of green tissue. The lower

1142

A. Alvarado

leaves of seedlings become pale green and are often shed prematurely, leaving a bare stem with a tuft of leaves at the top. The seedlings of E. grandis, E. saligna, and E. botryoides, in particular, produce larger than normal leaves at which have developed in the shade. — > Magnesium

Symptoms First Localized in Younger Leaves Uniform Yellowing Younger leaves show a uniform yellowing, later changing to a bronze-like color. Branches show a purplish tinge. — > Sulfur Mottling (a) Yellow mottling appears on the leaf blades, with the areas along the veins retaining their green color. Occurrence is mainly on calcareous soils unless foliar leaves >140 ppm Fe. — > Iron (b) Yellowing appears between the veins, but tissues near the veins remain green in color. As the deficiency becomes more acute, the tips and margins of the leaves begin to wither and show a sandy color, which spreads throughout the blade. — > Manganese Yellowing Between the Veins Leaves are normal in size and shape. (a) Yellowing of younger leaves occurs between the lateral veins, starting at the leaf margins and proceeding toward the midrib. Alongside the lateral veins the tissue remains green but later takes on a purplish tinge. The lower surfaces of the leaves become light green. This occurs mostly on acid granite-derived soils, on volcanic soils, and especially in dry climates unless foliar levels >46 ppm B. — > Boron (b) Yellowish spots occur between the lateral veins of mature leaves. Narrow bands along the veins remain green, with a purplish color along the leaf margin. — > Molybdenum Leaves are normal in size and abnormal in shape. (a) Younger leaves become yellow between the lateral veins, accompanied by deformation of the blade, with leaf margins irregular. This occurs often on sandy soils (Spodosols) unless foliar levels >6 ppm Cu; symptoms will be enhanced by N(P) application. — > Copper (b) Leaves are abnormally small in size and narrower in shape. In the beginning, the upper surfaces of the leaves show purplish areas between numerous discolored patches or spots. Small circular areas of lighter discolored tissue with brownish edges occur near the leaf margins far from the midrib. The whole

Plant Nutrition in Tropical Forestry

1143

leaf becomes pale green, the veins being darker in color, and later shortening of stem length forms a rosette of small, narrow, yellowish leaves. This occurs often on sandy soils (Spodosols) unless foliar levels >10 ppm Zn. — > Zinc Leaves are not abnormally discolored, but margins and veins are dying off. Leaves of seedlings are smaller than normal, often with crinkled surfaces and margins. Branching is pronounced, giving a bushy, round-topped appearance. — > PotassiumPinus: For Pinus foliar symptoms for micronutrient deficiencies are described by Hill and Lambert (1981). Macronutrient deficiency symptoms typical for pines are given by Bergmann (1988) and for conifers summarized by Weetman and Wells (1990). Hernández and Lombardo (1987) show photos of induced deficiency symptoms in seedlings under greenhouse conditions. N S P Mg K

Markedly reduced shoot growth and uniformly light-green to yellow needles, occurring first in the older foliage. Partly stunted needles. In general similar to those of the N deficiencies, but more on younger foliage. Untypical symptoms on older foliage: yellow (later dead) needle tips or purple-brown-tinged needles or thin and small needles. Yellow (“golden”) needle tips on older foliage with distinct borderlines to the rest of the needle. Variable symptoms (after tip yellowing) brown-reddish needle tips on older needles and no distinct borderlines to the rest of the needle or, e.g., bluishgreen needles.

With the exception of Mo, micronutrient deficiency appears first on the youngest parts of the trees: B

Cu

Fe Mn

Terminal and leader dieback often with orange-red discoloration. Resinous bud often fails to flush; main stem forks (multiple leadering) and becomes deformed. Black or dark brown pith and shortened needles. Deficiency (mature needles) between 5 and 12 ppm dependent on rainfall [the higher the moisture stress, the higher the foliar levels at which the tree develops deficiency symptoms; Lambert (1984)]. Often dark blue-green foliage, distorted (snake-tailed) shoots and branches, and bushiness. Some needle tips burnt. Prostrate growth in extreme cases. Deficiency range: less than 2–4 ppm. In Australia, in most cases (acid Spodosols) trees less than 5 years old are affected. Application of N (PK) fertilizer accentuates the deficiency. Older needles green. Increasing chlorosis toward younger foliage. White colors in extreme cases (only common on calcareous soils). Yellow-tipped needles, sparse, light green foliage, stunt. Deficiency range: 10). For multiple regression or discriminant analysis, the number of observations should be at least twice the number of parameters in the equation. Plot Size: For the study of mineral disturbances in tree plantations, e.g., by correlation analysis, a common plot size of 20  20 m (0.04 ha) to max. 30  30 m (0.09 ha) will be representative for the stands in view of growth, density, as well as topography and soils. For foliar sampling (see below) there must be several (co) dominant trees on the plot. In natural forests a plot size of 50  50 m will be more representative in view of botanical variability. In plantations with obvious variations in tree height, a larger representative plot size or several subplots representing the diversity are necessary. For experimental studies on mineral nutrition (e.g., fertilization), no rigid rules can be laid down for plot size. It depends on the number of trees necessary (see below), tree spacing, and on the number of buffer rows required. The latter may depend on tree age (Cellier and Oorrell 1984). In cases where there is a range in soil or plant properties within a treatment block, the plots should each contain the full range of the gradient conditions The following recommendations mainly concern tree plantations. Plot Description: The value of interpretation of soil and foliar data often depends on a careful plot description. The following data may be necessary and/or valuable in nutritional studies in established stands: – Location and plot number – Growth parameters (if available m3 ha1, site index or mean top height, DBH root collar diameter, etc., of the sampled dominant or codominant trees) – Planting year (or if coppiced, age of the stand) – Trees per hectare (former and recent spacing)

Plant Nutrition in Tropical Forestry

– – – – – – – – – –

1149

Provenance of the trees Previous land use (farmland, natural forest, fallow, etc.) Thinnings, possibly best stems removed Weeding (how long, how frequent) Occurrence and frequency of controlled fire or bushfire Degree of crown closure, understory, weeds (species, ground covering) Pests, diseases, termite, or fungal attack Deficiency or water stress symptoms of the foliage or shoot Fertilization Flowering/fructification at present

Site and Soil Description Nutrient availability depends on several parameters other than those detected by soil nutrient analyses, like limited water availability or excess of water. However, in the following paragraphs we generally deal with questions regarding sampling for chemical analyses. Soil analysis usually refers to topsoil samples (0–20 cm depth) collected at random with the help of a spade or an auger from at least ten points to conform a composite sample of half a kilo. Occasionally subsoil samples (20–40 cm) are taken since the subsoil might contain not enough water and nutrients for the deepest roots to absorb (i.e., Al accumulation that might reach toxic levels). Soil Pit Location and Description: The soil morphology should be studied on at least one lead profile (pit) in the middle of the plot. The representability of the lead profile(s) should be controlled during soil sampling with a soil auger (see below). If these samplings give information about a change of main soil properties, a second pit will be necessary. The depths of the pit should be about 120 cm. In the arid tropics deeper soil pits are necessary for trees with taproots; in the humid tropics over 90 % of the roots of most species are situated in the upper 100 cm. Limitations below a depth of 120 cm usually have no great influence on tree growth, until tree growth does not depend on taproot development (in arid areas for water supply). The pit should be situated in front of a dominant tree to study possible inoculation as well as the rooting behavior and rooting intensity as indicator for the most important depth for nutrition as well as root- and growth-limiting soil properties. In semiarid regions, legumes (such as Prosopis sp.) with taproots may have nodules only in greater depth near the groundwater table. Activity of nodules usually can be controlled by their reddish inside color. Soil and site characteristics which are routinely noted are described (e.g., by Hodgson 1978). From the nutritional point of view, some parameters are of special interest, depending on site and species: – Rooting depth and/or the physiological usable soil depth (down to a hardpan, a saline horizon, a pumice, or a clay layer) – Main rooting depths with the bulk of fine roots

1150

A. Alvarado

X: lead profile DT

DT

1–6: samplings with soil auger

DT

2 5

1

DT X

6 DT

DT

DT 3

DT 4

DT: dominant and codominant sample trees

Fig. 3 Schematic map of a 20  25 m sample plot in a teak plantation

– Depths with, for example, 20 % or 50 % mottles by stagnic properties (important with respect to, e.g., teak or Pinus radiata) – Bulk density and stoniness for nutrient storage calculations – Soil texture, drainage, and parent material – Darkness of the topsoil as indicator of organic matter content If there are no possibilities for laboratory analyses, at least the pH [and salinity (ECe or EC1:2) in semiarid areas] should be determined on the plots, using a field.

Soil Sampling and Analysis The difficulty of obtaining representative samples results from variability in soil properties in the vertical, horizontal, and time dimension, which may be greater than between-tree variations in foliar nutrient concentrations on the same area (Mead 1984). Samples should be located within plots according to a systematic sampling regime with a random start scheme rather than completely at random (Anderson and Ingram 1989). The number of soil samples to be taken depends on the parameter variability, the aim of the study, and the time consumption of sampling and analyses. Usually 20 samplings (soil auger) per ha are called the minimum for visual homogeneous areas; 5–6 sample points on 20  20 (3O)m plots are common. The samples are bulked to reduce analytical efforts, if on-plot variations are not of interest. Although it is very common to sample three sides of a “lead profile,” this is not recommended. Some data may explain this: in forest plantations in Togo, Benin, and the Ivory Coast, six soil samples per plot (ca. 0.05 ha) have been taken in two depths on visual homogeneous sites (Fig. 3). The mean standard deviation on 50 plots of organic carbon was 0.25 % Corg per plot and on 10 % of all plots 0.5 % Corg that means on a distance of 10 cm, Corg may increase from 0.8 % to 1.8 % in the topsoil (Table 3). One reason for these variations is frequent fires. The ash will be unequally distributed by surface flow. Generally, the mean coefficient of variation (std. dev./mean in %) was greatest in

Plant Nutrition in Tropical Forestry

1151

Table 3 Mean coefficients of variation (std. dev.) on 50 “homogeneous” teak plantation plots in West Africa (six samplings per plot) (Fig. 2, Drechsel and Zech 1993) Topsoil Corg Caex CECeff PH (H2O) Subsoil Caex CECeff PH (H2O)

19  13 24  14 21  13 43

N Mgex CECpot BSeff

19  12 22  11 18  11 55

30  20 24  17 43

Mgex CECpot BSeff

27  19 22  15 68

C/N Kex

64 31  20

Bspot

14  10

Kex

32  18

Bspot

17  14

more or less sandy soils, where a small increase in clay content causes high variability in depending parameters (CEC, exchangeable cations, etc.), while variance of pH or C/N ratio usually was small (care should be taken to avoid samples with charcoal rests or ash accumulations). Aweto (1982) found on plots of the same pedotopographic series in Nigeria mean coefficients of variation for P-Bray of 67 %. According to Oriola and Adesina (1988), a coefficient of variation over 33 % indicates heterogeneous conditions. The soil depths which should be sampled for chemical analyses depend on the main rooting zone of the species. In the humid tropics this is usually 0–10 cm, 10–20 cm, and 20–30 cm, where about 70 % of the fine roots of several common forest species are situated, as well as most (available) nutrient reserves. However, sampling up to 60 cm will give more reliable information and is recommended especially for subsoil acidity and storage calculations. In arid zones subsoil horizons of greater depths should be tested for alkalinity, salinity, or hard gypsum or limestone accumulations. Sample Preparation and Analysis: In total not more than 500 g of fine earth (2 mm (stones, iron concretions, nodules). Litter layer samples are collected with, for example, 20  20  5 cm squares. For all methods the reader is referred to the following literature: useful handbooks of soil nutrient analysis methods are published by CSTPA (1980), Anderson and Ingram (1989), and PAGE et al. (1982). Selected methods are described by, e.g., IITA (1979) and Cottenie (1980). However, before a method for “available” nutrients is used for soil evaluation, it must be checked if there are interpretation guidelines for tree species, if possible for nearly the same soil and species (see also section “Evaluation of Soil and Foliar Data by Correlation Analysis”). Guidelines for agricultural crops are given in the Tropical Soil Manual (Landon 1984) and by Walsh and Beaton (1973) or Westerman (1990).

1152 Table 4 Mean coefficient of variation (std. dev./mean in %) per plot of 50 West African teak plantations of different age and minimum number of (co)dominant trees which have to be sampled to detect differences (D) of 5 %, 10 %, and 20 % between means

A. Alvarado

Element N S P Si K Ca Mg Al Fe Mn Zn Cu

Coefficient of variation 13 14 14 16 15 17 14 20 18 21 17 16

5% 41 48 48 63 55 71 48 98 79 108 71 63

D 10 % 10 12 12 16 14 18 12 25 20 27 18 16

20 % 3 3 3 4 3 4 3 6 5 7 4 4

Number of Trees to Sample Generally, it has been shown to be more representative for a plot to sample only a few leaves from several trees than many leaves from a few trees. Measurements of between-tree variation have enabled estimates to be made of the number of trees that need to be sampled in order to detect differences of, for example, 10 % between element means of two populations ( p < 0.05). Table 4 shows the mean coefficient of variation (std. dev./mean in %) per plot of 50 West African teak plantations of different ages. In addition, the minimum number of (co)dominant trees is listed which has to be sampled in this region to detect differences (D) of 5 %, 10 %, and 20 % between means. Foliar sampling was strictly the same on all plots (Drechsel 1992). With increasing plantation age the coefficient of variation decreases slowly in view of Cu and increases some percent in view of Zn. Table 5 compares the number of trees of different species which must be sampled to allow a detection of 10 % difference between nutrient means. The data show that foliar N shows the smallest variations. For differences in the N and P nutrition between two stands, about ten trees should be sampled. For other nutrients, sampling of five to ten trees allows only interpretation of differences of 20 % or more. According to Ellis (1975), Brunck (1987), van den Driessche (1974), Weetman and Wells (1990), and Mead (1984), a rough guideline will be to sample not less than ten dominant or codominant trees per plot to detect differences between means of at least 20 %. Differences of 10–15 % need about 20 trees per plot for most nutrients; therefore, 20 trees are recommended to be adequate in nearly all cases. Differences of 5 % or less are for practical reasons not detectable.

Foliar Sampling and Analysis Guidelines for sampling of litter as well as of roots and aboveground biomass are given by Anderson and Ingram (1989). In fertilizer trials, for example, often all

Plant Nutrition in Tropical Forestry

1153

Table 5 Number of trees of different species to be sampled to allow a detection of 10 % difference of nutrient means. Eucalyptus deglupta (Lamb 1976), Pinus radiata (Mead 1984), teak (see Table 4), and Pinus elliottii (Mead and Pritchett 1974). With the exception of teak, the values are based on trees of a single site or plot. Both pine stands had been fertilized. Sample numbers calculated for Eucalyptus (Lamb 1976) concern the two crown positions (out of 12) with lowest coefficient of variation for nutrients N P K Mg Ca Al Fe Mn Zn Cu B

Eucalyptus deglupta 4 2–8 23–28 11–12 22–25 13–24 31–44 29–33 25–31 20–25

Tectona grandis 10 12 14 12 18 25 20 27 18 16

Pinus radiata 7 12 16 30 36

Pinus elliottii 2 7 26 18 32 27

69 48 11 52

27 3

trees per plot except on edges and margins are sampled. For nutrient inventory studies, trees representative of growth classes of the stand must be selected (Cochran 1977). For the study of possible nutrient disturbances, usually dominant or codominant trees are sampled in tree plantations to obtain site representative data and to avoid large tree-to-tree variations due to shading. Unfortunately, foliage sampling has not been standardized until now. Proposals on the part of the crown, age of foliage, sampling time, etc., depend on the species and in several cases on the research institute (van den Driessche 1974; Ellis 1975; Lamb 1976; Mead 1984; Brunck 1987; Zech 1991). However, there seem to be some agreements over a broad range of species. Age and Position of the Foliage: On pines, needles from the youngest fully developed shoot of the third (first to seventh) whorl below the terminal bud should be sampled. On broad-leaved species, the most recent fully formed (mature) non-shaded leaves of at least two branch ends of the upper third of the crown are taken. These leaves will have nearly the same color as the older leaves, often a somewhat paler green, but they are not as reddish green and small as immature foliage. To reduce in broad-leaved species within-crown variations of specific nutrient levels, other sample positions are possible (Ellis 1975; Lamb 1976). For several species, the differentiation between branches, twig, leaf, and leaflet is difficult (Casuarina spp., Grevillea spp., Cupressus spp.). In these cases, the exact sampled tissue should be sketched for comparison. The aspect will not influence nutrient concentrations near the equator (e.g., West Africa), but farther away (e.g., in South Brazil or North Thailand); foliage from the equator- or sun-exposed side of the trees should be sampled. Sampling Older Trees: It is easy to sample foliage of young trees. Since older dominant trees are normally not harvested for some leaves, they must be climbed to

1154

A. Alvarado

retrieve samples. The sampler has to cut one or two exposed branches of the distal upper crown, from which the leaves could be taken on the ground with care as to leaf position, age, and possible contamination (dust). Deficiency Symptoms: If there are leaves with visual symptoms outside the standard sample position, they should be sampled in addition separately and their foliar data compared with green leaves of the same position and age in healthy stands. In all cases, only to sample foliage with the most pronounced symptoms should be avoided, as due to interactions usually several nutrients are reduced. Samples of different symptom intensities can give more information on the initial limiting nutrient. All foliar samples should be described in detail (smaller than usual, uniformly chlorotic, intercostal/intervenal chlorosis, marginal necrosis, spot necrosis, etc., as well as the age/position of the affected foliage as indicator whether a phloem-mobile (older leaves) or phloem-immobile nutrient (young foliage) is limited). Sampling Time: The effect of the seasons of the year on foliar concentration of trees is important in defining the best moment to sampling; this is considered to be the one when larger concentrations with lower coefficients of variation occur (Vidal et al. 1984). Evergreen species, e.g., pines, often are sampled in the period of most stable nutrient concentrations that means at the beginning dry or dormant season, deciduous trees about 3–4 weeks before the onset of mobile nutrient translocation and coloration. Other authors suggest analyzing nutrient disturbances when nutrients are most in demand. Therefore, sampling for eucalypts in South Africa is carried out at the height of the growing season (Herbert and Scho¨nau 1991). Foliar samples should not be taken in the first 4 weeks of the rainy season and not within 12–36 h after a single rainfall. In the lower montane belt of Costa Rica, Birkelbach et al.’s (1996) sampling is done during the rainy season since significant correlations between foliar concentration of nutrients of five tree species and their soil availability exist during the rainy season, attributing the effect to the ability of the trees to explore deeper soil layer and nutrients associated than those the trees could absorb during the dry season. The rainy season is the most appropriated to sample leaves under natural conditions; in dystrophic seasonally dry tropical ecosystems, the foliar concentration of N, K, P, and S might be reduced to the half or to the third than the one of the rainy season, while Ca foliar concentration increases from the beginning of leaf formation until the half of the dry season, but Mg foliar concentration first increases and then decreases in the same period. The increase of foliar concentration of some elements is being attributed to the increase in their soil availability due to the mineralization of organic residues at the beginning of the rainy season. As a result lower foliar concentrations are found during the dry season and the higher values during the wet season coinciding with maximum tree growth, in spite of nutrients being rainwashed during the wet season (Sarmiento 1984). According to Scho¨nau and Herbert (1983), it seems that 12 months after planting is the optimum age for foliar sampling in order to detect meaningful responses to fertilization. Exceptions: Flowering trees or those with heavy cone or fruit production or trees which suffer from an insect or fungal attack should not be sampled. Trees near to

Plant Nutrition in Tropical Forestry

1155

laterite pists are usually contaminated with dust. If such avoidance is impractical, these factors should be taken into account in sample preparation, laboratory analysis, and data interpretation. Analysis of individual trees (rather than composite samples of the stand) is recommended by Weetman and Wells (1990) if sample trees vary with regard to these factors. Number of Leaves: Depending on the species and on leaf size, the number of leaves taken varies between 5 and about 30. On teak trees with a mean 10-leaf dry weight of up to 160 g in young trees, instead of four leaves, eight semi-leaves may be more representative, while on eucalypts about 20 leaves per tree may be representative. On legumes several leaflets consisting of 10–20 pairs of sub-leaflets will be sufficient for all analyses; on Pinus sp. often all needles from the sampled young shoot are taken. The analyses need only about 10 g dry weights for all common nutrients. If the on-plot variations between the trees are to be compared with differences between the plots, the foliar sample of each tree should be done in separate labeled paper bags (plastic bags can support fungal growth), which have to be marked with plot number, tree number, sample position, and possible symptoms. In other cases, the sampled leaves of all trees could be mixed per plot and a part of the bulk sample used for analyses. Sample Preparation: Usually the samples (especially those taken in the dry season) are gently washed twice with distilled water to remove dust (e.g., Al, Fe); for micronutrient analyses chloroform has partly been used with more success. However, even brief washing can remove significant amounts of soluble elements like potassium. In such situations the only answer is to divide the sample – one part washed and the other unwashed. Thereafter the samples have to be dried as soon as possible at 65–70  C for about 2 days. If no oven-drying facilities are available, thorough air-drying should be done (with protection from dust). If the analytical data are to be used for comparing foliar data between fertilizer treatments, for example, or if the foliage appears stunted, the mass (dry weight) of a constant number (e.g., 100) of needles or leaves has to be determined to calculate, for example, the N content per leaf. The leaf stalk has to be cut; on teak leaves the basal third part of the finger-thick midrib could also be removed. The samples have to be finely ground (at least 40-mesh screen). Foliar Analysis: Samples of the humid tropics (ca. >1,000 mm precipitation) should be analyzed at least for N, P, Ca, Mg, K, Mg, Al, Zn, Cu, and if possible S or SO4-S. Concerning N-fixing species, MO will be of interest as well. Foliar samples from arid regions (ca. 900 m asl), there is the station at Amani which during the period 1920–2000 shows a slight trend of declining annual rainfall, whereas the trend analysis for stations in the lower altitudes resulted in a slight increase in annual

Forest Hydrology in the Tropics

1937

rainfall (Yanda and Munishi 2007). Despite some problems with inaccurate readings, there have been assumptions on the effect of decreasing forest cover on the local climate in the Eastern Usambaras. Hamilton and Macfayden (1989) found evidence for increased annual rainfall variability since the 1960s. Rainfall has become less with rainfall events more concentrated in torrential episodes. This is in accordance with statements of long-time residents of the East Usambaras who reported that the climate on the Amani plateau has changed greatly since that time: “There is now much less mist, rain is less predictable and more concentrated in particular episodes, and it is warmer” (Hamilton 1989). However, to what extent a causal relationship exists between these reported changes in climate and large-scale deforestation which started simultaneously in the Eastern Usambaras remains unclear.

References Beven K (2012) Rainfall-runoff modelling: the primer, 2nd edn. Wiley-Blackwell, Chichester Beven K, Germann P (1982) Macropores and water flow in soils. Water Resour Res 18:1311–1325 Bonell M (1993) Progress in the understanding of runoff generation dynamics in forests. J Hydrol 150:217–275 Bonell M (2005) Runoff generation in tropical forests. In: Bonell M, Bruijnzeel LA (eds) Forests, water and people in the humid tropics – Book I. Cambridge University Press, Cambridge, pp 314–406 Bosch JM, Hewlett JD (1982) A review of catchment experiments to determine the effect of vegetation changes on water yield and evapotranspiration. J Hydrol 55:3–23 Brady NC, Weil RR (2008) The nature and properties of soil, 14th edn. Prentice Hall, Upper Saddle River Bruijnzeel LA (2004a) Hydrological functions of tropical forests: not seeing the soil for the trees? Agric Ecosyst Environ 104:185–228 Bruijnzeel LA (2004b) Tropical montane cloud forest: a unique hydrological case. In: Bonell M, Bruijnzeel LA (eds) Forests, water and people in the humid tropics – Book I. Cambridge University Press, Cambridge, pp 462–483 Bruijnzeel LA, Mulligan M, Scatena FN (2011) Hydrometeorology of tropical montane cloud forests: emerging patterns. Hydrol Process 25:465–498 Calder IR (2005) Blue revolution/integrated land and water resources management, 2nd edn. Earthscan, London Calder IR, Hall RL, Prasanna KT (1993) Hydrological impact of Eucalyptus plantation in India. J Hydrol 150:635–648 Chappell NA (2010) Soil pipe distribution and hydrological functioning within the humid tropics: a synthesis. Hydrol Process 24:1567–1581 Costa MH, Botta A, Cardille JA (2003) Effects of large-scale changes in land cover on the discharge of the Tocantins River, Southeastern Amazonia. J Hydrol 283:206–217 Da Rocha HR, Manzi AO, Cabral OM et al (2009) Patterns of water and heat flux across a biome gradient from tropical forest to savanna in Brazil. J Geophys Res Biogeosci 114:G00B12 Dingman SL (2002) Physical hydrology, 2nd edn. Prentice Hall, Upper Saddle River Dubrueil PL (1985) Review of field observations of runoff generation in the tropics. J Hydrol 80:237–264 Elsenbeer H (2001) Hydrologic flowpaths in tropical rainforest soilscapes – a review. Hydrol Process 15:1751–1759 Elsenbeer H, Vertessy RA (2000) Stormflow generation and flowpath characteristics in an Amazonian rainforest catchment. Hydrol Process 14:2367–2381

1938

S. Julich et al.

Falkenmark M, Rockstro¨m J (2004) Balancing water for humans and nature: the new approach in ecohydrology. Earthscan, London Germer S, Werther L, Elsenbeer H (2010) Have we underestimated stemflow? Lessons from an open tropical rainforest. J Hydrol 395:169–179 Giambelluca TW, Gerold G (2011) Hydrology and biogeochemistry of tropical montane cloud forests. In: Levia DF, Carlyle-Moses D, Tanaka T (eds) Forest hydrology and biogeochemistry – synthesis of past research and future directions. Springer, Dordrecht/ Heidelberg/London/New York, pp 221–259 Grip H, Fritsch J-M, Bruijnzeel LA (2004) Soil and water impacts during forest conversion and stabilisation to new land use. In: Bonell M, Bruijnzeel LA (eds) Forests, water and people in the humid tropics – Book II. Cambridge University Press, Cambridge, pp 561–589 Hamilton AC (1989) Climatic change on the East Usambaras – statements on climatic and environmental change. In: Hamilton AC, Bensted-Smith R (eds) Forest conservation in the East Usambara Mountains, Tanzania. IUCN, Gland/Cambridge, pp 115–116 Hamilton AC, Bensted-Smith R (1989) Forest conservation in the East Usambara mountains, Tanzania. IUCN, Gland/Cambridge, 392 pp Hamilton AC, Macfayden A (1989) Climatic change on the East Usambaras – evidence from records from meteorological stations. In: Hamilton AC, Bensted-Smith R (eds) Forest conservation in the East Usambara Mountains, Tanzania. IUCN, Gland/Cambridge, pp 103–107 Hewlett JD (1982) Principles of forest hydrology. University of Georgia Press, Athens Hofhansl F, Wanek W, Drage S et al (2012) Controls of hydrochemical fluxes via stemflow in tropical lowland rainforests: effects of meteorology and vegetation characteristics. J Hydrol 452–453:247–258 Kumagai T (2011) Transpiration in forest ecosystems. In: Levia DF, Carlyle-Moses D, Tanaka T (eds) Forest hydrology and biogeochemistry – synthesis of past research and future directions. Springer, Dordrecht/Heidelberg/London/New York, pp 389–406 Kumagai T, Saitoh TM, Sato Y et al (2004) Transpiration, canopy conductance and the decoupling coefficient of a lowland mixed dipterocarp forest in Sarawak, Borneo: dry spell effects. J Hydrol 287:237–251 Kume T, Tanaka N, Kuraji K et al (2011) Ten-year evapotranspiration estimates in a Bornean tropical rainforest. Agric For Meteorol 151:1183–1192 L’vovich MI (1979) World water resources and their future – translation by the American Geophysical Union. LithoCrafters Inc., Chelsea Levia DF, Keim RF, Carlyle-Moses DE, Frost EE (2011) Throughfall and stemflow in wooded ecosystems. In: Levia DF, Carlyle-Moses D, Tanaka T (eds) Forest hydrology and biogeochemistry – synthesis of past research and future directions. Springer, Dordrecht/ Heidelberg/London/New York, pp 425–443 Lin H (2010) Linking principles of soil formation and flow regimes. J Hydrol 393:3–19 Manfroi OJ, Koichiro K, Nobuaki T et al (2004) The stemflow of trees in a Bornean lowland tropical forest. Hydrol Process 18:2455–2474 Manfroi OJ, Kuraji K, Suzuki M et al (2006) Comparison of conventionally observed interception evaporation in a 100-m2 subplot with that estimated in a 4-ha area of the same Bornean lowland tropical forest. J Hydrol 329:329–349 Miyazawa Y, Tateishi M, Komatsu H et al (2014) Tropical tree water use under seasonal waterlogging and drought in central Cambodia. J Hydrol 515:81–89 Moglen GE, Maidment DR (2005) 15: digital elevation model analysis and geographic information systems. In: Anderson MG (ed) Encyclopedia of hydrological sciences. Wiley, Chichester Mun˜oz-Villers LE, McDonnell JJ (2013) Land use change effects on runoff generation in a humid tropical montane cloud forest region. Hydrol Earth Syst Sci 17:3543–3560 NERC (1994) Tanzania urban sector engineering project – yield estimates for Tanga and Morogoro. Report Natural Environment Research Council. Institute of Hydrology, Wallingford Newmark WD (2002) Conserving biodiversity in East African forests: a study of the Eastern Arc Mountains, Ecological studies 155. Springer, Berlin-Heidelberg

Forest Hydrology in the Tropics

1939

Nicholson SE (2000) The nature of rainfall variability over Africa on time scales of decades to millenia. Glob Planet Change 26:137–158 Roberts J, Cabral OMR, Fisch G et al (1993) Transpiration from an Amazonian rainforest calculated from stomatal conductance measurements. Agr For Meteorol 65:175–196 Roberts JM, Tani M, Bruijnzeel LA (2005) Controls on evaporation in lowland tropical rainforest. In: Bonell M, Bruijnzeel LA (eds) Forests, water and people in the humid tropics – Book I. Cambridge University Press, Cambridge, pp 287–313 Schellekens J, Scatena FN, Bruijnzeel LA et al (2004) Stormflow generation in a small rainforest catchment in the Luquillo Experimental Forest, Puerto Rico. Hydrol Process 18:505–530 Schlesinger WH (1997) Biogeochemistry: an analysis of global change, 2nd edn. Academic, San Diego Schultz J (1995) The ecozones of the world/the ecological divisions of the geosphere; with 48 tables. Springer, Heidelberg Shakesby RA, Doerr SH, Walsh RPD (2000) The erosional impact of soil hydrophobicity: current problems and future research directions. J Hydrol 231–232:178–191 Shougrakpam S, Sarkar R, Dutta S (2010) An experimental investigation to characterise soil macroporosity under different land use and land covers of northeast India. J Earth Syst Sci 119:655–674 Shuttleworth WJ, Gash JHC, Lloyd CR et al (1984) Eddy correlation measurements of energy partition for Amazonian forest. Q J Roy Meteor Soc 110:1143–1162 Sidle RC, Tsuboyama Y, Noguchi S, Hosoda I, Fujieda M, Shimizu T (2000) Storm-flow generation in steep forested headwaters: a linked hydrogeomorphic paradigm. Hydrol Process 14:369–385 Tresierra JC (2013) Equitable payments for watershed services: financing conservation and development- case studies on remuneration of positive externalities (RPE)/Payment for Environmental Services (PES). Prepared for multi-stakeholder dialogue FAO, Rome, 12–13 Sept 2013 Weiler M (2001) Mechanisms controlling macropore flow during infiltration: Dye tracer experiments and simulations. Ph.D. thesis, Swiss Federal Institute of Technology, Zurich Yanda PZ, Munishi PKT (2007) Hydrologic and land use/cover change analysis for the Ruvu River (Uluguru) and Sigi River (East Usambara) watersheds. Final report for WWF/CARE Dar es Salaam, Tanzania, 80 p Zalewski M, Janauer GA, Jolankai G (eds) (1997) Ecohydrology. A new paradigm for the sustainable use of aquatic resources. UNESCO IHP technical document in hydrology 7. IHP –V Projects 2.3/2.4, UNESCO, Paris, 60 pp

Range Management A. Swenne

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of Rangeland and Their Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Grasslands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Savannas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shrublands or Steppes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Desert Shrublands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shrub Woodlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Savanna Woodlands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Woodland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Forest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ecology of Rangelands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Ecosystem Concept and Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Functioning of the Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Succession and Climax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interactions Between Plants and Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principal Floristic Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rangeland Productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Physiology of Range Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Phenology of Range Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regulation of Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Morphology of Plants and Grazing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Growth Characteristics of Grasses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Effects of Grazing on Range Plants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Range Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Range Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Range Condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Grazing Management and Stocking Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1942 1943 1944 1945 1947 1948 1949 1949 1950 1950 1950 1950 1952 1954 1960 1962 1965 1966 1966 1967 1968 1968 1969 1971 1971 1975 1977

A. Swenne (*) Brussels, Belgium # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_161

1941

1942

A. Swenne

Range Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vegetation Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fertilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reseeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Provision of Water and Salt for Livestock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Animal Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Numbers and Distribution of Cattle, Sheep and Goats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nutrition and Feeding of Livestock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fodder from Trees and Shrubs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Importance of Fodder Trees and Shrubs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nutritive Value of Browse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Browse Intake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Propagation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Forestry Versus Range Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Range Development in the Tropics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Optimum Stocking Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Land Tenure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Socio-Economic Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Range Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1981 1982 1983 1984 1987 1987 1988 1988 1989 1995 1995 1997 1997 1998 1999 2000 2001 2001 2002 2003 2004 2005 2005 2006 2006

Introduction It is impossible to give precise figures on the areas covered by rangelands. Some indications on land use types can be found in the Production Yearbook (FAO 1989), and percentages calculated on the basis of those data are shown in Table 1. FAO defines permanent meadows and pastures, however, as “land used permanently (5 years or more) for herbaceous crops, either cultivated or growing wild” and forests and woodland as “land under natural or planted stands of trees, whether productive or not, and includes land from which forests have been cleared but that will be reforested in the foreseeable future.” To include all forests in the rangeland areas can be questioned, of course. On the other hand, large tracts of land that do not belong to the above-mentioned categories are in many cases grazed by nomadic flocks and should be considered as rangelands. Williams et al. (1968) estimated that 47 % of the earth’s land surface is rangeland, whereas Holechek et al. (1989) stated that if all the uncultivated land with the potential to support grazing by domestic animals is taken into account, rangelands comprise about 70 % of the world’s land area and are the major type of land found in all continents. What is less questionable is the importance of rangeland to animal production. Rangelands in the USA provide domestic ruminants with between 50 % and 65 % of their total feed needs. In Australia, one-third of the cattle and sheep populations

Range Management

1943

Table 1 Percentages of land area under permanent meadows/pastures and forest/woodland in 1988 (FAO 1989) World Africa North and Central America South America Asia Europe (-USSR) Oceania

Permanent meadows/pastures 24.6 26.8 17.3 27.3 25.3 17.7 52.1

Forests/woodland 31.0 23.1 32.2 51.1 19.6 33.3 18.5

Total 55.6 49.9 49.5 78.4 44.9 51.0 70.6

are supported by rangeland. On a world basis, rangelands contribute about 70 % of the feed needs of domestic ruminants, that percentage being close to 100 % in many parts of the tropics. Rangelands are important to foresters. Very limited areas of rangelands are totally void of woody vegetation. Actually, the density of trees is an important parameter in classifying different types of rangelands. In the wet tropics, there is usually no big problem in defining types of land use with regard to forestry and range. In the dryer regions, and especially in the arid and semi-arid areas, it very often happens that forestry and range development go together, or at least they should. If and when emphasis is put on forestry, there is a clear risk that the animal production linked with range development is totally neglected because of the lack of basic knowledge about improvement and management potentials and practices. This, in turn, leads to a situation where existing resources are not used to their maximum. It may also be at the root of conflicts with local populations whose socio-economic backgrounds are geared more towards animal production.

Types of Rangeland and Their Distribution The numerous combinations of factors affecting vegetation in the world have resulted in correspondingly numerous vegetation combinations typical of each set of conditions. Activities of man and his livestock, such as grazing, cultivation, timber cutting, control of fire, damming and diverting of water, have qualitatively or quantitatively changed almost all of the world’s vegetation. Several classifications of native vegetation have been made by ecologists that are, in part, applicable to range usage (Crowder and Chheda 1982). An international system of vegetation classification was put forward by UNESCO (1973). This retains the terms: forest, woodland, scrub (shrubland and thicket) and grassland. Different types are distinguished within each formation class. For grasslands, this is based on the height of the grass layer, and the density and nature of an eventual tree synusia. According to Stoddart et al. (1975), seven major categories of vegetation provide most of the world’s grazing lands: grasslands, desert shrubs, woodland shrubs, tropical savannas, temperate forests, tropical forests and tundra. Each of these may be

1944

A. Swenne

divided into smaller vegetation types. Taking into consideration the fact that grasses make up an essential component of the vegetational cover of more than half the land surface of the tropics and sub tropics (Rattray 1960) distinguishes six different plant communities according to the density of grass cover: grassland, savanna, steppe, woodland, forest and undifferentiated. Holechek et al. (1989) consider that grassland, desert shrublands, savanna woodlands, forests and tundra are the basic rangeland types of the world. Each of these types comprise several plant associations that support a slightly different biota because of variations in climate, soils and human influences. As there is such a large number of types of rangeland described in the literature, only the most common will be dealt with here.

Grasslands The term “grassland” is normally used in a broader context than the term “pasture.” Both terms are used for vegetation that is predominantly composed of grasses, but includes other legumes and herbs (forbs); most are grazed. Grassland includes barely managed rough grazing, rangelands and natural grasslands, whereas pasture is generally used more for managed agricultural grassland. The distinction between the two words is not clear-cut. Natural grassland, often called “climax grassland,” generally occurred in areas where the growth of trees was prevented or restricted by climatic or soil conditions. Large areas of natural grassland existed where the rainfall was inadequate for forests, e.g. steppes, prairies and pampas, and these graded into areas where rainfall was sufficient to support scattered trees, i.e. savanna. In both of these areas, the presence of trees was probably limited by fire also (Moore 1964). Natural grassland also occurred where temperatures were too low (e.g. high altitude and tundra), or wind velocities too high (e.g. coasts and exposed areas) for trees. Soil conditions also restricted tree growth in some areas that became dominated by grasslands, e.g. on saline soil (salt marshes and saline desert soils) and on frequently waterlogged soils (marshes). Many areas of natural grassland (e.g. salt marsh) occur as small islands in areas of forest. The most extensive semi-natural grasslands of the world are extensions of natural grasslands. Most of the productive agricultural pastures of the world are man-made, originally produced by felling or burning natural forest. Their continued existence depends upon grazing, burning, intermittent ploughing and resowing, the use of selective herbicides etc. The most important factors are grazing and burning. Unlike trees, most grassland species are well adapted to being grazed and burned; they usually recover rapidly and often are more productive and more competitive under moderate grazing or burning (Snaydon 1981). Without this management, most agricultural pastures would revert to scrub and eventually to forest, which is the climax vegetation of those areas (Moore 1964). Man-made grasslands are therefore stages in plant succession, i.e. they are serial communities. These pastures are often termed “deflected climaxes,” “disclimaxes,” or “plagioclimaxes.” Such grasslands are botanically unstable; even small changes in environmental conditions, in grazing management, and stocking rate can lead to large and rapid changes

Range Management

1945

in botanical composition (Davies 1960). For Crowder and Chheda (1982), a typical grassland is an open plain tract of land having a dense cover of tall or short grasses and associated herbaceous species. Shrubs and trees are absent or widely scattered, usually clumped into low-lying moist areas and spread along water courses. The grasses may be endemic or native to a given region, being referred to as “natural” components. They may have migrated into the region under the influence of man or grazing animals and are completely integrated into the natural flora, in which case they are said to be “indigenous” or “semi natural.” Many authors in different parts of the world have proposed subdivisions in order to give a more detailed classification of grasslands in their particular area. In South Africa, dry, Highveld and montane grasslands are differentiated from low to high elevations (Adamson 1938). Highland grasslands are recognized in Kenya between 2200 and 3000 m with a minimum of 1000 mm rainfall and frequent mist (Phillips 1959). Montane open grasslands occur in Ethiopia at about 2500 m and upward, under 750–1275 mm annual rainfall (Keay 1959). The same types are described as cool, mountain grasslands in Latin America (Roseveare 1948). Wet and dry grasslands have been identified in Sri Lanka (formerly known as Ceylon) (Holmes 1946). These two have been subdivided into arid, sub arid, mild arid, sub humid and humid in South Africa (Phillips 1959). Tussock and hummock grasslands occupy parts of the steppe and sub desert regions of Australia (Wood and Williams 1960). In Papua New Guinea, grasslands are identified on the basis of plant height (Heyligers 1965). In many regions open grassland is replaced by grass-woody plant associations. A proposal was made in East Africa to use canopy cover as a criterion in measuring the contribution of trees and shrubs to grazing lands (Pratt et al. 1966). Areas dominated by grasses, but with widely scattered or grouped trees and shrubs having a canopy cover no greater than 2 %, were called grasslands. Bushed and wooded grasslands had scattered or grouped shrubs or trees, respectively, with less than 20 % canopy cover. Areas of arid or infertile land sparsely covered by grasses and dwarf shrubs not exceeding 1 m in height, but sometimes with widely scattered larger shrubs or stunted trees, were called dwarf-shrub grassland. Subtypes suggested for these categories were manifold in describing associations on a local level (Crowder and Chheda 1982). Grasslands are a minority type but are productive grazing lands where they occur. Dry land cultivation of large areas in unsuitable climates has been attempted, resulting in serious weed invasion. In the USA, Argentina and Africa, the grasslands are mainly a fire-induced disclimax, and shrub invasion follows overgrazing or other forms of fire prevention. In Australia, grasslands grow on a cracking clay unsuitable for tree growth (Snaydon 1981).

Savannas The majority of the grasses present in savannas are hemicryptophytes, and the annual dormant period occurs during the dry season. The woody layer is variable, and it is thus possible to recognize a number of facies: grass savanna (without any

1946

A. Swenne

woody layer), shrub savanna, tree savanna and savanna woodland. A formation very similar to savanna woodland is woodland that has a denser and taller tree layer (covering over 50 % of the area) and a reduced grass cover, which may still be continuous enough to permit the passage of fire. It is also necessary to describe the transition between savannas and forests as well as the dynamic relationship between these two formations. It is generally accepted that, apart from certain special soil conditions, the climate of savanna regions is normally that where forest could occur, and savannas constitute a non-natural state maintained by fire. Thus, savannas are the result of the degradation of former extinct dry ecosystems (dry woodlands, etc.), or are the result of the destruction of dense humid forest and its replacement by a savanna flora (periforest savannas). In these areas bordering the forest in humid climates, the ligneous flora of the savanna is often poor and reduced and is always very different from that of the dense humid forest. In recently deforested areas in particular, savannas are totally herbaceous formations with species such as Pennisetum purpureum or Panicum maximum in Africa and Imperata cylindrica in many areas. These can be considered as “forest fallows” that could, given the circumstances, return to forest or evolve to true savannas. The forest/savanna boundary is clear-cut and maintained by fire. There is no transitional formation but a mixed landscape occurs formed of a mosaic of areas of forest and savanna (and often secondary formations and fallows). In many regions, especially in Africa, the evidence of former dry forest formations has almost disappeared because of the length of time of human intervention. Thus, two-thirds of the African forest has disappeared in several centuries. The very rapid disappearance of tropophilous forests is perhaps due to the frequency and intensity of fires in savanna regions with a marked dry season. Elsewhere, as in Brazil for example, the relicts of these forests occupy large areas (cerrado). The Brazilian campos cerrados derive from the degradation of these forests, and they have the same ligneous flora. Although the Brazilian cerrados often have a particular physiognomic structure because of the presence of a very dense woody layer, they can be considered as true shrub savannas that have not been subjected to a strong degradation by fire or by overgrazing as in Africa. In Asia (India, Malaysia and Southeast Asia in particular) the existence of savanna ecosystems is, in the majority of cases, a direct result of short- or long-term degradation of various forest formations. In regions having very old civilizations, such as India, true annually burnt savannas occur only on mountains and plateau that are relatively sparsely populated (UNESCO 1979). The tropical savanna grades into a forest on its more humid boundary and into grassland on its more arid boundary. The climate of savannas is largely monsoonal with distinct wet and dry seasons. Plants make rapid growth at the onset of rains with lush vegetation occurring within days of the first monsoonal rain. During the dry season, only the woody plants escape dormancy, and even some of them shed their leaves. The largest savanna type is the Acacia-tall-savanna of Africa. It stretches in a broad belt from east to west across the center of Africa between the desert scrub to the north and the tropical forests to the south. Similar

Range Management

1947

savannas occur in eastern and southern Africa. Acacia and Combretum are the main tree genera. Herbaceous vegetation, mainly grasses, provides a dense ground cover and most of the annual dry -matter production. Stoddart et al. (1975) concur with many authors that savannas are difficult to identify because many areas between forests and grasslands exhibit the savanna structure. Due to external factors, changes may occur. On the humid boundaries, increased grazing intensity and fire control may lead to woodland encroachment. On the dry boundaries, desert encroaches following drought, heavy grazing and removal of trees for charcoal. The descriptions of the many types of savannas have led to a complexity and synonymy of terms and nomenclature in attempts to separate categories and classifications of plant formations. The distinction between savanna and steppe has not been clearly defined since the classification of both is based on the nature of the herbaceous layer and on the density of the woody vegetation (Cole 1963).

Shrublands or Steppes The term “steppe” was originally used for open and treeless plains of xerophilous vegetation in Russia and Asia, i.e. short, wiry, tufted perennial grasses that developed under cool, temperate, low-rainfall or arid conditions. A warm steppe exists in tropical and sub-tropical regions where xerophytic plants developed under low-rainfall or arid conditions. Steppe vegetation was described in Africa (Keay 1959; Rattray 1960) and Australia (Wood and Williams 1960) but has less common usage in South America and Southeast Asia. The warm steppe refers to vegetation cover of low-growing trees or shrubs with widely spaced annual or perennial grasses, as well as treeless grass-herb sub-desert formations. Steppe formations occur in areas with a rainfall of less than 500–600 mm/year and a dry season of 8–9 months. This implies a fundamental difference from savannas in the ways of utilization: in the savannas, forage is largely from grasses grown during the wet season and also from the smaller amount of regrowth in the dry season after burning; in the steppe, all the forage is provided during the brief wet season. Steppe can be defined according to the Yangambi classification (CSA 1956) as: “open herbaceous formations, sometimes mixed with woody plants; generally not subjected to fire; perennial Gramineae with a widely spaced distribution, generally not exceeding 80 cm high, with narrow, rolled or folded leaves, originating mainly from the base; annual plants are often abundant between the perennial plants.” Thus steppe can be distinguished from savanna by the less dense grass layer as well as by the xeromorphic characters and annual nature of many species. Woody plants are often thorny or succulent and belong to different floristic groups from those of the savannas. Tropical steppes may be considered as derived from degraded scrub such as the Brazilian caatinga or the thickets of southwestern Madagascar. They cover large areas of sandy and saline soils in northwestern India and Pakistan. Different types may be recognized according to the characters of the woody layer:

1948

A. Swenne

– grass steppe (hemicryptophyte steppe); – dwarf-shrub steppe (chamaephyte steppe); – succulent steppe (Cactaceae, Euphorbiaceae; halophilic plants, e.g. Salsolaceae); – tree or shrub steppe (nanophanerophyte or phanerophyte steppe, e.g. Acacia spp. in Africa and Eucalyptus spp. in Australia) It is appropriate to mention an intermediate type of vegetation that occurs as an almost totally herbaceous formation dominated by perennial grasses belonging principally to the genus Aristida. These have the morphology of steppe plants with narrow basal leaves that are strongly xeromorphic. This type of formation has the appearance of a steppe, but it can be found in regions with a relatively long wet season, i.e. in savanna climate, but where edaphic conditions are particularly unfavorable. They are of low pastoral value and cover considerable areas, for example in west central Africa (over the Kalahari sands), on the high plateau of Madagascar and in the Campos Limpos of central southern Brazil. The term pseudo-steppe was proposed to describe them. These formations are the result of excessive overgrazing in peninsular India (UNESCO 1979). In the steppe climates of Australia, bunch grasses predominate and are usually associated with scrub trees on thorn bushes and annuals (Moore 1964). A cool grass steppe described in the eastern region of La Pampa in Argentina is an extension of the humid pampa to the north and a transitional zone abutting on a xerophilous region of woodland to the west.

Desert Shrublands Desert shrublands are the driest of the world’s rangelands and cover the largest area. Woody plants less than 2 m in height with a sparse herbaceous understorey characterize vegetation of this type. Desert shrublands have received the greatest degradation by heavy grazing of the rangelands’ biomes and show the slowest recovery from degradation. In some cases, desert shrublands have been created by degradation of arid grasslands by heavy livestock grazing. Desert shrublands generally receive less than 250 mm of annual precipitation. The amount of precipitation varies much more from year to year than in the other biomes. In hot desert shrubland areas, precipitation occurs as infrequent, high-intensity rains during a short period (less than 90 days) of the year. This results in long periods where the water content of the soil surface is below the permanent wilting point. This provides highly unfavorable conditions for short, fibrous-rooted plants (grasses). Coarserooted plants (shrubs) can collect moisture from a much greater portion of the soil profile than those with short, fibrous roots near the soil surface. Desert shrub roots extend considerable distances laterally as well as downward. The sparse spacing of desert shrubs permits individual plants to collect moisture over a large area. This explains why they can survive long, dry periods much better than grasses. The shrubs are important as a stabilizing influence both environmentally and as a fodder reserve. When in good condition, the main sources of fodder are the

Range Management

1949

associated perennial grasses although the extremely arid areas can support only ephemeral or annual herbs. Flexible grazing systems are essential in the desert shrublands to avoid overuse during the frequent drought. Overuse of the pastures first removes the perennial herb component and annual or ephemeral herbs become dominant. Grazing use of the shrubs increases, and the more palatable shrub species die out, overland waterflow and soil erosion increase and a more xeric vegetation establishes itself, which is dominated by succulent species (Harrington 1981).

Shrub Woodlands Shrub woodlands usually occur in about the same annual rainfall belt as grasslands, but low-growing trees (usually less than 10 m) and dense shrubs are the dominant vegetation. Some of the main reasons for the dominance of trees are: lack of periodic fire, poor rainfall distribution during climatic cycles, shallow and rocky soils, and heavy use by grass-eating animals. Many of these woodlands have a high potential for range improvement. The tropical woodlands are found where the forests are flanked by dry regions. In Africa, they may give way to tropical savannas in more favorable situations or where fire has held the woodland in check. Thorn forests occur on the Indian subcontinent in the drier monsoonal area. Tropical woodlands occur in the northern portion of Australia where a summer rainy period produces 420–750 mm of rain followed by a dry period of 3–9 months. Woody species are mainly Eucalyptus with a mixture of other shrubs such as Acacia, Hakea and Terminalia (Stoddart et al. 1975).

Savanna Woodlands Savanna woodlands are dominated by scattered, low-growing trees (less than 12 m tall). They have a productive herbaceous understorey if not excessively grazed. Heavy grazing usually results in loss of understorey grasses and an increase in the density of the trees and shrubs. Typically, savanna woodlands occur as a transition zone between grassland and forest. Shifts towards grassland or forest take place continually in this biome, depending on grazing intensity, fire control, logging, and drought. Shrub and tree densities on many savanna woodlands have increased substantially because of fire suppression and heavy grazing of the understorey. Rocky, thin soils favor woodlands in grassland climatic zones because the long, coarse roots of woody plants can grow down into cracks in the rocky layer where moisture is collected. Further, many woody species have long lateral roots that can absorb moisture over a large area of very thin, rocky soil. Without periodic fire, most of the wetter portion of the tall grass type with loamy to sandy soils is quickly invaded by trees and shrubs because considerable moisture reaches that portion of the soil profile below 2 m (Holechek et al. 1989).

1950

A. Swenne

Woodland The woodlands with which grasses are associated are open forests with deciduous and semi-deciduous trees, often having their crowns touching, and with a sparse undercover of tall grasses that thicken when the trees are removed. In areas of high rainfall, grasses may be absent because of the closed tree canopy. Burning is usually practiced to remove the old accumulated material and to maintain an open woody plant formation, otherwise the derived grassland reverts to woodland (Crowder and Chheda 1982).

Forest Tropical forest of dense, complex mixtures of trees, vines and epiphytes are found in tropical areas of all continents. Those areas are of little importance for grazing. Hot, wet conditions endanger the health of domestic animals, and diseases and parasites are common. Little herbaceous vegetation is produced under dense canopy, and it is usually of poor forage quality. Where commercial cattle operations exist in the tropical-forest type, they are dependent upon improved pastures from which the forest has been removed and grass established (Stoddart et al. 1975). According to Holechek et al. (1989), forests are distinguished from savanna woodlands by having trees over 12 m in height that are closely spaced (less than 10 m apart). In many areas, forests are managed primarily for timber production and are too dense to have any grazing value. They can produce considerable forage for both livestock and wildlife, however, when thinned by logging or fire or when in open stands. A large percentage of the present grasslands developed from forests and are considered fire subclimaxes (Crowder and Chheda 1982).

Ecology of Rangelands The Ecosystem Concept and Components An ecosystem is a “functional unit consisting of organisms (including man) and environmental variables of a specific area” (Van Dyne 1966). It contains living and non-living elements and there is an exchange of energy and matter among these elements or components (Lewis 1969). These components are the abiotic (non-living) factors, primary producers, consumers, and decomposers. The living and non-living elements comprising a piece of rangeland on which man has placed boundaries for management purposes is referred to as a rangeland ecosystem (Holechek et al. 1989). Each component affects, and is affected by, each of the other components. The magnitudes of the various actions and reactions differ and are indicated broadly by the thickness of the arrows in Fig. 1. The grazing animal is a part of the plant’s environment and the plant a part of the animals. So long as the two live together, the welfare of each is dependent upon the

Range Management

1951

Fig. 1 A diagrammatic representation of the interactions that occur between the components of a grazed pasture ecosystem (Snaydon 1981)

other. This concept is fundamental in range management. Never can the forage and the animal be considered separately. Each of these must be looked upon as part of a great and intricately related biological complex. The abiotic factors form the setting or environment in which the biotic factors operate. With the exception of fire, the manager has little control over them. They enter into his decisions in determining the suitability of a site for various uses, but they are not easily manipulated. The biotic factors can be controlled more easily. Their manipulation is the basic tool used in determining productivity and usefulness to man. Four basic functions are performed by organisms in the biotic portions of an ecosystem. Producer organisms are plants that capture the energy from the sun. They are the only major agent for converting the sun’s energy into food for animals. The number of livestock or wildlife that a particular range can support depends upon the plants ability to synthesize food by fixing light energy through photosynthesis. Consumer organisms are animals that eat, rearrange and distribute the energy captured by plants. Primary consumers are herbivores that live directly off plants such as some insects, livestock and large ungulates. Secondary consumers are those animals that eat herbivores, i.e. carnivores (Stoddart et al. 1975). Decomposers function primarily in the decomposition process and are responsible for preventing accumulations of organic matter. Without decomposers, ecosystem functioning would not be possible because there could be no nutrient cycling, and elements would eventually be tied up in undecomposed organic material. The decomposer microorganisms are generally bacteria and fungi as well as actinomycetes, algae, and lichens, which possess the enzyme systems necessary to break down resistant organic materials. Several other groups of organisms also function as microconsumers and are active in detrital food chains (Paris 1969). Manipulators are organisms that deliberately rearrange the factors of the ecosystem for their own benefit. Man is the master manipulator. The abiotic components consist mainly of the soil and climatic factors and are not usually manipulated by the range manager. Grazing animals, pasture plants and microorganisms form the biotic components of the ecosystem. All three of these components interact with one another. For example, grazing animals affect the pasture through defoliation, through return of nutrients and through treading. Conversely, the pasture affects grazing animals through the amount of feed available, the seasonal pattern of production, and

1952

A. Swenne

Fig. 2 Diagrammatic representation of energy flow through successive trophic levels of an ecosystem (Stoddart 1975)

through pasture quality. Microorganisms interact, directly and indirectly, with both pasture plants and grazing animals; microbial pathogens attack both plants and animals, while microbial symbionts (e.g. rumen microflora, N-fixing bacteria, and mycorrhiza) stimulate the growth of animals and plants, and microbial saprophytes decompose plant and animal remains, thus releasing nutrients. In addition to these two component interactions, there are also three-component interactions. For example, pasture plants, grazing animals and soils interact in the cycling of mineral nutrients and so do pasture plants, saprophytic micro-organisms and soils. Similarly four-component and five-component interactions can be envisaged.

Functioning of the Ecosystem Range ecosystem function can be viewed mainly from two standpoints: energy flows and chemical cycles. These really represent physiological processes within the ecosystem. Energy flows throughout the ecosystem and operates under the first law of thermodynamics (Holechek et al. 1989). Energy Flow. Interactions between plants, animals and the environment are considered most simply by investigating the fate of individual elements (or compounds) and energy within pasture ecosystems. The various nutrient elements (N, P, K, and S etc.), carbon and water are all cycled in ecosystems. In contrast, energy flows through ecosystems; it enters as solar radiation, is progressively dissipated as it passes through the ecosystem (Fig. 2) and is finally lost by re-radiation into outer space (Snaydon 1981).

Range Management

1953

Studies have shown that less than 1 % of the usable solar radiation received by plants in range ecosystems is utilized in photosynthesis (Sims and Singh 1978). Snaydon (1981) calculated the efficiency of each trophic level in a grazed pasture ecosystem in the humid temperate region: – efficiency of the pasture (producer level), 0.7 %, – efficiency of use of the photosynthetically active solar energy, 1.4 %, – efficiency of the grazing animal (first consumer level): relative to energy content of food consumed, 10 %, – relative to energy content of herbage produced, 2.5 %, – relative to total solar energy receipt, 0.017 %. We see that the energetic efficiency of both herbage production and animal production is very low. Herbage production, and hence the energetic efficiency of herbage production, can be increased by additional inputs, such as fertilizers and irrigation, and perhaps also by grazing management and by using improved species and cultivars. Animal production (and so energetic efficiency) can be increased by increasing herbage intake, e.g. by improving nutritional value or by increased stocking rate, if it is suboptimal, or by using breeds that convert feed more efficiently (Snaydon 1981). Communities in climax condition are more stable than those of lower stages. The maximum diversity of the high stages of succession offers maximum resiliency and freedom from degradation. Thus it may be desirable to manage a critical watershed for climax conditions even though forage production may not be maximized. The range manager’s job is to minimize energy and nutrient drain on the ecosystem and maximize system health. One of the major ways to insure ecosystem health is through manipulating the producer level of the ecosystem (Stoddart et al. 1975). Chemical Cycling. The second basic functional process of range ecosystems is chemical cycling. Unlike energy, chemical elements cycle through the various compartments and can be reused (Fig. 3). The source of many elements, except for nitrogen, is the soil parent material. In many cases, the soil acts as a sink or reservoir for chemical elements. Nitrogen Cycling. Although there is a great abundance of nitrogen in the atmosphere, most plants and all grazing animals cannot metabolize elemental nitrogen directly. It must be “fixed” or transformed by “free-living” microorganisms or those living in symbiotic relationship with certain plants in nodules on their roots. These organisms convert atmospheric nitrogen into forms that can be used by the plants. In natural grassland systems, the input of nitrogen by nitrogenfixing micro-organisms is reduced by competitive interactions with the many other types of organisms present (Clark and Paul 1970). Thus the growth of pastures and grazing animals is frequently limited more by the availability of nitrogen than by that of any other essential nutrient (Henzell and Norris 1962). Nitrogen input and flow through the different components of the pasture system are controlled by interacting processes that create a complex ecosystem (Fig. 4).

1954

A. Swenne

Fig. 3 Relations of inputs, stores, and products from a rangeland ecosystem (Stoddart et al. 1975) (Adapted from Perry 1970)

Succession and Climax Definition Succession is the orderly process of community change. It is the process whereby one association of species replaces another, until the final community is reached. This final, somewhat stable community, is often called the climax. Such a succession usually is gradual and involves a series of changes that follow a more or less regular course. Succession results from a change in habitat and invasion of new species. Change of environment or habitat results in change of the plant cover adapted to the area. Change in habitat sometimes results from action of plants upon soil and microclimate. Thus the plants themselves may initiate the change that ultimately will result in succession and their own destruction. Succession may be either natural or induced. Natural succession takes place until climax conditions are reached. It results from soil changes in the process of soil succession. Also, both before and after climax is reached, advancement or recession of the process may result from fluctuation in the habitat. Distinguishing between natural and induced succession is important. Induced succession results generally from the action of man, and hence is not a condition imposed by nature. As such, it can be modified by man more readily than can natural succession. Succession not only involves a change in species composition but also a change in plant abundance (Stoddart et al. 1975). In some cases, succession may be looked at simply as a re-arrangement of species that were present during initial stages, perhaps only as

Range Management

1955

Fig. 4 A flow chart of major nitrogen pathways in animal production from a grazed pasture (Simpson and Stobbs 1981)

seeds or other propagules. The proportion of the various species changes during succession, but different groups of species may dominate different stages for different time periods. These shifts or rearrangement of species over time emphasize the initial floristic composition theory of succession (Egler 1954). Another type of succession involves immigration of new species to the site from other sites with time progression. Different groups of species dominate the site for various periods. Egler (1954) used the term “relay floristics” to describe this type of successional change. Primary succession starts from bare areas and proceeds to the development of a somewhat stable climax vegetation. Such changes require extremely long periods, on the scale of hundreds or even thousands of years. Consequently, primary successions may be of interest, but they play a small role in range management. It is important to recognize that other ecosystem components undergo succession as well as vegetation.

Primary Succession The term primary succession generally is applied to natural plant succession on previously un-vegetated areas leading to a climax. Large areas of rangeland originated from deposits of wind- or water-moved soil. Under such conditions, soil formation and modification were not extensive. Other rangelands originated from dry rock surfaces (Stoddart et al. 1975). On bare ground, there may be microsites

1956

A. Swenne

where lichens, algae or moss are supported. As the rock is weathered, and water and organic matter are added from the lichens or algae, a rudimentary soil is formed. Seeds from nearby plants may be available to germinate and to support vascular plants. These are often annuals that can survive under harsh conditions. With further weathering and soil formation, some perennial plants may become established. These are generally herbaceous plants, but eventually if the climate will support them, woody plants will become established. Each assemblage of plants influences soil and microclimate, sometimes making it more suitable for plants that need more water, nutrients, and so on. Thus, some plant species alter the environment such that it is no longer suitable for them, and they are replaced by other plants. Total biomass (plants and animals), total energy storage, diversity and rate of mineral cycling increase as succession proceeds (Lewis 1969). Holechek et al. (1989) summarize the processes of primary succession as follows: – The development of soil from parent materials. – Increasing longevity with successional advance. – Replacement of species with broad ecological requirements by those occupying narrow niches complementary with other species. – Greater accumulation of living tissue and litter per unit area with successional advance. – Modification of micro-environment extremes. – Change in size of plants from small to large. – Increase in the number of pathways of energy flow. – More nutrients tied up in living and dead organic matter. – Greater resistance to fluctuation in the controlling factors.

Secondary Succession Secondary succession refers to a succession, usually induced, on land previously occupied by more highly developed vegetation destroyed by unusual circumstances, such as fire. Range managers routinely deal with secondary succession, but rarely with primary succession; however, sometimes erosion does change the initial soil surface conditions. Generally, we are concerned mainly with vegetational changes in secondary succession and how these changes influence habitats for other organisms. Secondary succession generally occurs much faster than primary succession, and generally, in a more predictable fashion. The variability in secondary succession is reduced as the climax is approached (Huschle and Hironaka 1980). Climax The final stage of succession is the climax. The climax has been viewed in different ways by different authors. Some have considered that the climax is “stable,” others that it is in “dynamic equilibrium” with the environment. Clements (1916) viewed the climax as controlled primarily by the macroclimate. This he referred to as the “climatic climax.” He viewed large areas of landscape as having the same climax governed by the climate. Development of climax vegetation was considered a very slow process on the same time scale as geologic changes.

Range Management

1957

Tropical Rangelands and Climax Vegetation The herbaceous climax formations in tropical regions are most frequently edaphically determined. Most grazing land ecosystems have been transformed by annual fires or by overgrazing, with changes in the floristic composition leading to a reduction in their fodder value. Nevertheless, some ecosystems seem to be in balance with present-day environmental conditions, and they are described as a pseudo-climax, periclimax or disclimax, according to the author. Grazing land formations closest to climax vegetation are: • The tree or shrub steppes of arid and semi-arid regions having a short period of vegetative growth and a high proportion of annual grasses and being difficult for fire to sweep through. Their limiting factor is the low annual rainfall concentrated in a few months. Physiological and phenological adaptation of the species to the arid conditions also con tributes towards maintaining their equilibrium. These are the areas of nomadic pastoralism. They are very sensitive to all types of abuse: overgrazing around watering points or villages; and removal of woody material. • Edaphic savannas or grasslands on marshy or temporarily flooded soils. These formations are common in humid areas but also exist under dry climates in interior deltas (e.g. Niger and Senegal), around large lakes (Chad) and along major watercourses (UNESCO 1979). The climax in the tropical humid zone is forest. The majority of authors recognize the anthropogenic origin of grazing land ecosystems. Grass, shrub or tree savannas play a large role in tropical montane areas (e.g. Mt Cameroon, mountains of southern India and Sri Lanka). Their ever increasing extent, at the expense of the forest, is due to repeated burning, even in areas with low population density. The ecological conditions of this zone (large daily variation in temperature and available moisture) make forest formations vulnerable and their reestablishment in savanna areas more difficult (Blasco 1970).

Plant Retrogression Any of a great number of actions may disturb the climax plant cover and bring about retrogression leading away from the climax. Retrogression may be caused by drought, fire, or grazing. If this action is temporary, a succession leading back to climax follows. Causes of Retrogression. By far the most important of the factors bringing about retrogression on range is improper grazing. The retrogression of a plant cover under grazing does not follow in the reverse order to the succession that gave rise to it, because the retrogression is usually of vegetation and not of soil. Since the climax soil is damaged less easily, it is more permanent than the vegetation, and its retrogression lags far behind. The stages of grazing retrogression in vegetation, then, are determined not by climate or soil, but by the newly introduced biotic factor, usually livestock. Unfortunately, grazing continues to weaken the soilprotecting vegetation, and soil deterioration also occurs. Water or wind may

1958

A. Swenne

move away the developed surface soil to the point that exposed sub-soil is incapable of again supporting climax plants. Succession to the climax, therefore, must again await development of a new soil mantle. Especially in arid areas, soil formation is a very slow process involving hundreds, or even thousands, of years. Soil retrogression caused by erosion and trampling may progress so far that vegetation may be held in a sub-climax stage, even though grazing has ceased entirely. Absence of a rapid secondary succession following good management often confuses the range manager, since vegetation cannot respond to improved grazing conditions as he expects. Retrogression of vegetation under grazing may follow a multitude of courses, depending on vegetation and type of grazing. Grazing during a restricted season may harm only certain species, whereas others may benefit because of reduced competition. If a short grazing season results in use of a certain species during a critical growth stage, that species may disappear. Similarly, because of foragepreference differences among kinds of livestock, grasses may increase on a sheep range at the expense of forbs and brush; conversely, on cattle ranges, grass may disappear. Too intensive grazing is marked by a disappearance of the preferred plants or of those physiologically less resistant to grazing. Retrogression thus involves plant competition. The removal of climax plants by abuse beyond their endurance leaves space for other plants. Less preferred or more resistant plants may survive and replace the removed plants. These species are sometimes referred to as increasers, because they increase under heavy grazing. Continued grazing will cause an influx of species, often annual, which are not part of the climax. These are called invaders. The most preferred climax plants, under stress of heavy grazing, lose vigor, as evidenced by reduction in annual growth; reduction or complete absence of reproduction activity; and, in woody plants, abnormality of growth induced by removal of the growing tip and excess stimulation of lateral buds. Continued physical disturbance of the preferred plants results in their death. Death and disappearance may result from starvation following reduced photosynthesis, competition from other plants less weakened by grazing, natural old age accompanied by lack of reproduction, or drought made more serious by a weakened root system. Composition change on the range usually is gradual, marked first by a decrease in the most preferred plants, and the plants physiologically and anatomically most susceptible to grazing injury (decreasers). Accompanying the decrease in numbers is a decrease in competition, which results in an increase of less preferred or more resistant individuals (increasers). Animals change their diet, because of increasing shortage of desirable species, to those less preferred. Thus, succession continues, with better climax plants progressively becoming fewer. Following, or simultaneous with, these composition changes comes the invasion of new species. The first invaders are mobile annuals; later, the invasion of herbaceous or woody perennials of low grazing value often takes place. The annual invaders may be highly preferred by stock for a short season, but they often are adapted to thrive despite grazing. Most invading perennials are not highly preferred by stock. Climax plants ultimately may disappear: They leave first from the most

Range Management

1959

accessible and, hence, most grazed areas, and soon are evident only under the protection of a stout shrub or thorny cactus. Continued heavy grazing forces stock to consume the invading species, which suffer as did the climax species. The most preferred and most susceptible species are removed first, and the less valuable temporarily increase in numbers. As grazing continues, these may bear the brunt of the grazing, and their numbers will decrease. If these are not followed by new invaders such as shrubs of low palatability, the land approaches a barren state, with soil deteriorating rapidly. Secondary Succession Following Retrogression. The secondary succession following improved grazing conditions usually differs from primary plant succession since good soil conditions may remain. Frequently, however, soil retrogression follows plant retrogression, because of erosion and trampling. In such cases, secondary succession may be almost as slow as primary succession, and may follow in very similar steps towards the climax. When soil has not deteriorated along with vegetation, succession of vegetation, upon removal of grazing stress, may be very rapid, especially in areas of high precipitation. It is especially rapid if climax plants have been removed. Practical management may maintain climax cover but, often, a vegetation cover lower than climax proves most practical. It is impossible to obtain the best use of a range without some disturbance, and the range manager cannot always have climax vegetation as his goal. When maintenance of maximum soil protection is wanted, it may be desirable to manage for near climax conditions and capitalize on the diversity of higher stages. Changes in plant composition may not result in reduced plant cover. Cover may actually increase. Forage Value of Invading Species. The preference that an animal displays for a plant is not an accurate index to its value for grazing. Animals can be forced to eat almost any plant, and some of the less liked species are as nutritious as are the preferred ones, and animals sometimes do as well on them. A slight decrease in palatability on the plant cover after excess grazing may, in itself, be no indication of reduced value. Usually, however, invasion of less preferred species is accompanied by marked reduction in grazing capacity, independent of volume yield. This, is probably attributable to the fact that animals eat less of feed they do not like rather than to nutritional difference. Nutritional studies have failed to show consistent differences between climax and invading species, except that invading ephemerals are likely to become dry earlier and to deteriorate more upon drying than are long-lived perennials. In some cases, particularly in the leached soil of monsoonal tropics, the climax vegetation is low in nutritive value during the dry season. The introduction of weedy annuals results in improved forage quality during the dry season. Unfortunately, most rangeland climates are hazardous for establishment of young plants and, since this process is a frequent necessity for short-lived species, their dependability is greatly reduced. Annual plants depend upon a favorable period each spring during which they can germinate and send their roots to the moist subsoils. Such a period is far from a surety over most ranges, and, hence, failures are common. Fluctuation in forage volume from year to year is much greater on ranges high in annual plants. Annuals are, likewise, most variable in their season of growth. Perennials, having deep roots already

1960

A. Swenne

established, are less dependent upon current precipitation and more upon temperature, which is less variable, for their start of growth. Annuals are dependent upon precipitation for initiation of growth and may reach their period of productivity at vastly different dates from 1 year to another. Annuals are short lived and are best grazed during their green period, often only 6–8 weeks, which may not fit well with the management scheme. Most poisonous plants are low in palatability, hence increase upon heavy grazing is inevitable. Retrogression following misuse is the greatest single factor contributing to poisoning. Gradual invasion of low quality species and decline of good forage, indeed even serious decrease in total production, may escape notice of the range manager. This decrease in forage ultimately forces animals to eat plants that normally remain untouched (Stoddart et al. 1975).

Interactions Between Plants and Animals Ecosystem Stability and Grazing Ecosystems may develop some degree of stability over time with certain levels of herbivores present. These herbivore levels may fluctuate and at times may be destructive of the vegetation. Hence climax equilibria must encompass considerable variation in producer and consumer organisms. In many cases, livestock numbers were in excess of the capacity of the resources to support them. In these cases, induced regression occurred, which resulted in less primary production, accelerated erosion, and so on. The impacts of grazing by domestic livestock are varied. It is extremely difficult to generalize because of differences in climate, resistance of different species to grazing, stocking levels, composition of vegetation, grazing season, and many other factors. In some cases, shifts in species composition may be minor, whereas in other cases, they may represent changes in life forms (Holechek et al. 1989). Plant communities change in an orderly way when grazed by a particular kind of animal. Those plants most preferred by the animals are the first to show signs of grazing stress. They lose vigor, little annual growth is produced, and reproduction is almost absent. As grazing is continued, these palatable plants, (or decreasers) die. With the death of the more palatable plants, the less palatable members of the plant community (increasers) increase in abundance, resulting in a change in community composition. As the community progresses toward the less palatable plants, the grazing animals must change their food habits, move to new areas, or die. In turn, populations of new animals may develop that prefer the dominant species of the altered community. In that case, the increaser plants may become decreasers under the grazing pressures of the new herbivores. Range vegetation, then, is being influenced constantly by the kind and number of animals present. Indeed, the native rangelands of the world are the result of different types of grazing pressures. In Africa, where more than 20 species of large herbivores raze the same range, niche segregation among different species is highly developed. Talbot and Talbot (1963) reported that in Kenya, there are more

Range Management

1961

than 70 species of grasses for wildebeests (Connochaetes); however, 5 plant species comprised from two-thirds to three-fourths of the diet and 10 species made up 90 % of the diet. Talbot (1962) reported that the diets of animals grazing the East African rangelands were complementary and non-duplicating. Most of the large species ate different plants, but where they did eat the same plants, they ate different portions of the same species. When the preferred forage species of the highly selective African ungulates are utilized, the animals move to new areas rather than eat forage not normally eaten.

Organic Reserves and Grazing Organic reserves are reserve compounds that are elaborated by the plant from the simple sugars produced by photosynthesis. They are then stored, either passively or actively, and ultimately utilized by the plant at some later date for maintenance or growth. The usual organs of storage are plant roots, rhizomes, stolons and tiller bases. In pasture plants, the most important reserve compounds are the carbohydrates, i.e. sugars, fructosans and starches. Davidson and Milthorpe (1965a, b) have pointed out that these reserve carbohydrates are in equilibrium with a pool of labile structural and nitrogenous compounds. In general, the evidence strongly suggests that organic reserve compounds, both protein and carbohydrate, are utilized during the first week of regrowth and that later regrowth is dependent on leaf area and photosynthesis. From an animal production point of view, it is necessary to consider the influence of the carbohydrate reserves on ruminant growth and nutrition. As pointed out by Blaser et al. (1966), forages can be low in utilizable energy, and fertilizer nitrogen applied to increase the yield may result in an additional lowering of utilizable energy per unit weight of forage. On the other hand, imposition of grazing managements (for example lax and infrequent grazing) to produce high carbohydrate reserves will be accompanied by reduced dry matter yield. Thus, better quality of forage to improve animal nutrition may have to be balanced against reduced pasture and animal production per unit area. Root Reserves Although the grazing animal is only capable of utilizing the above-ground herbage, the magnitude of the root reserve pool will influence the regrowth after a grazing or defoliation. Further, as grazing affects the development of new root mass, it directly influences the capacity of the pasture to withstand periods of soil moisture stress and to compete for soil nutrients, in particular soil horizons. In a dynamic situation, the plant attempts to maintain a fixed relationship between its chemical composition and weight, but as environment forces it to move away from this position, it can only do so by altering its root-to-shoot ratio. Thus, an increase in photosynthesis initially causes an increase in root weight, which ultimately results in more mineral absorption and further shoot growth. Following defoliation, the organic reserves in the root system are utilized to restore the absorptive and photosynthetic capacities of the whole plant, reducing root weights.

1962

A. Swenne

Species Composition Livestock numbers exert a powerful influence on the vegetation of intensively managed pastures. Vegetational changes in pastures occur as a result of the action of one or all four possible mechanisms. Plants may possess attributes such as low palatability or accessibility, which prevent removal by grazing (Kydd 1966). Animals differ in their grazing behavior with some species exercising greater dietary selection than others. Botanical composition can be influenced by the level or spatial distribution of plant nutrients (Hilder 1964; Rossiter 1966; Wolfe and Lazenby 1973a, b). Finally, climate and the physical environment exert powerful and interactive effects with the previous three mechanisms. The value for animal production of a particular botanical change will be small in the short term in many cases. When the grazing pressure is low, and the pasture supply is plentiful, then grazing animals are able to select forage of much higher quality than the average on offer. In the long term, where persistence of the vegetation type is considered, it is obviously necessary to maintain a particular species composition in order to sustain a particular animal production system. In grasslands, animal selectivity, grazing frequency and defoliation intensity all operate together. Under infrequent defoliation, erect species such as grasses are able to grow taller to shade and suppress more prostrate species such as legumes, whereas under frequent defoliation, the competitive advantage of the grass is removed and the more prostrate species can dominate the sward. Thus, infrequent defoliation tends to maximize dry matter yields in both mono-specific and mixed pastures, but in the latter, there is also a change in species composition towards grass dominance. The legume, however, is not always the suppressed species. The stage of growth at which defoliation takes place will also influence the outcome of defoliation, but although grass dominance is favored by light or infrequent stocking, which results in an erect habit and a large number of fertile tillers (Stem and Donald 1962), this sequence in which the prostrate species are shaded and reduced may be modified by the particular characters of the dominant species. Such a character may allow the species to survive grazing or it may be a character that allows it to grow aggressively under the particular physical or chemical conditions prevailing in the niche it occupies (Vickery 1981).

Principal Floristic Groups Grazing land ecosystems have a diversity due both to their belonging to a particular floristic region and to edaphic and climate factors of the sites. As they are often secondary or derived formations, their flora is usually less rich than that of the climax ones. For example, the flora of the Guinea-type savannas is much poorer than that of the dense evergreen forest. On the contrary, the steppe areas of northern Mexico have a remarkable floristic diversity. Moreover, the flora of grazing land ecosystems has generally been profoundly influenced by the action of fire and

Range Management

1963

grazing; this leads to additional diversification, through a succession of vegetation types, but it is marked by a severe selection of species, of which the most tolerant, often having the lowest fodder value, are encouraged. Many autochthonous species have thus been eliminated and replaced by pantropical species that are adapted to the conditions and have been introduced either accidently or willingly by man (Imperata cylindrica, Heteropogon contortus, Hyparrhenia rufa and many others). In the grazing lands of annual grasses occurring in arid regions (Sahel steppes) the passage of fire before seed dispersal may have profound effects on the replacement of the grass cover, whilst, in more humid regions, fire is often responsible for maintaining a balance, and its suppression can lead to shrub invasion and a succession to climax forest.

Woody Strata Many grazing land ecosystems are mixed formations having herbaceous cover and a woody stratum, and the relative importance of these two layers is very varied. In the arid and semiarid climates, the woodlands and thickets are often of thorny species (Acacia spp., Prosopis spp. and Cactaceae). In zones having a dense dry forest climax, many of its species are adaptable to life in savanna or in woodland. The tree density of these formations depends on edaphic conditions (e.g. in the Brazilian cerrado) or on anthropogenic action (the Miombo of central Africa). In the zones having a dense evergreen forest climax (the Guinea-type forest-margin savannas and in Amazonia), the woody cover is always poor and only a small number of species can adapt to these conditions (Curatella spp., Byrsonima spp. in America; Hymenocardia spp. and Annona spp. in Africa). On waterlogged soils, palms may be the dominants. The woody layer influences the grass cover through its shade; under a humid climate, shade favors shade-tolerant species of low fodder value and low productivity; thus, ecological niches favorable for development of forest species are created, and it is often necessary to keep the woody vegetation under control by periodic fires. In all savannas in forest climax areas, the suppression of fire and competition of herbaceous plants due to overgrazing lead to rapid woody plant invasion. On the contrary, in a semi arid climate, shade may favor the development of mesophytic species having high productivity; for example, in the African Sudan zone, the grass biomass in sunlight zones varied from 70 to 250 g/m2 (according to the soil type), whereas it is >300 g/m2 in shaded areas; in the Sahel steppe areas, the grass biomass may be multiplied by 2.5 in shaded areas. Grass Stratum This is the essential element of pasture productivity. In humid zones, with long periods of active growth, grasses are usually almost only grazed in their green state during their growth period in the wet season and as secondary regrowth in the dry season. In arid zones, with a much shorter growth period, the shorter

1964

A. Swenne

grasses are largely consumed in a dry state after seed dispersal. Some graminoid species or some genera, with a wide geographical range, play a predominant role; these include: Andropogoneae • Hyparrhenia spp. and Andropogon spp., which are very well represented in all humid and sub-humid regions (especially H. diplandra, H. rufa, A. gayanus); • Saccharum spontaneum, Dichanthium spp. and Themeda spp. are important species of some grazing lands in Africa, Asia and Australia, Sehima nervosum, Cymbopogon spp. and Chrysopogon spp. are important in grazing lands of monsoonal Asia, especially India; • Heteropogon contortus and Imperata cylindrica are two pantropical species that are predominant in many types of pasture. Paniceae • Panicum maximum and Pennisetum purpureum, two species originating from the Guinea region of Africa, which are cultivated in all the tropics; • Cenchrus spp. represent one of the essential fodder resources of areas with long dry seasons; • Pennisetum clandestinum, originating from upland areas of eastern Africa, has been introduced in many regions, especially in South America (Brazil and Uruguay) and in Indochina; an excellent fodder species that forms an essential part of upland pastures; • Paspalum spp. has many species and is important in many grazing lands of humid and subhumid America; • Echinochloa spp. (in flooded areas), Panicum spp., Brachiaria spp., • Digitaria spp., etc. Arundinelleae Although of lesser value, many species of Loudetia spp., Tristachya spp. and Trichopteryx spp. play an important role in grazing lands on poor soils. Aristida spp. has many species and is found in all ecosystems having an unfavourable water balance due to climate or edaphic reasons. Many other species are capable of playing a varying role (Chloris spp., Cynodon spp., Eragrostis spp., Sporobolus spp., etc.). Genera having holarctic affinities (such as Bouteloua spp. in the New World and also Agrostis spp., Bromus spp., Poa spp., Stipa spp.) are found only at high altitude or in regions in contact with areas outside the tropics (as in northern India, the Cape region, southern Brazil, Mexico). The fodder value of forb species and their role is generally not so well known. Many species with a growth period longer than that of the grasses may play an essential role during dry periods in semi-arid pastures; this is true in the Sahel where it is the non-graminoid species (forbs or shrubs) that supply almost all the nitrogen. Legumes do not always play as important a role in tropical grazing lands as in temperate or Mediterranean regions. Some species are widely used, especially as crops; they mainly belong to Phaseoleae (Phaseolus spp., Vigna spp.,

Range Management

1965

Centrosema spp. and Pueraria spp.) or the Hedysareae (Stylosanthes spp., Desmodium spp., etc.). Their role in natural pastures is very variable; in the Guinea-type savannas of Africa, not only are they few in number but they are practically inedible (Indigofera spp., Crotalaria spp., Tephrosia spp., etc.); they have a more important role in America, as, for example, in northeastern Argentina (Desmodium spp.) or in the llanos (Stylosanthes spp.). A study of the semi-arid ecosystem of the Serengeti area (eastern Africa) has also shown that some legumes (Crotalaria spp., Indigofera spp. and Glycine spp.) were frequent and may even be predominant over several square kilometres; however, they are disregarded by the large herbivores and only Grant’s gazelle eats them during the dry season. The importance of some Trifolium spp. in upland areas, especially in eastern Africa, should also be mentioned. The possible role of plants belonging to other families is very little known; they may be important in arid zone pastures (e.g. in the Sahel many species remain green during the dry season and are very palatable) (UNESCO 1979).

Rangeland Productivity Biomass and standing crop have been used nearly synonymously to refer to the weight of organisms at a given time. The essential distinction between productivity and biomass or standing crop is that productivity is a rate process with a specified time interval, whereas biomass and standing crop refer to quantities at a particular point in time. Herbage is a term often used by range workers and refers to the biomass off all herbaceous vegetation at one point in time (Pieper 1978). Not all the herbage is usually eaten by livestock or other herbivores, since some may be unavailable (out of reach or protected by a shrub or spiny plant) or not readily acceptable at conservative stocking rates. Forage, although defined in various ways by different authors, generally refers to herbage available and acceptable to grazing animals (Pieper 1978). Thus forage is always less than herbage. Browse has been defined as “that part of leaf and current twig growth of shrubs, woody vines, and trees available for animal consumption” (Duvall and Blair 1963). Thus browse is comparable in some ways to forage. Herbage or browse biomass is usually expressed in terms of dry weight per unit area. The quantity of herbage and browse is never stable on rangelands within or between years. Most rangelands are characterized by a single growing season, when soil water and temperatures are suitable to support plant growth. Early in the growing season plant growth is slow, then reaches a peak during mid-growing season, and finally slows and then ceases during the dormant season. It is assumed often that herbage standing crop is nearly stable during the dormant season. Many processes, however, such as respiration, translocation, shattering, herbivory, and so on, continue during the dormant season and contribute to the decline in herbage biomass (Pieper et al. 1974). Thus herbivores face a declining food supply as the dormant season proceeds even without considering their own consumption.

1966 Table 2 Grass genera grouped according to the most abundant kind of non-structural carbohydrate stored in stem bases (Smith 1972; Ojima and Isawa 1968); quoted in Stoddart et al. (1975)

A. Swenne

Fructosans Agropyron Agrositis Alopecurus Arrhenatherum Bromus Calamagrostis Dactylis Elymus Festuca Hordeum Lolium Phalaris Phleum Poa Triticum

Sucrose Avena Sorghum Zea

Starches Andropogon Bouteloua Buchloe Cynodon Distichlis Echinochloa Eragrostis Leptoloma Muhlenbergia Oryza Orysopsis Panicum Paspalum Phragmites Sorghastrum Spartina Sporolobus Stipa

Physiology of Range Plants Most grasses fall into one of two groups with respect to the kind of non-structural carbohydrate present in vegetative parts those in which starches predominate and those in which fructosans predominate. The former are of tropical or subtropical origin; the latter are of temperate origin. Some contain neither starch nor fructosan in any quantity (Table 2). Starch is the most abundant carbohydrate found in trees (Kramer and Kozlowski 1960).

Phenology of Range Plants Growth Patterns In the tropics, grasses and legumes develop as annuals or as perennials. There are no true grass or legume biennials in the tropics. With some species, such as Chloris gayana, Cenchrus ciliaris, Hyparrhenia rufa and Panicum maximum, a short vegetative period of 4–6 weeks is followed by stem elongation and emergence of floral parts. Unless plants are cut, new tillers develop and produce inflorescences throughout most of the growing season. A number of species, such as Pennisetum purpureum, Tripsacum laxum, Andropogon gayanus, Axonopus scoparius, Melinis minutiflora and Brachiaria ruziziensis, remain vegetative over a longer period of time. Stem elongation occurs during the latter part of the rainy season in these species, with a fairly uniform development of inflorescences and seed maturity in the dry season. Perennial tropical legumes are more drought tolerant than grasses. Most perennial species produce flowers during the latter part of the rainy season or early dry season. Generally, flowering is non-uniform and spread over several weeks, or even

Range Management

1967

months. The growth pattern of grasses and legumes is different, and adds to the complexity of compounding mixtures for sowing, as well as to the difficulty in management of mixed swards. Furthermore, the imposition of grazing or cutting treatment during one phase of the growth cycle of any given mixture may indirectly affect subsequent cycles to a great extent.

Stages of Growth and Development – Crowder and Chheda (1982) identify the following general stages of growth: Seedling- time from emergence to tiller formation or axillary branching. – Establishmental – a transitional period when young plants are producing leaves and tillers, nodal and secondary roots. – Vegetative – production of leaves, shoots, stolons and rhizomes with no visible, or relatively few, floral stems. – Floral stem elongation – shoots having flower primordial increase in height, or length, as internodes lengthen. – Reproductive – floral and seed production. Leaf, Stem and Root Growth In the early stages of growth, the herbage consists entirely of leaves. As grasses age, stems comprise a greater percentage of the bulk forage. As the aboveground plant parts increase in size, the root system also enlarges. Roots generally comprise less of the total plant weight than the shoots. They usually constitute a greater proportion of the total plant weight in the juvenile stage than in the more mature plant stage, but exceptions occur. Species differ in their shoot: root ratios. Legumes produce a primary or taproot that develops vertically and reaches depths of 8 m or more, depending on the species and soil type.

Regulation of Growth Soil Moisture. Rainfall is the greatest single factor affecting growth and herbage dry matter production. Temperature. The dry matter yield of plant tops and roots of tropical grasses increases markedly with an increase in temperature to an optimum that lies between 30  C and 35  C. Tropical legumes have a lower optimal growth temperature than tropical grasses. Light Intensity. Plant growth increases as light intensity increases, up to the point of light saturation of the leaves in a canopy exposed to full sunlight. Shading would reduce plant growth and development. Burton et al. (1959) showed that reduced light decreased herbage yields, production of roots and rhizomes, nutrient reserves for regrowth and total available carbohydrates (TAC) in the herbage of Cynodon dactylon Coastal (Table 3). The effect was even more dramatic with high rates of applied N, with forage yields being decreased proportionately as light was reduced, i.e. a reduction of light by 30 % decreased yields by about 30 %. In full sunlight, applied N consistently

1968

A. Swenne

Table 3 Influence of light intensity on the growth and production of Cynodon dactylon Coastal (Burton et al. 1959) Available light (%) 100.0 64.3 42.8 28.8

Seasonal dry matter (t/ha) 15.5 14.1 10.6 8.1

Roots and rhizomes (t/ha) 5.17 3.51 3.44 2.39

Reserve index (g) 2.2 1.6 0.8 0.1

TACa (%) 15.8 14.0 10.5 9.0

Lignin (%) 9.2 9.7 10.2 10.4

a

TAC total available carbohydrates

increased plant density and leaf area. Both declined with shade and within 2 years many plants had died after receiving only 28.8 % sunlight. Shade significantly increased the lignin content of the herbage, which would decrease digestibility. Thus animals consuming forage produced under cloudy climatic conditions could be expected to make less live-weight gains. Photoperiod. The influence of day length on plant growth is usually over shadowed by its conspicuous effect on flowering in many grasses and legumes. When short-day plants, such as Stylosanthes humilis and S. guianensis (Mannetje 1965) and Hyparrhenia rufa (Agregeda and Cuany 1962), are grown in long days, many or all plants remain vegetative and accumulate dry matter.

Morphology of Plants and Grazing Morphological as well as physiological characteristics determine the ability of plants to tolerate removal of foliage. Specific differences in the number and position of perennating buds and growth sequences make plants differentially susceptible to injury from cropping. The location of meristematic tissues is important; greater impacts occur where active tissue can be readily removed. Shrubs can be eaten back into previous year’s growth, but with herbs, only the removal of the current year’s production is possible. Under repeated use, however, shrubs may become so compacted that animals can obtain forage from them only with difficulty, an effective means of protecting buds against removal. Plants with tough fibrous stems survive grazing better than do those with weak soft stems. Protective structures, such as thorns and spines, limit use and promote survival.

Growth Characteristics of Grasses The differences in growth and bud characteristics are of great importance for the range manager. Grasses that have a high proportion of vegetative stems, those that delay the elevation of the apical bud, and those that sprout freely from axillary buds are most resistant to grazing and most productive under use. It is these characteristics that influence differential responses among species such as the reduction of mid-grasses and the increase in low, turf-forming grasses on the short-grass plains.

Range Management

1969

Effects of Grazing on Range Plants The ability of a range plant to survive and produce forage under grazing cannot be explained on the basis of carbohydrate content, morphology, root growth, or reproduction alone. All are interrelated, and what adversely affects one affects another. For example, low carbohydrates and few vegetative buds may go together, and lowered forage production integrates all these separate factors together into a practical guide useful to the range manager when based on consecutive years of records. With some exceptions, experience has shown that total forage yield from grasses on arid ranges decreases with increased frequency of clipping and closer harvesting in any 1 year. Even more important is the cumulative effect of close and frequent forage removal over a span of years. Any grazing, whether moderate or heavy and whether early or late, has a measurable influence upon the metabolism of a plant. With reduction of photosynthetic tissue comes reduction in carbohydrate and nitrogen reserves and decreased forage production. This is one of the most important effects of incorrect grazing. Critical to the range plant is the influence of grazing upon volume and depth of the root system. A reduction of food reserves slows the growth of the entire plant, and root growth is no exception. The acceptance of these facts is basic to range management. Provided that grazing is neither too frequent, too close, nor improperly timed, plants have great ability to survive use.

Effects of Time of Forage Removal on Forage Production At any given level of use, forage production is affected greatly by the time of forage removal, for plants are more vulnerable at some periods than others. If forage removal occurs in the early growth stages and while moisture is available, the healthy plant quickly replenishes lost foliage and there is little disruption of plant functions. The same level of clipping in midseason is much more critical. In general, defoliation early in the growing season is less detrimental than at a later time. In dry climates, time cessation of grazing may be more important to plant welfare than time of beginning of grazing (Stoddart 1946). Grazing that removes herbage just prior to the onset of the dry season prevents normal food storage, development of roots, and formation of buds. Grazing or clipping after the foodstorage cycle has been completed has the least effect. Combinations of drought and heavy grazing are particularly detrimental to plants. Effects of Grazing on Production from Shrubs and Forbs Most shrubs and forbs, unlike grasses, are not well adapted to regenerate forage removed through grazing. It is well known that browsing or clipping shrubs encourages twig growth at the expense of flowers and fruits, but the effects only become apparent in following years. By thus keeping shrubs in a vegetative condition, increased forage production may result (Garrison 1953). Certain shrubs can withstand heavy utilization year after year when cropping takes place during the winter. Conversely, heavy and repeated clipping during the growing season can cause rapid declines in subsequent forage yields and increase plant mortality (Lay

1970

A. Swenne

1965). Generally, forbs are less resistant to grazing than are grasses, although many leguminous plants withstand repeated use. Many of the more productive forbs are weak-stemmed and are readily damaged through breakage and lodging by trampling (Stoddart et al. 1975).

Effects of Grazing upon Reproduction of the Plant Decrease in valuable forage plants on the range does not result entirely from the death of established plants, although death is a significant factor (McKell et al. 1966), but also from a decrease in reproduction and consequently a smaller number of young plants available to replace those that are dying. The life span of a range plant varies from a few weeks in annuals to 50 years or more in shrubs and, possibly, in perennial grasses also. When it becomes evident that mature plants are not being replaced by young plants, maintenance of the range requires an immediate change in livestock management to allow normal reproduction. A healthy range has a mixture of many classes of plants. Old and senescent plants die and are being replaced constantly by new ones. The reproduction process must not be interrupted to the point where no new plants are available to replace those that are dying. The influence of grazing upon seed production is twofold. The animals may graze the plant so heavily as to consume the seedstalks prior to dropping of seed, or cropping may disturb the physiology of the plant as to inhibit seed formation. Probably the latter is far more important than the former, despite the general opinion to the contrary. Very intense grazing would be necessary to consume all the seed produced by healthy plants. Because nature is so lavish with the number produced, the seed remaining after the usual grazing may produce all the young plants for which there is space. Some plants protect their seeds so as to make them inaccessible to ordinary grazing; on others they are easily accessible and highly preferred. Since seed formation requires large quantities of concentrated food reserve, any depletion in the reserve of the plant interferes with normal seed formation. Seed production is especially important to annuals, since it is the only way they can reproduce. It has been shown that seed production in annual grasses can be greatly reduced by clipping, especially late in the growth season. It is unlikely, however, that grazing can reduce seed production below the amount needed for production (Stechman and Laude 1962). Grazing may influence species composition, however, because of its differential effect on seed production and other responses among various species (Stoddart et al. 1975). Other Influences of Grazing Upon Vegetation Animals also influence vegetation indirectly. Grazing animals have an influence upon the soil, tending to compact it to surprising depths, especially during the spring or other moist season. Not only are compact soils poor absorbers of precipitation, they also restrict normal root development. The roots are sometimes only half their normal length. Compaction is greatest near the surface; thus the hard soil increases the difficulty of seedling emergence and establishment, and depresses vigor (Barton et al. 1966). Conversely, when the soil is not wet, animals are believed to be beneficial in loosening the soil surface and in covering seeds that

Range Management

1971

have accumulated on the surface. The mechanical action of animals in loosening seed, in carrying burs, awned seeds, and the like, in their hair, in distributing hardcoated seeds through the feces, and in loosening bulbs, corms, and bits of rhizome so that they may be transported elsewhere is probably of unexpected importance. There are instances in which total protection of range from livestock has failed to result in the expected revival of the vegetation, presumably because of the lack of animal action in aiding reproduction. In arid regions, it is possible that grazing induces better moisture relations since, with the removal of herbage, the transpiring surface is reduced and plants may be better able to withstand drought. Where great quantities of plant material accumulate, production may be lowered (Stoddart et al. 1975).

Range Management Range Inventory Ecological surveys deal with the physical and environmental factors such as precipitation, topography, soils, and broadly defined vegetation communities. They provide useful basic information to other kinds of range inventories, though they are not strictly essential to them. Commonly, they are undertaken to provide a framework for other more intensive inventories. These broad surveys often form the first step in range management planning (Stoddart et al. 1975). Range forage inventories are most commonly for the purpose of determining the grazing capacity of rangeland either for domestic livestock or wild herbivores. Special consideration is given to plant species and types that are preferred as forage, though other physical and biological data bearing on range productivity and management procedures may be included. Utilization surveys are ways of assessing the current grazing pressure exerted by foraging animals as a means of determining appropriateness of current stocking levels or management systems. Condition and trend analyses are at present the most important range evaluations. The data compiled from them enable the range manager to judge the adequacy of stocking and management practices. Based on successional and community dynamics concepts, they are designed to assess whether range sites within the range ecosystem are at, or depart from, accepted standards and capabilities based on their potential for production. Multiple use surveys are whole spectrum analyses done to determine the entire biological and physical resource base with the objective of integrating all capabilities and uses of the range into a comprehensive and coherent plan. Rangeland appraisals are for the purpose of determining economic productivity of a range area. The physical inputs come from information developed in inventories described above. These are related to other factors outside the range area, but intimately associated with it.

1972

A. Swenne

The initial determination of the quality of grazing land considered primarily as a system for meat production is through species inventories, association inventories and the morphological and behavioral adaptive structures of plants. The whole question is to define production potential in terms of the total matter available for transformation by livestock. The mode of growth and production of plant communities or individuals indicate the present production of the ecosystem and the way in which the ecosystem evolves under the influence of human interventions. It is important that all aspects of primary production be studied in detail for effective management; primary production processes control the growth rate of animals and thus determine the total yield of the ecosystem (UNESCO 1979).

Vegetational Attributes A vegetation inventory provides data on the absolute or relative abundance of plant species by vegetation types. Data may be estimates or they may be quantified by numeration, by ground cover, by volume, or by weight. Normally, the data are derived from sample plots positioned throughout the area sampled. Vegetative types most often form the basis of the sampling unit (Stoddart et al. 1975). There are many methods available for determining vegetational characteristics (Brown 1954; Mannetje 1978; Pieper 1978; Risser 1984). Weight or Biomass Increases in biomass through the growth process over time are generally considered productivity estimates that include a time dimension. Most estimates of plant biomass or standing crop include only that above the soil surface. Belowground biomass is very important for plant functions, but it is difficult to measure and generally not included in inventory or monitoring procedures. Direct harvesting is considered the most reliable method of determining aboveground biomass, but it is too time consuming to be of practical value for inventory or monitoring of extensive range areas. Several weight estimate techniques have been developed for rapid and fairly reliable determination of herbage weight (Pechanec and Pickford 1937; Shoop and McIlvain 1963). These procedures involve estimating herbage weight by species from small quadrants in the field. Training of observers in the field is necessary. This can be done easily by checking the estimates with clipped quadrants. The method is considered reliable enough to be used on detailed research studies (Shoop and McIlvain 1963). Weight estimates can be adjusted by clipping a portion of the quadrants that have been estimated. Double sampling procedures involving regression adjustments have been outlined by several workers. Area or Cover Aerial or canopy cover refers to the area covered by the vertical projection of the crown of plants onto the soil surface (Brown 1954; see Fig. 5). Basal cover or area refers to the area occupied at the intersection of the plant and soil surface. Woody plant cover is often expressed in terms of canopy cover since the basal area of trees and shrubs is very small in relation to the role of these plants in the plant community. Basal area is most often used for herbaceous plants since it was

Range Management

1973

Fig. 5 The manner in which vegetation is projected for purposes of cover determination. If other than forage cover is being estimated, the foliage in the top of the shrubs would also be included (Modified from Stoddart et al. 1975)

assumed that basal area was not influenced greatly by seasonal precipitation and temperature. These assumptions concerning the relative stability of basal area of grasses, however, may be misleading (Young 1980). Cover determination is often conducted for inventory and monitoring purposes. The two methods that appear to meet time requirements for inventory and monitoring procedures are estimation (Stewart and Hutchings 1936; Daubenmire 1958), and the point-step method (Evans and Love 1957). Estimating procedures usually involve estimating cover by species in relatively small plots. Often, cover classes are used instead of whole percent unit estimates (Table 4). In this case it is only necessary to estimate cover in the nearest cover class and then to use midpoints for data summarization. The point-step method was developed as a rapid, objective method of determining cover and species composition of large range areas (Evans and Love 1957). The method involves cutting a notch or marking a spot on the observer’s boot. The observer paces across the range area, recording whatever is directly beneath the notch or mark on his or her boot. Individual species, litter, bare ground, rock, and so on, can be recorded. Other devices, such as a fine rod or tripod, can be used to make placement of the point more objective (Owensby 1973). Care must be taken to make the point as small as possible, to avoid overestimation of cover (Holechek et al. 1989). In older range literature, cover was called density, which should not be confused with its present usage.

1974 Table 4 Cover classes rated according to percentage of ground surface covered by vegetation (Daubenmire 1958)

A. Swenne

Class 1 2 3 4 5 6

Coverage (%) Range 0–5 5–25 25–50 50–75 75–95 95–100

Midpoint 2.5 15.0 37.5 62.5 85.0 97.5

Density or Frequency The simplest inventory is a species list in which consideration is given not to relative amounts, but only to presence of a plant. Number lists are made by counting the number of individual plants of each species occurring in sample plots (Stoddart et al. 1975). Density is defined as the number of individual plants per area (Cooper 1959). In some cases, it is difficult to identify an individual plant for sod-forming species (Dix 1961). In these situations, it may be necessary to use a plant unit such as an individual shoot. Density can be determined by counting the number of plants in quadrants, but quadrant size is critical. Large quadrants serve well for vegetation with low density but may be too time consuming for areas with high density. Frequency sampling is fast and easy to conduct in the field. If one determines density from quadrants, frequency can be calculated from the same data since frequency represents the percentage of the quadrants in which the species occurs. Quadrant size is critical with frequency sampling also. If the quadrant is too large, many species will have high frequencies, whereas if the quadrant is too small, frequencies will be too low, especially for the less abundant species. Hyder et al. (1963) provide guidelines for frequency sampling. Sample Plots Sample plots vary in size, depending primarily on the kind of vegetation studied. Tree and shrub stands require larger plots than herbaceous vegetation. The most effective sampling of an area can be obtained by the use of numerous small plots, rather than fewer and larger plots, but the plot chosen must be large enough to encompass individual plants of the larger species present. Spacing of individual plants and the number and distribution of species are important in determining plot size. The plot size required increases both as distance between plants and the number of species increase. Plots may be round, square, or rectangular. Sometimes rectangular plots are elongated greatly in length and narrowed in width (belt transect), sometimes to a mere line (line transect). Transects are especially useful to sample across ecotones where one vegetation type intergrades into another. Permanent plots are commonly square or rectangular since marking the corners of the plots with stakes ensures more accurate placement of a plot marker at subsequent visits. Iron rods hurried in the soil can be used also for plot marking. Subsequent detection is then made by using a metal detector. Irrespective of shape, permanent plots are commonly referred to as quadrants especially when

Range Management

1975

the position and area of each plant are mapped. Round plots are used more frequently for temporary than for permanent plots. Whatever size of plot and the area sampled, effort must be made to obtain a representative sample. Though carefully selected plots may fulfill this criterion, statistical theory requires that plots be randomly located, since all measures of statistical reliability are based on chance occurrence. As a practical matter, a purely random arrangement of plots makes plot location much more time consuming than a mechanical arrangement in which plots are spaced regularly at fixed intervals. Consequently, elements of the two methods are combined. This is done by randomly selecting lines through the area under study and randomly locating plots along these. Thus, plot location is simplified, and chance is permitted to operate so that statistical analyses are possible. Since the number of plots required to obtain an adequate sample depends on the heterogeneity of vegetation, this number can be minimized by reducing the heterogeneity of the population being sampled. This can be done by dividing the area under study into subareas on the basis of differences in the vegetation. Plant communities differ from place to place because of edaphic factors such as slope, exposure, and soils. Recognizing these differences beforehand and making these subareas the sampling units well reduce the variability among individual plots which, in turn, enables one to attain a given degree of reliability with few plots (a smaller sample). The outlines of these sampling areas should be drawn on a base map, preferably aerial photos, before collection of vegetation data is begun (Stoddart et al. 1975).

Photography and Remote Sensing Photographs are a useful aid to range analyses because they provide visual evidence that is difficult to convey by data alone. They are especially useful in reconstructing changes in vegetation over a period of years. Close range photography has not, however, proved helpful as an analytical device because of the distortion resulting from varying proximity to the camera of vegetation occupying different height strata. Aerial photographs are more useful, their quality and usefulness varying with the type of imagery used. Most useful are large-scale (1:1000–1:5000) aerial photographs taken for the specific purpose of making vegetative resource surveys (see chapter “▶ Climate Aspects of the Tropics,” this volume).

Range Condition In classifying rangeland for potential productivity under good management, it is necessary to know whether the vegetation is improving or deteriorating. This requires a knowledge of the desirable natural and naturalized plants, their competitiveness and desirable densities, acceptability to animals, tolerance of drought and trampling etc. Weedy species can serve as indicators of the degree of deterioration or depletion. Information about soil characteristics, particularly inherent fertility, is important with regard to maintaining vegetative cover for sustained herbage growth and as a safeguard against erosion (Crowder and Chheda 1982).

1976

A. Swenne

The term “range condition” to the range manager has a special meaning relating current condition of the range to the potential of which the particular area is capable. It should not be confused with immediate availability of forage. The range manager attempts to discover whether the plants that should grow in a particular situation are present and in good vigor. He notes the quantity of each species present as a basis for determining the degree to which the productivity of the range has been impaired. Range condition, in this sense, is best described as the state of health of the range.

Assessment of Range Condition Climax Approach. This method of rating range conditions is applicable to perennial grasslands and is based on a comparison of the present vegetation with that of the previous composition at a given interval of time (Dyksterhuis 1949). As indicated earlier, range plants, whether desirable or undesirable, can be classified as “increasers,” “decreasers” and “invaders.” The first two can be valuable herbage types but the latter are mostly undesirable and are generally abundant on overgrazed, unstable rangeland. In developing the base from which future comparisons are made, individual species are placed in one of the three groups and relative percentages of each group (inclusive of all species) are recorded. This is usually tabulated so as to rate the range condition as excellent, good, fair or poor. Such data should be taken periodically throughout the season to provide records under varying conditions. Thus, by knowing the current range condition, and comparing it with past situations under similar circumstances, adjustments in stocking rates and distributions can be made. This scheme is more satisfactory where bunch type grasses comprise the vegetation rather than creeping and trailing types (Crowder and Chheda 1982). Palatability Rating. In this system used for annual-type rangelands, ratings are made of plants highly acceptable to the livestock. Data are collected early in the season before moving animals on to the range. Excellent to good conditions indicate a large proportion of highly preferred plants, a relatively dense cover, a thin mulch on the soil surface and no active erosion. Ranges in fair to poor condition are dominated by less palatable species and a greater number of undesirable plants. Those in poor condition consist of sparse soil cover, plants of poor growth, many weedy species and heavy soil erosion (Crowder and Chheda 1982). Range Potential. This approach attempts to express the current herbage production in relation to the ultimate potential (Humphrey 1949). It requires prior knowledge of the range vegetation and output, with emphasis given to ratings of botanical composition and density of cover, plant vigor (potential production), quantity of mulch and degree of soil erosion. In predicting potential herbage production and stocking capacity, it is assumed that (1) range condition is not a temporary state but repeatable under comparable environmental circumstances, (2) a current rating differing from the expected normal would not necessitate reclassifying the range condition on an annual basis and (3) excellent to good range will produce more than fair to poor condition, even though the current rating might suggest a modification (Crowder and Chheda 1982). Score Cards. In this method, the range evaluator has before him a list of important factors such as (1) general growth and vigor of the desirable herbage

Range Management

1977

species, (2) density, composition and overall grazing value of the vegetation, (3) indicator plants, including annual grasses, weeds and poisonous plants, (4) soil erosion indicators, such as quantity of mulch, extent of erosion and formation of gullies and (5) animal indicators, which include weight gains of the livestock and appraisal of the wildlife population. Numerical values assigned to the various items are scored and summarized to provide a rating of the range condition (Parker and Woodhead 1944).

Condition Trend Condition ratings, even though they are accurate, are of little use without knowing the trend in condition. A range in poor condition that is still deteriorating requires different treatment from a poor range that is in the process of improving. Trend has been defined as the direction of change in range condition. Generally trend is considered upward (or improving) or downward (declining) or stable. Determining trend is highly important. Generally, livestock reductions and wide-scale changes in management are unnecessary if condition trend is upward, although, by improved management, rate of range improvement may be increased. Poor range condition does not mean that current management practices are wrong. Only trend of condition will reflect correctness of current grazing practice. Trend can be determined only by careful analyses of the range. Judging range trend is even more hazardous than judging range conditions, because there are few objective means for assessing trend. Soil and vegetative factors are commonly listed to ensure consideration of a uniform set of factors. Soil factors include, among others, presence of litter, evidence of soil trampling, and presence and condition of gullies. Plant factors include such things as plant vigor, seedling establishment, degree of present utilization, and evidence of past utilization, especially on browse plants. Obviously, each of these factors may be judged positively or negatively. Gullies may be new and raw indicating deterioration, or they may be filling in, indicating improvement. If seedlings of better species are present, conditions are favorable. Seedlings of poor or invader species are an indication of deterioration (Stoddart et al. 1975). Vegetation measurements on one site at intervals of time can indicate whether the trend in condition is up or down, provided there is opportunity for comparison with a reference site that is known to have had little or no grazing use. This comparison is essential because seasonal changes can induce temporary vegetation responses, which mask the effects of management. If details of stock management of the site are known, it is often possible to identify the cause of the trend. Vegetation changes due to grazing can usually be related to the percentage of the current season’s growth that is utilized by the animal (Heady 1973).

Grazing Management and Stocking Rate Stocking Rate and Grazing Capacity Stocking rate is defined by the Society for Range Management as the amount of land allocated to each animal unit for the entire grazeable period of the year.

1978

A. Swenne

Carrying or grazing capacity are terms commonly used when discussing stocking rate. These terms refer to the maximum number of animals that can graze each year on a given area of range for a specific number of days without inducing a downward trend in forage production, forage quality, or soil. Although actual stocking rates may vary considerably between years because of fluctuating forage conditions, grazing capacity is generally considered to be the average number of animals that a particular range will sustain over a period of time. Selection of the correct stocking rate is the most important of all grazing management decisions from the standpoint of vegetation, livestock and economic return. Although this has been the most basic problem confronting ranchers and range managers, specific approaches to this problem are still generally unavailable. It is generally agreed that there is no substitute for experience in stocking-rate decisions on specific ranges. Regardless of the technique used, all methods so far developed based on vegetation analyses yield only an estimate of grazing capacity. True grazing capacity can be determined only by stocking with an estimated number of animals and watching the range trend (Stoddart et al. 1975). A practical solution is recognizing vegetational changes that lead to range decline and making adjustments of stocking rates before deterioration becomes severe. The importance of range condition and trend studies have been stressed as guides in the determination of carrying capacity and assessment of management. There is, however, a lack of information for most tropical rangelands concerning productivity of major range types and their stocking capacities under different treatments. Information on potential herbage production is needed rather than preconceived concepts of vegetation climax and succession. In many places, the livestock numbers already exceed the potential stocking capacity, especially in localities of communal grazing. Thus, overgrazing is the common practice and destocking becomes difficult to achieve, because it demands a radical change in the pastoralists’ way of life. Under such conditions, completely new approaches must be devised to modify the land tenure system and the social structure before range management practices can be imposed (Crowder and Chheda 1982).

Grazing Systems Grazing systems are alternative means of grazing by deploying stock on the pasture year round. They usually include a period of rest from grazing. They may incorporate periods of intense grazing to increase use of less palatable species and to reduce competition for the palatable species (Acocks 1966). In other systems, the animals are moved almost continuously on the principle that overgrazing occurs plant by plant and that if animals are prevented from regrazing a plant it will recover and may even be improved in vigor because of tillering or branching induced by removing of apical dominance (Goodloe 1969). Some systems employ rest periods that are long enough to ensure that seeds are set either periodically or every year, or that desirable plants attain sufficient size to withstand another grazing. A very important use of the rest periods is to promote sufficient fuel for a controlled burn and to prevent damaging overuse of the recovery growth. Traditional grazing societies frequently employ forms of deferred grazing to make hay or

Range Management

1979

reserve fodder for winter grazing. The following characteristics of a good grazing system are listed by Stoddart et al. (1975): – It is based on the physiology and life history of the plants. – It is suited to the kind of plant present. – It is adapted to soil conditions, and erosion, for instance, will not result during heavy grazing. – It will move plant succession toward higher productivity by favoring the desired plants. – It is not detrimental to animal gain. – Its implementation is practical in a ranching operation. The last point is vital. Heavy investment in fencing, water installations, etc., is neither economically nor socially possible in most parts of the arid zone. The improvements perceived by the introduction of a grazing system have frequently been ascribed to the fact that, to operate it, considerably more attention had to be paid to the condition of the pasture than was normal, thus improving pasture management irrespective of the system employed. Grazing systems are secondary in importance to the regulation of animal density and require considerable understanding of plant phenology and long-term experimentation (Harrington 1981). Grazing systems can be classified into the following categories: Continuous: livestock are placed on the range and allowed to remain indefinitely, as is the case of year-round grazing even with seasonal herbage growth. Animals have free access to any part of the range. Deferred: the range is divided into camps or paddocks (sometimes called ranges) so that a long period of rest is systematically allotted to each, the deferment falling at different times of the year over a predetermined number of years. The resting period coincides with a fixed time (by calendar or season) before and after burning and may include an interval to allow for seed set and maturity. Variations of the deferred grazing system have been employed in various parts of Africa in order to provide fodder reserves during the dry season and to accumulate flammable material for late and effective burns. A certain number have been described in the literature (West 1955; Heady 1960; Rains 1963; Naveh 1966; Howell 1978). Rotational (sometimes called divisional rotation): separation of the range into 2, 3, or 4 equal, or nearly so, areas and rotating the entire herd from one division to another at systematic intervals. Deferred rotation: a division of the range with a given portion deferred at some critical period of the year, generally during seed set and maturity, with sufficient time allowed for seedling establishment.

Choice of Grazing System The choice of grazing system depends on such factors as the condition of the range and the trend of condition, rainfall and its distribution, length of the dry season; vegetative cover and the potential production of grasses, forbs and browse plants; objectives of the livestock enterprise; the land tenure system; type of livestock and

1980

A. Swenne

quality of desired product; and above all, managerial skill. The system should be simple in design and implementation since the more complicated grazing schemes require closer supervision and greater attention. If a deferred plan is employed, a small number of paddocks and infrequent movement of cattle simplify range and grazing management. With less available moisture, the value of continuous grazing becomes more pronounced. In regions with less than 250–375 mm yearly rainfall, the formalized deferred and rotational schemes are not appropriate, so that traditional nomadic and transhumance rotations between wet-dry and dry-season grazing, under controlled stocking rates, are probably more desirable. Rotational grazing is usually more beneficial in the more humid areas where there is wider range of species and greater differences in animal acceptability of plants.

Influence of Stocking Rate on Livestock Production Generally, as stocking rate is increased, productivity per animal declines. Differences in animal productivity between light and moderate stocking rates are much less than between moderate and heavy stocking rates. Although productivity per animal unit declines as stocking rate increases, productivity per unit area increases up to a point. It then decreases as scarcity of forage reduces nutrient intake by livestock. Maximum gains per animal and per unit area are not possible concurrently. Animals will cease to gain weight as forage becomes increasingly scarce and often lower in nutritive quality. During drought periods, heavy stocking can be economically disastrous because the complete lack of forage will necessitate that all livestock be removed from the range and fed hay or sold at low prices. Ranchers using light to moderate stocking rates have much higher levels of forage standing crop throughout the year and generally more vigorous plants than do those using heavy stocking rates. This forage reserve permits much less adjustment in animal numbers than would be necessary under heavy stocking rates. The decline in livestock performance per animal unit as grazing intensity increases is explained by reduced forage intake and diet quality. Decreased forage availability reduces animal selectivity and forces them to consume diets lower in quality. It also forces animals to spend more energy on foraging activity that could otherwise go into production. Influence of Stocking Rate on Economic Returns Heavy grazing generally maximizes gross economic returns, but net economic returns are maximized by moderate grazing. Death losses and supplemental food costs are higher for heavy grazing compared with moderate grazing (Klipple and Costello 1960; Shoop and Mcilvain 1963; Abdalla 1980). On cow-calf or ewe-lamb operations, weaning percentages (percentage of female animals in the herd producing a marketable offspring) are lower for heavy compared with moderate grazing. Livestock can make high gains on heavily grazed range for a few years, particularly if they are given supplemental feed and precipitation is average or above. In drought years, however, the reduction in livestock productivity both per animal

Range Management

1981

Fig. 6 Effects of overgrazing

unit area is far more severe than on moderately grazed ranges. Death losses from poisonous plants are much higher on heavily grazed ranges because the nonpoisonous, palatable species are less available. Continued heavy grazing results in gradual degradation of soil and vegetation resources (Fig. 6).

Range Improvement Drastic manipulations of range ecosystems are sometimes required or desired. The invasion of unwanted plants, severe droughts, past abuses by grazing animals, or the desire of the operator to change botanical composition or productivity on all or part of the range unit can make practices to revegetate with useful plants desirable. High management inputs are required once these risky, costly practices are used if the land manager wishes to realize a reasonable return on his investment. The most economical method for reclaiming deteriorated grazing lands is through use of methods not requiring planting of desirable species. This may be accomplished by control of unwanted plants, concentrating moisture or harvesting precipitation, and/or grazing management. Proper grazing use of desirable plants is very important. If natural revegetation is not feasible, planting of desirable vegetation may be needed (Holechek et al. 1989). According to Harrington (1981), however, capital-

1982

A. Swenne

intensive measures involving herbicides, fertilizers and machinery are generally not justified in the arid zones except as once-only corrective measures.

Vegetation Control A notable increase in stocking capacity can be achieved by clearing land of unwanted woody growth so as to allow the growth of native and naturalized grasses. Bush control operations should first be carried out on land of high production potential. These should be followed by measures to prevent reinfestation. The costs and expected economic returns should be carefully calculated before the decision is made to launch a bush control program (Crowder and Chheda 1982). The control of unwanted plants is necessary to make more water available for the reproduction and production of desirable vegetation. This may be accomplished by chemical, biological, or mechanical means; by judicious use of fire; or by use of different animal species. Plant control in range management is simply the reduction of unwanted or undesirable plants that have invaded or increased in a plant community. Plants “out of place,” or the movement of certain species out of their normal range or habitat, is one of the major problems on rangelands of the arid and the semi-arid regions (Holechek et al. 1989). Fire is the most common method of bush control on most rangelands. Handslashing can be a rather cheap and effective way of controlling vegetation where labor is available. In the humid regions, however, rapid regrowth occurs from stems and roots, so that other control measures are often needed or else the benefits may not last even for 1 year. In that case it can be useful to combine slashing with burning. Slashing with sufficient time for drying of woody material prior to burning generates more intense and uniform fires when heavy bush covers the area than burning without slashing (Crowder and Chheda 1982). Considerations in selection of mechanical methods (bulldozer, holt breaker, chaining, cutters etc.) are availability of equipment, the size and stand of the plants to be eliminated, whether the target plants have sprouting or nonsprouting characteristics, soil conditions and the type of terrain (Holechek et al. 1989). In general, mechanical control methods are costly and are used on land of high productive potential along with other practices for range improvement. Satisfactory control of unwanted plants and considerable improvement in the grazing capacity of rangelands may often be obtained by applying herbicides (Gibbens et al. 1986). Herbicides may be classified as contact, translocated, selective, nonselective, and soil sterilant. A contact herbicide kills only those plant parts that are directly exposed to the chemical (for example, diquat and paraquat). A translocated herbicide is applied to one part of a plant but is carried to other parts of the plant by plant tissues (for example, 2, 4-D, 2,4,5-T). A selective herbicide (e.g. the herbicides listed as translocated herbicides) kills or damages a particular species or groups of species with little or no injuries to other plants. A non-selective herbicide kills or damages all plant species (e.g. paraquat). A soil sterilant is an herbicide that kills or damages plants when it is present in the soil.

Range Management

1983

Most of these herbicides are selective at low rates and non-selective at high rates. Broadcast spraying is the method of herbicide application most commonly used on rangelands. Since the herbicide is applied to all plants, desirable as well as undesirable, selective herbicides are generally required. Applying granulated or pelleted herbicide is also used to control unwanted plants. The latter method is less dependent than foliar sprays on stage of growth but does require precipitation to dissolve the granules or pellets so that the herbicide may penetrate into the soil. Fundamentals to consider are: – Proper kind of herbicide. – Proper rate of application. The amounts of herbicide required to provide adequate control vary among plant species. Higher rates than those needed for adequate plant kill cause damage or death to leaves and branches, so that herbicides are not translocated to the proper site and death of the plant does not result. – Proper volume. The volume is dictated by the target species. It is important to obtain adequate coverage but not excessive amounts that will seriously contaminate the adjacent environment. – Proper time. The phonologic development of the target species, or associated plants, is a reliable index to seasonal susceptibility. Plants are most sensitive to foliar sprays when they are growing vigorously, and the leaves are fully expanded (Holechek et al. 1989). Results from herbicides and arbocides have been variable, with kills up to 80 % or more for first applications. Repeated treatments, and combinations with other control measure such as burning, are needed for more complete control. Browse plants are likely to be seriously damaged or killed (Crowder and Chheda 1982).

Fertilization Fertilizers have rarely been used on rangeland outside the USA and even there, the economic advantages are rare (Stoddart et al. 1975). Phosphorus has been successfully used on the infertile soils of Australia to establish rangeland legumes (Shaw and Bryan 1976). It is unlikely that economic benefit will be derived from applying fertilizer to arid and semi-arid rangeland. Response of natural and semi-natural grasslands to added fertilizer nutrients has been demonstrated, but more favorable results occur when combined with other improvement practices (Crowder and Chheda 1982). Nevertheless fertilization has advantages over other means of range improvement. It requires no highly specialized equipment; costs are less than for seeding, and a period of rest is not necessary. Fertilization is undertaken primarily to increase forage production, but it can also provide a more varied forage mix, more palatable and nutritious forage and a longer grazing season. The most widely applied elements are nitrogen, phosphorus and potassium.

1984

A. Swenne

Nitrogen. Most tropical soils are deficient in nitrogen and heavy applications are required to produce high yields of grass with high protein content. Differences in response to applied nitrogen has been observed. This is due to the following factors: species, stages of growth, amount and time of application, soil moisture and climatic condition. Unfortunately, there is little or no response to nitrogen fertilization in the arid and semi-arid regions. Nitrogen is subject to leaching by water and in areas of heavy rainfall, split applications are necessary to avoid nitrogen losses. Grasses respond better to nitrogen applications than legumes. Therefore, continued use of high nitrogen fertilizers causes a rapid decline in the legume component of grass-legume mixtures. According to Vallis et al. (1968), this is due to the fact that grasses grow faster and are more aggressive than legumes and that they are more persistent under frequent defoliation. Furthermore, applied nitrogen reduces nodulation of the legume, which reduces its competitive capability. Response of native species in depleted rangelands to nitrogen application is usually low, except if phosphorus is supplied at the same time. Phosphorus. Highly weathered tropical soils are generally deficient in phosphorus. Phosphorus fertilizers become efficient only when the P-fixation capacity of the soil is reached and residual phosphorus supply accumulates. Phosphorus is little subject to leaching, and it is thus possible to fertilize heavily with expectations of carryover effects for several years. When pastures are grazed, a large portion of the soil P taken up by grasses and legumes is returned in the excrement of grazing livestock or in plant residue. This also reduces the need for high applications of fertilizer after initial fertilization. Unlike that of nitrogen, the phosphorus requirements of grasses depend more on soil properties than on the grass species. Once the fixation capacity is satisfied, one or two applications per year is sufficient to meet the nutrient requirements of grasses. Phosphorus is essential in nodule development and nitrogen fixation of legumes. Successful establishment and growth of legumes cannot be expected if phosphorus is not available in adequate quantities. Response of native species in rangelands is usually low, but legume density can increase. Potassium. Response to potassium application depends on the species fertilized as well as on the availability of nitrogen and phosphorus. Since about 80 % of the potassium consumed by animals is returned to the soil via excreta, the need for maintenance application is low.

Reseeding Whether natural revegetation or artificial planting is used depends on the residual vegetation. For further revegetation to be effective, there must be residue of desirable plants to take over and dominate the site. In instances where undesirable vegetation is competing severely with the establishment of desirable vegetation, it is necessary to reduce or eliminate such undesirable vegetation (Holechek et al. 1989). Range seeding is practiced in the USA in regions where the natural flora has been degraded by overstocking or where the natural woody vegetation has been

Range Management

1985

removed. Mature plants compete with seedlings for moisture and nutrients, which results in most seedlings failing to establish. Thus, the first plants to establish themselves can resist subsequent germination events even under very favorable conditions. If artificial seeding is carried out in such a way as to favor the desired species and the resulting plants are properly managed, the benefits, and thus the costs, may be spread over many productive years without fear of competition from undesired species (Harrington 1981). In deciding whether an area should be seeded, the range manager should ask the following four questions: – Is seeding absolutely needed? Range can be rehabilitated more positively and at lower cost by better livestock distribution, better systems or reduced stocking. Only where the desirable native perennial forage plants are almost completely killed out is seeding essential. – Are proven methods available for the site? Where not available, projects should not be undertaken until satisfactory procedures have been developed. – Can prove methods be used? On many sites the procedures are known for the general type but cannot be applied because excessive rocks, steep slopes or other factors prevent use of the types of equipment or procedures needed. – Can the area be given proper grazing management after seeding? Seeding should not be started until proper grazing management can be assured.

Basic Criteria for Successful Revegetation – Change in plant cover must be necessary and desirable. – Terrain and soil must be suitable for seeding. Deep fertile soils on level. – To-gently sloping land are preferred sites for seeding. – Precipitation and water concentration must be adequate to assure establishment and survival of seeded species. – Competition from unwanted plants must be removed or reduced. Most plants used for revegetation are perennials. Seedlings of these species are often slowgrowing and cannot compete with existing, unwanted plants. A good seedbed will provide the best possible moisture conditions for existing plants before seeding. In addition, it is sometimes necessary to control unwanted plants that are competing with the seedlings of the desirable plants. – Adapted plant materials should be used. The plant species selected for seeding must be compatible with management objectives. It is important to use only those species and varieties well adapted to the soil, climate, and topography of the specific site being revegetated. If native plants are being revegetated, species of local origin are used. – Mixtures of plant types rather than single species should be seeded because: all areas have variable conditions of soil, moisture and slope and each species produces abundantly on the site more nearly supplying its needs; seasonal forage production is likely to be more uniform; a mixed diet is more desirable to livestock; some plants of the mixtures may have favorable influence on others (e.g. legumes). Mixtures, however, require greater management skills since they

1986

–

–

–

– – – –

–

A. Swenne

are subject to differential utilization by animals and tolerance to grazing. Highly palatable species may be killed out and less palatability species may be underutilized. Seed treatments should be used. Various microbial treatments (e.g. nitrogen fixing bacteria or mycorrhizal fungi) may enhance seedling survival. Dormancy of most seeds can be reduced by special treatments (e.g. scarification). Proper seeding rates should be used. It is important to use enough seed to get a good stand, but not more than necessary. Too much seed can produce a stand of seedlings so thick that individual plants compete with each other. Proper depth of seeding is necessary and determined by the plant species. Optimum depth of seeding is roughly four to seven times the diameter of the seed. Seeding equipment should be used that provides positive seed placement at the desired depth. More stands are lost because seeds are planted too deep than too shallow. Correct seeding dates are important. The most desirable time to seed is Immediately before the season of the most reliable rainfall and when temperature is favorable for plant establishment. Uniform distribution of seed is essential. Seedbed preparation is essential also. The major objectives of preparing seedbeds for seeding are to (a) remove or substantially reduce competing vegetation, (b) prepare a favorable micro environment for seedling establishment, (c) firm the soil below seed placement and cover the seed with loose soil, and (d) if possible, leave mulch on the soil surface to reduce erosion and to improve the micro environment. Revegetated areas must be properly managed. All seedlings must be protected from grazing by animals through the second growing season, or until the seeded species are well established.

Methods of Direct Seeding Drill Seeding. Drilling is by far the best method of planting seed where site conditions permit. The seed is covered to the proper depth by the drill control, distribution is uniform, the rate of seeding is positively controlled, and compaction can be utilized if necessary. Broadcasting. is any method that scatters the seed directly on the soil without soil coverage. The seed, however spread, must be covered in some way if it is to germinate and become established. Limitations to broadcasting seeding are: – – – – –

a heavier seeding rate is required; covering of seed is poor compared to drilling; distribution of seed is often poor; loss of seed to rodents and birds can be great; establishment is generally slower. This method should be avoided if possible (Holechek et al. 1989).

Range Management

1987

Interseeding. An alternate method is to cut furrows or otherwise destroy the existing vegetation in strips placed across the area at fixed intervals into which the desired plants are sown. This technique offers the following advantages: there is less disturbance to the site; the species introduced can be those that complement existing forage; forage production remains high during the treatment period; where legumes are introduced, this may result in higher production of the existing species, and it is less costly than complete cultivation (Stoddart et al. 1975).

Water Conservation The most universal factor limiting production in arid and semiarid zones is lack of adequate soil moisture. In those areas, any mechanical modification of range sites that will improve infiltration into soil will reduce soil-moisture stress and increase production. Several techniques exist to reduce surface runoff and improve infiltration. Pitting involves the creation of small basins to catch and hold precipitation. Specialized equipment have been developed in the USA and in Australia for that purpose. Pitting is usually combined with seeding operations. Chiseling can be used on heavy clay soils and where hardpans form beneath the soil surface. Not only is water penetration low on these areas, but plants find it difficult to establish themselves in the compacted soils. Water-spreading structures consists of dams and dikes that intercept surface runoff and convey it out of natural drainage areas at low gradients across the land surface where it can be absorbed. Similar systems of water management are used by African pastoralists for crop production and for rangeforage supplementation. Dams and brushwood deflectors in the ephemeral streams of the Sahelian Zone divert water onto “run-on” areas. These small areas, periodically irrigated, may produce almost as much forage as the vast area of surrounding rangeland (Stoddart et al. 1975). Contour furrows can be ploughed or listed strips placed close together and generally not smoothed after plowing.

Provision of Water and Salt for Livestock On arid and semi-arid rangelands, adequate daily water supply is rare. The daily water intake of range livestock reaches about 8.21/100 kg of live weight in the dry season, and about half this amount during the wet season. Actual water intake varies with the moisture content of herbage and climatic conditions. Animal output is greater with daily watering than every second or third day, and the distance walked between grazing and the watering site affects productivity. With widely spaced watering points, the area around the supply is seriously trampled arid overgrazed, whereas the more distant herbage is not utilized and generally of low quality. Opening and closing of water points can be used to draw animals to or away from sections of the range as desired. Salt licks can be used similarly. Left to themselves on the open range, stock will overgraze favored parts of the range whilst

1988

A. Swenne

leaving substantial areas ungrazed. Government attempts to control human nomadic movements using a similar technique are difficult to achieve in practice because closing of watering points is very unpopular. Construction of new watering points is often part of attempts to develop arid-zone environments by opening up unused or little-used pastures. In the absence of control of stock numbers, the result is usually an increase in the area of depleted range with no increase in animal production or decrease in drought susceptibility (Harrington 1981).

Animal Production Numbers and Distribution of Cattle, Sheep and Goats The world’s cattle population is increasing more slowly than the world’s human population. The total is now more than 1000 million. Somewhat more than one-third of the world’s cattle population are to be found in the tropics. Within the tropics, the largest concentrations of cattle are found in northeast and east Africa, South America, the Indian subcontinent and tropical Australia. Stocking rates are very high in the Caribbean and the Indian subcontinent and very low in the humid region of central Africa, the semi-arid region of western Asia and the humid region of Papua New Guinea. On the other hand, the number of cattle per 1000 inhabitants is very high in tropical Australia (Where cattle are extensively ranched), moderately high in west, northeast and east Africa (where cattle are often managed in a nomadic or transhumant system), and in South America, where cattle are ranched. There are small numbers of cattle per 1000 inhabitants in the humid region of central Africa, where the tsetse fly is prevalent, in Southeast Asia where the farming system is mainly a subsistence one, and in Papua New Guinea (Fig. 7). Approximately one-quarter of all African sheep are to be found in South Africa. Within tropical Africa, sheep are important in Ethiopia, Kenya, Mali, northern Nigeria, Somalia, the Sudan and Tanzania. Within tropical America, the largest national flocks are to be found in Mexico, Bolivia, Brazil and Peru. Within tropical Asia, with the exception of India, sheep populations are relatively small, but there has been a very marked increase in the sheep population of Myanmar. Virtually all the sheep in tropical Oceania are to be found in the drier areas of tropical Australia. More than 50 % of the world’s goat population is found in the tropics. The largest concentrations are in Africa and in the Indian subcontinent. Goats are possibly the most widely distributed of domestic livestock. They are found in countries representing the climate extremes of the tropics, from the arid and semiarid areas of South America to the wet and humid tropics of Southeast Asia. Their wide distribution is partially explained by their ability to survive and thrive in environments where vegetation is extremely sparse. Their rustic and hardy qualities enable them to withstand dry environmental conditions much better than cattle. They perform best, however, in the drier tropics and on light sandy soils. In Africa, for instance, the greatest concentrations of goats are to be found in East Africa, northern Nigeria and Morocco. This pattern of distribution is also true of the

Range Management

1989

Fig. 7 Stocks of cattle and buffaloes, and sheep and goats (2010) (FAO 2013)

Cattle and buffaloes

Sheep and goats

2.0

billion heads

1.5

1.0

0.5

W

or

ld

ia an ce O

Eu

ro

pe

ia As

ic er Am

Af

ric

a

as

0.0

Indian subcontinent, western Asia, South and Central America and the Caribbean. Dwarf goats are found throughout the humid tropics, and it could be that they are especially adapted to this type of climate.

Nutrition and Feeding of Livestock Dry Matter It is important to know the amount of dry matter since animals eat a certain amount of forage, the dry weight of which is proportional to the size of the animal (ca. 2.5 kg/day/100 kg live weight for cattle). The percentage of water present in vegetation varies considerably during the seasons. Non-climatic factors can prolong or attenuate these variations; these include the vegetative stage of the plant, waterregulating mechanisms and buffering effects of surrounding vegetation. The amount of dry matter present is generally fairly low at the beginning of the growing period and varies between species, but is ca. 30 %. It then increases fairly rapidly and remains for a long time at ca. 50 %, reaching 60 % at the time of fructification. Leaves that are still alive in the middle of the dry season only have ca. 25–30 % water. Standing dead matter during the wet season has a very varied water content (15–40 %) depending on the amount of rain preceding harvesting; the dry matter

1990

A. Swenne

content of straw during the dry season is nearly 95 %. The dry-matter content of the tree leaves (initially 30–40 %) increases in the same way. It appears that grass growth is disrupted if the water content is reduced to below 60 %, but some dicotyledonous appear to be more resistant. The rate of desiccation of living herbaceous matter is governed by the density of the vegetation cover. The percentage of water increases from 30 % in herbaceous savanna to 40–45 % in shrub savanna (according to the extent of woody cover) and never falls below 50 % in savanna woodland (UNESCO 1979).

Energy The major nutrient need for body functioning in terms of quantity is that of energy. This is provided by carbohydrates (starch, cellulose, hemicellulose, and sugars) and fats. Energy provided by the metabolism of these nutrients is necessary for maintenance of body heat and for work, growth, fattening, and reproduction. Proteins may also provide energy if supplied in excess of the animal’s needs for muscle and tissue formation (Stoddart et al. 1975). Several measures of energy value of forage have been proposed: digestible energy, metabolizable energy and net energy. Successively, each of these represents a further refinement, and theoretically, each gives a more accurate index to nutrient value than the preceding index. Digestible, Metabolizable and Net Energy Of the total energy available from a forage (between 4.19 and 4.90 Cal/g on a dry-matter basis in tropical pastures, Butterworth 1964), a portion is lost as faecal energy and the remainder is considered as the usable digestible energy. About 20 % of this digestible energy is lost as combustible gases, mostly methane during digestion and in the production of urine, leaving metabolizable energy available for the animal’s metabolic and productive activities. Additional energy is lost as heat when the animal digests and metabolizes the forage. The energy that remains is the net energy of the forage available to meet maintenance and production needs of the animal. Energy Content The energy content of forage is a fairly variable concept. The choice of nutrition units and their representative value have been widely discussed by livestock rearers; some were concerned with digestible energy content, but this is generally expressed in terms of forage units (one forage unit is the energy equivalent of 1 kg of barley = 1648 Cal) or in starch units (1 starch unit = 1.43 forage unit). The energy content is, ecologically speaking, the calorific value that can be used in the study of energy flows within the ecosystem. Plant matter contains slightly more than ca. 4 Cal/g dry weight. It is necessary to make measurements if more detailed analysis is required since this value varies widely between species, organs and seasons. The distribution of energy within the plant biomass markedly differs from the distribution of the biomass itself. Maximum values are attained by seeds (>5 Cal/g dry weight), less in green and standing dead leaves and are generally least in litter and roots (sometimes ca. 3 Cal/g). Marked variations occur during the growth cycle, especially if calorific value is determined from dry matter without ash.

Range Management

1991

Maximum values in aerial organs are reached at the time of flowering and fruiting; significant energy concentrations may also occur at other times in perennial plants, when there is accumulation of reserves in twigs. The creation of energy and nutrient reserves also explains the range of variations observed in roots (up to 1 Cal/g) (UNESCO 1979). Calculation of Energy Total Digestible Nutrients – If digestibility is available: DOM ¼ DCP þ DEE þ DCF þ DNFE Where DOM is digestible organic matter; DCP digestible crude protein; DEE digestible ether extract; DCF digestible crude fiber; DNFE is digestible nitrogenfree extract; TON is total digestible nitrogen; DM is dry matter; CF is crude fiber; SE is starch equivalent. TDN ¼ DOM þ ðDEE  1:25Þ – If digestibility is not available: fresh grasses: TDN ¼ 54:6 þ 3:66Loge CP  0, 26CF þ 6:85Loge EE hay: TDN ¼ 51:78 þ 6:44Loge CP Energy – Metabolizable energy: ME ðCal=kg DMÞ ¼ TDN  3:65 – Net energy: NE ðCal=kg DMÞ ¼ ME  ð1 Cal  DMÞ NE ¼ FU=kg DM In fodder units (FU): 1648 NE In starch equivalent: 2360 ¼ SE=kg DM

Feeding Requirements Tables 5, 6, and 7 give the feeding requirements for cattle, sheep and goats. Forage Intake by Animals An animal consumes forage in varying amounts. Therefore, assessment of forage quality depends not only on the nutritive value of the forage but also on the quantity of that forage voluntarily eaten or, in other words, on the total quantity of digestible nutrients consumed by the animal. The gross or total energy content of tropical pasture is relatively constant, varying between 17.2 and 18.7 MJ/kg of dry matter (Minson and Milford 1966). Because of this constancy, the results of most intake studies are expressed in terms of dry matter. The quantity of dry matter eaten depends not only on the quality of the pasture but on the size of the animal eating the pasture. To eliminate differences in body size on the measurements of feed intake most results of intake studies are quoted in terms of grams of feed dry matter eaten per unit metabolic weight, where metabolic weight is the 0.75 power of body weight in kilograms. Feed intake expressed in this way varies from 30 g/day per kg 0.75 for mature tropical pasture to 140 g/day per kg 0.75 for immature temperature pastures.

1992

A. Swenne

Table 5 Feeding requirements of cattle

Energy in FU

DCP (Digestible Crude Protein)

Ca P NaCl

Maintenance (kg) 100: 1.2 150: 1.6 200: 2.0 250: 2.3 300: 2.6 400: 3.2 500: 3.8

Growth Per kg gain Weaning: 1.2–1.7 6–12 months: 2.1 12–18 months: 2.7 18–24 months: 3.0 24–36 months: 3.2 0.6 g per day per Total in g per kg live weight FU weaning: 130–140 6–12 months: 100–130 12–18 months: 80–100 18 months +: 80 5 g/100 kg live 15–25 g/kg gain weight 3 g/100 kg live 10–20 g/kg gain weight 5 g/100 kg live 2 g/kg gain weight

Fattening Per kg gain Beginning: 3.0 Middle: 3.5–4 End: 4–5

80–120 g/UF

Milk production 0.38 per kg milk at 4 % fat

60 g per kg milk at 4 % fat

Gestation 7th month: 0.1/100 kg live weight 8th month: 0.2/100 kg live weight 9th month: 0.3/100 kg live weight

100/FU

3 g/kg milk

6 g/100 kg live weight 1.5 g/kg milk 5.5 g/100 kg live weight 2 g/kg milk

In ruminants, intake depends largely on the capacity of the digestive tract, particularly the rumen. The animal stops eating when a certain degree of “fill” is reached and starts to eat again when the “fill” is reduced as a result of digestion and movement of the residue through the digestive tract. Even though with forage of very high digestibility (over 70–75 %), blood metabolite level can control intake by ruminants, under most conditions (particularly in the tropics where forage generally has lower digestibility), gastro-intestinal fill controls intake. Therefore, some authors (Raymond 1969; Cordoba et al. 1978) have concluded that involuntary physiological reflexes, rather than subjective preference, control voluntary forage intake by ruminants. Animal intake of tropical forage, particularly grasses, is, in most cases, considerably lower than temperate species. This low level of intake is generally considered to be the major cause of low animal productivity in tropical environments. Ingalls et al. (1965) suggested that 70 % of the variation in animal productivity can be accounted for in terms of voluntary intake differences, as compared to 30 % accounted for by digestibility differences. A higher rate of intake is directly related to the shorter time that ingesta are retained in the rumen (Poppi et al. 1981). Intake of tropical legumes is considerably higher than that of tropical grasses.

Range Management

1993

Table 6 Feeding requirements of sheep

Energy in FU

DCP (Digestible Crude Protein) Ga P NaCl

Maintenance (kg) 10: 0.26 20 0.38 30: 0.47 40: 0.53 50: 0.58 60: 0.64

Growth Per 100 g gain 1st months 0.16 2nd month 0.21 3rd month 0.27 3rd month +: 0.32 Per kg live Total in g per weight FU 3 months: 2–3 g/day (lamb) 150–190 g 0.8–1.2 g/day 3–5 months: (adult) 135 g 0.5 kg/100 kg 1.8–7.5 g/day live weight 0.3 kg/100 kg 1.2–4.5 g/day live weight 0.5 kg/100 kg 5 g/kg gain live weight

Fattening 3rd month +5 % 4th month +20 % 5th month +50 % of maintenance requirements

Milk production 0.6 FU per kg milk at 8 % fat

Gestation 0.40–0.55 per 100 g gain

3 months -: 0.8 g 110 g/l Total: per kg live weight milk 60–70 3–5 months: g/FU 1.3–1.8 g per kg live weight Total: 3.5–5 g 4–5 g/milk 2.5–3.5 g 3–4 g/milk 2 g/milk

Table 7 Feeding requirements of goats

Energy in FU

Maintenance (kg) Growth 10: 0.43 Per 100 g gain: 20: 0.50 0.15–0.30 30: 0.57 40: 0.64 50: 0.71 60: 0.78

DCP (Digestible Crude 30–55 g/day Protein)

Total in g/FU: 100–170

Ca P

2.0–3.2 1.3–2.0

0.7–3.0 0.5–1.8

Milk production Per kg milk at: 3 % fat: 0.32 4 % fat: 0.36 5 % fat: 0.4 Per kg milk at: 3 % fat: 50 4 % fat: 55 5 % fat: 60 4 g/kg milk 3 g/kg milk

Gestation 4th month: maintenance +0.25

4th month: maintenance +20 g

Maintenance +1.5 g Maintenance +1.8 g

Sensory Responses and Palatability The critical decision of an animal about whether or not to accept a particular food item is controlled by information relayed by the sense organs of sight, smell, touch and taste, probably in that order. The term palatability may be used to refer to the sensory properties of the food that influence its acceptance by the consumer. The ability to distinguish basic properties of sweetness, sourness, bitterness and saltiness seems to be common to all mammals, but animal species vary in their

1994

A. Swenne

thresholds of detection and tolerance for these tastes. Domestic ungulates reject herbage contaminated by feces of their own species, but can be induced to accept it if the odor is masked by spraying the herbage with molasses (Odberg and FrancisSmith 1977). This response is presumably adaptively related to avoiding parasite contacts and may be of significance to the formation of localized defecation sites by those grazing ungulates that occupy restricted ranges. In experiments with sheep, Theron and Booysen (1966) found that the tensile strength of grasses was the most important determinant of preference. In other experiments with sheep involving impairment of different senses; however, Krueger et al. (1974) demonstrated that taste had the greatest influence on forage preferences, with other senses appearing supplementary.

Deterrents The significance of secondary compounds of plants to plant-herbivore relationships has become clear only in recent years. Cates and Rhoades (1977) propose that these secondary chemicals can be divided into two categories: – Digestibility-reducing compounds, mostly tannins, which form complexes with protein. These inhibit the digestive action of proteolytic enzymes and possibly interferes directly with the form of these enzymes. This binding is most effective at low pH, below pH 8 for condensed tannins and below pH 5 for hydrolysable tannins. Their action is associated with an astringent sensation. – Toxic substances that interfere directly with the physiology of the consumer or possibly with that of its symbiotic microbiota. For example, pyrrolizidine alkaloids and oxalates affect liver and kidney function, other alkaloids and amines attack the central nervous system, saponins destroy erythrocytes by lysis, cardenolides act on muscle systems, cyanide inhibits the action of cytochrome oxidase and non-protein amino acids become miss-incorporated into proteins. Monoterpene alcohols, but not their esters or hydrocarbons, exert an inhibitory effect on the rumen microbiota. Most of these compounds seem to be associated with a bitter taste. Volatile terpenes (essential oils) produce a variety of characteristic odors. Tannins seem characteristic of woody plants and are much less common in herbs. An estimated 80 % of dicotyledonous woody perennials contain tannins, compared with 15 % of herbaceous dicotyledons (Cates and Rhoades 1977). Tannins are particularly abundant in mature tissues, where they may be present in quantities up to 60 % of dry mass. A difference of 7 % in the tannin content of the North American legume Sericea lespedeza caused the food intake of cattle to drop by 70 %. Alkaloids have been reported from several 1000 plant species of a wide variety of growth forms, whereas significant quantities of cyanogenic glycoside have been found in several 100 species. The proportion of alkaloid-bearing plants seems to be higher in tropical regions than in temperate areas. Toxins are generally present only in small amounts, and reach highest concentrations in immature tissues, which are the most susceptible to herbivory. Generally, secondary

Range Management

1995

compounds seem much less prevalent among grasses. The widespread occurrence of various secondary chemicals in dicotyledons, and their general absence from the Gramineae, could be an important factor in the grazer-browser dichotomy.

Fodder from Trees and Shrubs Importance of Fodder Trees and Shrubs Herbage for livestock taken from trees and shrubs is known as browse. It may be eaten directly from the natural growth of the plants or from regrowth of sprouts after cutting near ground level (known as coppice). In addition, woody branches can be cut from taller shrubs and trees, thus falling to the ground, where the twigs, seeds, pods and even the bark are eaten. This is known as pollarding and is a common practice with Acacia senegal, Terminalia brownei and Baphia bequaertii, for instance (Dougall and Bogdan 1958; Lawton 1968). In Australia, Brachychiton populneum, Heterodendrum oleifolium, Ventilago viminalis, Flindersia maculosa and Casuarina cristata withstand lopping. Regeneration takes place from roots or stems, but Acacia aneura seems incapable of regeneration unless a large amount of leaf is retained. The value of certain trees and shrubs as fodder in times of drought has been widely recognized (Dougall and Bogdan 1958; Everist 1969; Wilson 1969). This is particularly the case in arid and semi-arid zones, where leaves, pods and flowers may form a high portion of the diet at certain times of the year (Lawton 1968; Gray 1970). As most browse plants are deep-rooting, they are able to exploit the soil moisture and fertility to greater depths than grasses and forbs, maintain green leaves for longer or initiate new growth ahead of the wet season. Some are deciduous and lose their leaves at the onset of dry weather; the leaves are eaten from the ground by livestock. Some leaves may have senesce and drop just ahead of new leaf development. Both legumes and non-legumes are browsed, although the leguminous species have the added ability to nodulate and enhance soil fertility. Other characteristics distinguish fodder trees and shrubs from herbaceous species: – Their perennial nature. All ligneous and most sub-ligneous species are pluriannual, and, as a result, provide a constant biological resource. – Their long growth phase. The phenology and development of ligneous species do not depend to any marked degree on the rainy season. Some species produce new leaves well before the return of the rains, and foliage lasting on the tree long after the end of the rainy season is very frequent. Some species are also able to provide a feed source throughout the year. Whether in the form of leaves or fruit, these plants enable a supplementary feed source to be found during difficult periods. – Their relatively delicate character when young. The fact that ligneous, plants are delicate when young is linked with their perennial nature, which is not conducive

1996

A. Swenne

to regeneration by natural sexual processes. While the plants are young they are vulnerable to constant browsing and sensitive to fire (Piot 1980). An estimated 75 % of the trees and shrubs in Africa are browsed to some extent by domestic animals and game (Whyte 1947). In the savannas in the arid, semi-arid and sub humid zones, ligneous species are an important component of livestock diet. Many of them are browsed or lopped as dry-season feed. Ligneous species also have an important effect on the quality, seasonality and productivity of the grass cover growing in their shade. The main browse species found are the following: Acacia albida, A. senegal, A. seyal, A. mellifera, A. etbaica, A. bussei, A. ehrenbergiana, A. laeta, A. tortilis, A. flava, A. giraffae, A. sieberiana. Many legume species are of great importance also. This is the case of Bauhinia rufescens, Entada africana, Pterocarpus lucens, Albizzia amara, Dichrostachys cinerea, Prosopis africana, Tamarindus indica, and Dalbergia melanoxylon. Other families have important browse species such as the Capparidaceae, the Combretaceae, the Tiliaceae, the Sterculiaceae, the Rhamnaceae, the Anacardiaceae, the Rubiaceae, the Zygophyllaceae and the Salvadoraceae. Ruminants cannot meet their maintenance needs on dry grass alone. Since in the dry tropics, the dry season lasts 6–9 months, and there is usually no supplementary feeding, livestock and wildlife often depend entirely on browse to balance their diet in protein, phosphorus, calcium and vitamin A during this season. Many pastoral groups in the arid and semi arid tropics habitually lop branches from various forage species to make the forage accessible to livestock during the dry season. In Latin America, shrub-dominated ecosystems used as grazing lands cover huge areas. This is the case of the Cerrado and Caatinga of central eastern and northeastern Brazil, the coastal deserts and subdeserts of Peru and Chile, the Chaco of Argentina, Paraguay and Bolivia, the Monte of Argentina, the Andean Puna of Peru, Bolivia, Chile and Argentina. As in other parts of the World, browse is mainly used by livestock outside the growing season when the weather is too cold or too dry (Gasto and Contreras 1972; Soriano 1972). Roseveare (1948) listed 385 trees and shrubs as being eaten by cattle in South America. Some important species are the following: Acacia spp., Prosopis spp., Cercidium spp., Capparis spp., Caesalpinia ferrea, Cassia excelia, Ziziphus jozeiro, Piptadenis spp., Lycium spp., Chenopodium peniculatum, Atriplex coquimbana, A. atacamensis, A. repanda, A. sagit tifolium, Maytenus spp. and Opuntia spp. The arid zone covers 70 % of the surface of Australia. About one-third receives less than 250 mm of rainfall and is only used for extensive grazing. Browse is an important component in extensive grazing systems, and some 200 species have been reported to be browsed by livestock (Everist 1972), although only 40 of them are widespread and play a major role in the livestock industry. The browse ecosystems cover some 2.5 million km2 or 30 % of Australia’s land surface. Some of the Australian browse species have been introduced to other continents and are planted as forage, especially several species of Acacia spp., Atriplex spp. and Maireana spp.

Range Management

1997

Lists of browse species can be found in many publications (Van Rensburg 1948; Everist 1958; Kadambi 1963; Wilson and Brendon 1963; Lawton 1968; Le Houerou 1980; FAO/UNEP 1983).

Nutritive Value of Browse Browse plants are less subject to seasonal variation than grasses in terms of nutrient content. Furthermore, they leaf out at the end of the dry season, before the rains and before other forage plants appear. This occurs at a time when animal need is maximal for feed of a higher nutrient content, as they are grazing on low-quality grasses. Browse plants alone keep healthy animals in fair condition, but may be inadequate as the sole feedstuff. A mixture of several species for browsing is superior to a single species. The pods of some trees are highly nutritive and readily consumed by livestock and game. In general, browse plants have consistently higher crude protein levels than grasses, ranging from about 10 % to more than 20 % (leguminous shrubs) on a dry weight basis (Rose Innes and Mabey 1964). Fiber content and lignin is higher than in grasses, and minerals are high also with an average of 10 %, although calcium and phosphorus are usually low. Dry matter content ranges between 30 % and 60 % while trees and shrubs are growing compared to 60–80 % for dry grass. Dry matter digestibility varies from less than 30 % to over 70 % according to species, the part of the plant and the phenological stage. Browse plants are low in energy and added energy content evaluated on the basis of digestibility averages 3 MJ/kg fresh matter (0.42 FU) but varies in the same way as digestibility (0.28–0.86 FU).

Browse Intake The intake of browse varies with the type of animal, season, the alternative vegetation, availability and palatability of the ground flora. Species of animals differ in the amount of browse they eat. With regard to feeding habits, livestock species are usually classified as being grazers or browsers. According to circumstances, however, grazers can satisfy part of their nutritional needs by browsing woody species; there are no 100 % browsers among livestock. The percentage of feed from trees, shrubs and grasses in the diet of livestock varies according to availability as well as to the seasons. Cattle are considered to be typical grazers whereas sheep and goats are transitional between extreme grazers and extreme browsers. Goats, however, definitively use more fodder from trees and shrubs than do sheep. The goat is the domestic animal that is best adapted to woody forage in view of the anatomy and physiology of its digestive tract. As indicated earlier, browse is on average poor in energy. In consequence, the animal needs to ingest greater quantities to satisfy its needs. The goat has smaller maintenance requirements, per unit of live weight, than cattle or sheep:

1998

A. Swenne

18.64 Cal/kg live weight, against 21.94 for sheep and 31.35 for cattle. It can ingest more than 6 kg of dry matter per day per 100 kg live weight, whereas the corresponding figures for sheep and cattle are 3.8 and 2.9. Thus, taking into account the utilization coefficient of the food, the lower limit for energy content in order to meet maintenance needs is 31.35 kcal/kg of dry matter for goats, 57.75 for sheep and 107.25 for cattle. One can conclude from these figures that the goat is the only domestic ruminant that is capable of growth using browse exclusively, thus explaining its adaptation to and proliferation in poor vegetation (Le Houerou 1980). Browse can ensure the maintenance needs of sheep but does not allow production. Where browse is the principal diet of heep and cattle in the absence of herbage, either seasonally or fulltime (e.g. in Somalia and some parts of Australia dominated by Acacia aneura), productivity is low. Animal selectivity plays a major role in determining the usefulness of browse plants. Acacia senegal, readily eaten by goats and camels, spreads on extensive areas in the absence of these animals. The thorny nature of A. brevispica and A. mellifera makes them little accessible to cattle but they are freely eaten by goats. A. nilotica and Combretum spp. and Albizzia spp. are preferred at certain periods or during the absence of other food; at times, only the flowers, which shed, are eaten. In Australia sheep prefer A. aneura, whereas Capparis mitchellii inflorescences are eaten in preference to leaves. Acacia cambagei is only eaten when dry. Some species of browse, although regarded highly as fodder, have been reported as toxic at times, e.g. Heterodendron oleifolium (Everist 1969), Acacia nilotica (Glover et al. 1966). Eucalyptus populnea trees are regarded as unpalatable and poisonous but may be safely eaten when young. The fruits of Atalaya hemiglauca contain toxins, but the lower leaves are eaten by sheep and cattle without toxic effects.

Propagation Techniques The main methods used are: – – – – –

direct drilling planting out of young plants raised in the nursery planting rootstocks planting out cuttings, before or after the roots have been set grafting.

Direct Drilling. Direct drilling is used only for a small number of species, generally those with low height and a high rate of germination with rapid development from the seedling stage. This is the case for Atriplex semibaccata, A. glauca, A. canescens, Haloxylon persicum, H. aphyllum, Artemisia herba-alba. The advantage of direct drilling is obviously its low cost. Planting out Young Plants. Planting out young plants raised in the nursery, although much more expensive, is often preferred. This is especially true in arid

Range Management

1999

areas since the chances of success are far higher. Competition between the trees and shrubs planted can be more easily reduced or eliminated, and planting densities are low so that it is possible to provide one or more waterings, which will help the young plants to recover. The best technique is to concentrate runoff water on the beds to be planted several months before planting, to ensure that the minimum necessary reserves of water are in the soil. This can be done by creating microcatchment areas during land preparation. The young plants should be between 75 and 100 days old when they are transplanted. They should not exceed 20–25 cm in height nor have a diameter of over 5 mm at the neck. In the arid zone, it is recommended that the resistance of nursery plants should be strengthened by low and infrequent irrigation at least a month before planting out. Seed scarification before planting is often necessary in order to prevent inhibitions to germination caused by seed coats, especially for Acacia and Prosopis species. Propagation by Cuttings. Many species are suitable for propagation by cuttings from the stem or roots, from lignified plant material, from root stalks or from young branches (e.g. Atriplex halimus and A. nummularia). Propagation by cuttings can be practiced in the nursery or out in the field. In the latter case, rooted cuttings may or may not be used. In many cases it is advantageous to use rhizogenic hormones (indol butyric acid). Plant Spacing and Density. Plant spacing and density depend on the species under consideration, on the development expected and the management and utilization methods planned. For instance, shrubs such as Medicago arborea and the most common Atriplex are planted at densities of 500–5000 stocks per hectare depending on the environmental aridity, the nature of the soil and the management method. Maintenance. During the first few years, it can be advisable to reduce competition from weeds. This allows better establishment in terms of survival rates and more rapid growth. Species like Faidherbia albida and Prosapis, which tend to start their development in a bushy way, can be pruned in order to develop a single trunk. Other species like the Atriplex need to be cut back to remain within reach of the animals. In most cases, it is necessary to enclose plantations to avoid damage caused by uncontrolled grazing by livestock and wild animals. This can be done by using live fences which is cheaper than barbed wire, but they should be established at least 3 years before planting fodder trees or shrubs.

Management The management of a plantation depends on many factors such as the species, the production type (fruit or leaves), the type of utilization (gathering by man or direct browsing by livestock) and whether the production system is extensive, intensive or semi sedentary. In some cases, the fruit are collected when they fall on the ground and may even be sold as is the case in Mali, Senegal and Mexico, for instance. Most species are browsed directly during periods of scarce feed supply or may even be utilized continuously (Leucaena). It is advisable, however, to avoid browsing during the period of active growth of the shrubs. Fodder shrubs and trees should

2000

A. Swenne

usually be planted in single species stands. Different species have different biological and edaphic requirements and react differently to utilization. As a consequence it is very difficult to obtain a rational management for each one of them when they are mixed together. If diversification is desired this than be done by having monospecific plantations within the same management unit. The species used for direct browsing generally need periodic cutting back, either for the purpose of rejuvenating and revigorating of aging plantations or to bring the consumable biomass within reach of the animals. Coppice growth produces an abundance of very palatable forage. It should be used carefully and with moderation to avoid exhausting the shrubs and killing them. As a rule of thumb, a forage utilization rate of 60 % constitutes a maximum not to be exceeded at each utilization. The duration of each utilization should be as short as possible and, as a general rule, should never exceed 1 week or normal regeneration might be impaired. Generally speaking, trees have a productive life of 100 years or more. Shrubs remain productive for a shorter period of time (up to 40 years for Atriplex).

Yields Ligneous species are characterized by a relatively regular, if not constant, interannual production. As such, they provide a stabilizing element in pastoral or agropastoral systems. Their root system, which is generally powerful, plays a particularly important role. The roots of trees and shrubs can often penetrate to a depth of 10 m and sometimes well beyond. This root system enables these species to reach levels of subsoil in which water is present (inaccessible to grass species) enabling them in particular to make use of ground water at a fairly deep level. The perennial nature of these species also enables them to use early or late rainfall or rainfall occurring at unseasonal times. Browse plantations established on good soils are often highly productive. In Zambia and Zaire respectively (formerly known as Rhodesia), 10–20 large Acacia albida trees per hectare produce from 1100 to 2200 kg of pods without serious yield reduction of surrounding grass (West 1950). Smaller trees such as A. subulata, when spaced at 25–50 per hectare yielded from 550 to 1100 kg of pods per year. For Opuntia spp., yields are usually between 3000 and 10,000 kg/ha/year (Monjauze and Le Houerou 1965). Atriplex spp. plantations in North Africa and the Near East produce 1250–5000 kg DM/ha/year of forage (Franclet and Le Houerou 1971; Le Houerou 1975). Plantations of Acacia cyanophylla, A. salicina, A. victoriae and A. ligulata have a forage production of 1500–6000 kg DM/ha/year. In the semi-arid zone, Medicago arborea produce 3000–6000 kg DM/ha/year (Hamrouni and Sarson 1974). In the arid and semiarid zones of West Africa, plantation shrubs can produce 500–2000 kg of consumable DM/ha/year (Le Houerou 1980). In the sub-humid and humid African tropics, production of Leucaena spp. ranges from 6 to 12 t of DM/ha/year of forage (Savory and Beale 1974; Taylor 1980). Generally speaking, well-established and managed plantations have an annual per hectare production of 5–10 kg of consumable DM for each millimeter of rainfall.

Range Management

2001

Economics A study on the economic viability of browse plantations (Acacia, Opuntia, Atriplex and Prosopis) in Africa (De Mongolfier and Le Houerou 1980) reached the following conclusions: operating costs are fairly low owing to the extensive nature of the plantations, the fact that labor requirements to protect them and ensure their productive output are fairly modest, and that labor costs in Africa are low in any case. Investment costs vary widely but unit costs (cost per plant) vary much less. The lowest costs occur when the planting density is high (as for Opuntia spp. and Atriplex spp.) or when planting is carried out by direct drilling and not by transplanting young plants raised in the nursery (Leucaena leucocephala and A triplex semibaccata). Direct drilling, however, cannot be used in the semi-arid and arid areas of tropical Africa, where it is too risky. In areas where rainfall is unreliable, even transplanting nursery plants is not without risk of failure and requires special care (the plants must be watered) to ensure survival. As a result, tree and shrub plantations in these areas are fairly expensive. A second factor further increases the establishment cost of plantations in Africa, namely, the cost price of barbed wire fencing (in some cases it might double the establishment cost). An alternative solution is to plant thorn edges, but it has the drawback of prolonging the non-productive pre-development period. The IRR (internal rate of return) appears extremely sensitive to variations in investment costs and browse yields, especially when these are low. The IRR falls by 0.6 % when investment costs rise by 1 %. The IRR doubles when browse production moves from 500 to 1000 FU/ha and increases by almost 70 % when production rises from 1000 to 2000 FU/ha. If a level of 10 or even 15 % is identified as the minimum IRR, it appears that the profitability of browse plantations is assured at all production levels (except in the case of Opuntia spp. and Acacia cyanophylla) when investment costs do not include enclosures. Opuntia plantations appear to have the lowest IRR for a given level of browse production and do not appear to produce browse at a sufficiently low cost to justify the level of investment.

Forestry Versus Range Development Increasing demographic pressure has led to the situation where more and more marginal lands are used for agriculture, infringing on areas that were previously covered with rangelands or forests. This is particularly the case in the arid and semiarid regions of the world. As a consequence, conflicting situations arise with regard to whether priority should be given to range or forestry development. Forestry and range are considered by many people as being totally different and distinct disciplines. This feeling is reinforced by the fact that from the administrative point of view, there are usually distinct departments dealing with each of them. This is true not only at the ministry level in most of the developing countries, but also within the national and international cooperation agencies. Furthermore, land use planners almost never consider land use units where forestry and range development should be carried out jointly.

2002

A. Swenne

This is a most unfortunate situation because foresters and range managers have quite a lot in common. They both have to deal with the “left-overs” of agricultural development. In other words, they have to work in the most challenging conditions. This is why they usually are the most knowledgeable people with regard to understanding fragile ecosystems and integrating their various components. Trying to make the best out of degraded land and to curb erosion is the common lot of foresters and range managers, especially in the arid and semi-arid areas. Provision of fuelwood for the local populations and fodder for their livestock are two of the major issues that are frequently addressed by development projects. Because there is no proper integration of both in the planning stage, however, one issue very often receives priority over the other. This leads to situations where, for instance, a forestry project has to take into consideration some form of range management after a while because of the pressure of local herds that have less grazing land without an increase in their fodder supply. At that stage, it is very difficult for a range manager to produce an effective management system, if only because foresters in the dry areas tend to plant mixed stands of trees and/or shrubs with different management requirements. Furthermore, in stands where trees bear fruit of high value for livestock (e.g. Prosopis spp. or Acacia spp.), the maximum yields are reached at a time when fuelwood should be collected as well. Even at the planning stage, foresters should avoid the temptation to recreate climax vegetation at all costs. As stated earlier, it is quite difficult to define exactly what climax vegetation should look like, this is particularly true in areas where extensive soil degradation has taken place because of erosion. Range managers almost never deal with climax vegetation because it is not the most productive from the livestock production point of view. Another problem is that foresters are usually reluctant to let any livestock graze or browse in reforested areas. This attitude is quite legitimate in young plantations. It is less so in pre-mature or mature plantations. There is still a strong feeling among many foresters that livestock is detrimental to tree growth. That feeling comes close to revulsion when goats are concerned, even though goats are mainly present in the driest and most degraded areas. Goats are blamed for destroying whole forests and accelerating the desertification process. There was a time when pictures showing goats climbing on trees were quite common in some publications. There is a growing number of scientists, however, that believe that mismanagement is the major cause of forest degradation by livestock. As a consequence, during the past years, integration of animal and forest production has been a growing concern among some foresters as well as range managers. From the forester point of view, livestock can be useful in controlling the growth of the understory to obtain better timber production, to reduce the risk of fire or to accelerate the cycling of nutrients.

Range Development in the Tropics Rangelands represent an important resource in many countries around the world. About 30–40 million people in arid and semi-arid regions have “animal-based” economies. Over 50 % of these people live in Africa, and they are commonly

Range Management

2003

referred to as “pastoralists” (Sandford 1983). They derive most of their income and sustenance from livestock grazing in arid and semiarid areas. In the Sahel, for example, large portions of the population are directly dependent on their livestock for food and cash to purchase alternative food and other necessities (Simpson and Evangelou 1984). Rangelands in many sub-tropical and tropical areas are being stressed as animal numbers expand to meet a growing human population dependent on a shrinking base. As a consequence, many countries have experienced destructive grazing to varying degrees. In many cases, range deterioration is most pronounced around watering points and other areas of livestock concentration. The primary reason for this situation is overstocking due to an excessive number of livestock or a decrease in available grazing areas or a combination of both. A growing number of experts, however, emphasize the importance of the human factor. According to Harrington (1981), for instance, analysis of the reasons for the widespread degeneration of arid zone pastures in both developed and traditional economies indicates that management decisions are strongly influenced by sociological and economic factors, that condition of the animals is often secondary and that condition of the pasture is rarely given any consideration at all. Only when pastures fail to recover after a drought is concern for the plant life exhibited by arid zone inhabitants. Such damage is not repaired in a human life span and thus there is little motivation to make the necessary effort and sacrifices. Arid zone plant life is adapted to drought, but the perennial plants are at their most vulnerable in such conditions and can suffer permanent damage if overutilized by stock. The breeds of cattle, sheep and goats common in the arid zone are also adapted to drought. They can move long distances away from droughtstricken areas, subsist on poor quality dead plant matter, suffer over 25 % loss in body weight and show compensatory gain when better conditions prevail. The worldwide downward trend in condition of arid zone pastures is evidence that it is man who is ill adapted to the arid environment, and not the animals and plants on which he depends. The reasons for heavy stocking are many and complicated.

Optimum Stocking Rate Determining optimum stocking rate is complicated by extreme temporal and spatial variability in herbage and browse production on rangelands. Coping with these variations is a challenge to the most progressive operator as well as the nomadic pastoralist in developing countries. Providing flexibility in livestock operations to meet these variations is difficult under the best conditions. In most range areas, precipitation patterns are unimodal, with a fairly short rainy period with limited opportunities for plant growth. Consequently, at the end of the dry season and early in the rainy season, livestock are under maximum stress from the standpoint of intake and nutritive quality of their diet. In many cases, pastoralists have little flexibility in adjusting to these stress conditions and must suffer outright mortality of their animals.

2004

A. Swenne

When droughts occur, the situation is even worse. The inhabitants of the arid zone are relatively deprived of normal standards of living. In an age of increasingly effective means of communication, these relative disadvantages are more apparent and are affecting the traditional strategies for surviving in the arid zone. Governmental influence in nomadic and transhumant societies, often reinforced by the people’s changing aspirations, is encouraging a more settled way of life. This intensifies grazing pressure in the settled areas and makes the people more prone to disaster.

Land Tenure In most of the developing countries tenure rights are usually nil or customary only. Where grazing rights exist, they are usually secondary to any other usage and cropping unsuitable land has devastated the traditional grazings. This situation is exacerbated by increasing government control due to better communications, which favors the villager and attempts to constrain the nomad (Widstrand 1975). The primary type of land tenure for extensive range areas around the world is open grazing. Herders who graze on these unrestricted ranges are sometimes divided into three classes: – Nomadic. Herders who have no permanent base; they take all their provisions with them as they move with their livestock. – Transhumance. Herders who have a permanent base to which they return each year. They move with their livestock during certain parts of the year. – Sedentary. Often farmers who also raise stock on the side. They have a permanent home and graze livestock in the vicinity of their permanent base. It is easy to visualize nomadic grazing systems in arid and semiarid areas as a mechanism with maximum flexibility for herders to provide feed and water for livestock. Often, these patterns were seasonal, with movements away from permanent water during the rainy season and the reverse in the dry season. The impact of local droughts can be minimized by altering some of these movements to those areas less influenced by the drought. If the dry season extends longer than normal, herders can move to other areas in search of forage. These nomadic and seminomadic systems were well adapted for many rangelands around the world when human populations were relatively small and forage and water were plentiful for the livestock that supported the human population. As the human population increased, however, the number of livestock needed to support the human population also increased, putting additional stress on fragile vegetation and resulting in considerable range deterioration. Many Western range people consider communal grazing as one of the prime factors responsible for destructive grazing with excessive livestock (Hardin 1968); but not all writers agree with Hardin’s argument. Artz (1985), for example, argues that in most communal systems, there are various regulations within the group to control abuse. Shifts from communal grazing to private ownership have not

Range Management

2005

alleviated all problems of destructive grazing (Runge 1981; Sandford 1983). On the other hand, Gilles and Jamtgaard (1982) have listed examples in Peru and Africa where communal rangelands have been grazed for years without serious range deterioration. Other critical resources, under the control of a family group, grazing association, and so on, may indirectly control grazing pressure (Gilles and Jamtgaard 1982).

Socio-Economic Environment The keeping of large numbers of livestock as signs of prestige and wealth has often been blamed for being a major cause of overstocking without any productive purposes. This is true in some societies, but it would be a mistake to generalize to all the existing pastoral systems. In ancient pastoral economies, pastoralists do not hold their wealth “on the hoof” simply because of a social desire to have large herds. Their policy is a rational one because there is normally no alternative form of investment open to them. In many cases, offtake from the herds is determined by ecological constraints and increased prices do not change the numbers of male animals sold for slaughter (Wilson and Clarke 1976). Granting such pastoralists security of land tenure will not cure the overgrazing problem unless alternative forms of investment are also provided. On the other hand, no rational pastoralist will reduce his breeding herd if he is grazing communal land unless he is guaranteed that his neighbors will do likewise. Organized destocking would actually increase both animal productivity and the number of people able to live in some nomadically grazed areas. Animals require maintenance rations before they can produce a surplus. Thus the greater the number of animals on a range, the greater the number of maintenance rations required, and, after a certain optimum, maintenance can only be supplied at the expense of production. On this basis, it has been calculated that if the Sahelian drought of the 1970s reduces animal populations in that region by 50 %, the production of meat and milk might actually double (USAID 1974).

Range Projects Many project analysts, when approached in private, confess their reluctance towards range development projects, going as far as to declare that they know very few cases where that kind of project has proved successful. Similar feelings are expressed by many range management experts although it is very often for different reasons. Range development projects as such are not very numerous. More often range is a component of broader-scale projects such as rural development, animal production, forestry development, watershed management or soil and water conservation. Animal production projects usually have three components: animal health, animal production and range management. The animal health program aims at reducing the losses due to endemic diseases by implementing vaccination campaigns as well as

2006

A. Swenne

improving productivity by carrying out treatments for internal and external parasites. The main objective of the animal production component is to improve animal husbandry systems. In principle, better health and productivity of livestock should lead to a decrease in livestock numbers, especially where overgrazing is a major problem. This is rarely the case, and, consequently, any improvement of range production is short lived.

Conclusion The immediate prospect of reducing destructive grazing in sub-tropical and tropical countries is not bright. Heavy stocking and deterioration of basic resources remain one of the most serious problems facing the range livestock sector worldwide. With a few exceptions, the arid zone grazing societies are hinterlands of cities or humid agricultural societies and the economy and politics are geared to the advantage of the latter (Stamp 1961; Perry 1970). If the arid zone is to continue to be productive, the pastures to maintain health, and social conflict to be avoided, it is essential that adequate social and economic studies be undertaken, that social services be provided and prices for arid zone products be commensurate with their value to the national economy. Range development projects should be carefully planned and designed, and realistic goals should be set. In the planning phase, all factors should be taken into account, and integration at the regional and even national level is a prerequisite. Careful analysis of the socio-economic conditions should have priority over technical approaches if lasting success is hoped for. Remedies for problems relating to excessive livestock numbers on rangelands are complex and not easily resolved. Continued exploitation of the basic range resources, however, may lead eventually to the elimination of livestock production for thousands of nomadic and seminomadic people as well as to a dramatic reduction in the availability of animal products.

References Abdalla SH (1980) Application of simulation techniques to evaluate grazing management policies in the semidesert grasslands of southern New Mexico. PhD thesis, New Mexico State University, Las Cruces Acocks JPH (1966) Non-selective grazing as a means of veld reclamation. Proc Annu Grassl Soc South Afr 1:33–39 Adamson RS (1938) The vegetation of South Africa. British Empire Vegetation Committee, Kew, London Agregeda O, Cuany RL (1962) Efectos fotoperio´dicos y fecha de floracio´n en jaragua (Hyparrhenia rufa). Turrialba 12:146–149 Artz NE (1985) Must communal grazing lead to tragedy? In: White LD, Tiedeman JA (eds) Proceedings of the 1985 International Rangeland Resources Development Symposium. Cooperative Extension, Department of Forestry and J Range Management, Washington State University, Pullman

Range Management

2007

Barton H, McCully HM, Box T, Box JE (1966) Influence of soil compaction on emergence and first-year growth of seeded grasses. J Range Manage 19(3):118–121 Blasco F (1970) Montagnes du Sud de l’Inde: savannes, forets, ecologie. Trav Sect Sci Tech lnst Fr Pondichery 10:77–103 Blaser RE, Brown RH, Briant HT (1966) The relationship between carbohydrate accumulation and growth of grasses under different microclimates. In: Proceedings of the 10th International Grassland Congress, Helsinki, pp 148–150 Brown D (1954) Methods of surveying and measuring vegetation, Commonwealth Bureau of Pastures and Field Crops, Bulletin, 42. Commonwealth Agricultural Bureaux, Farnham Royal Burton GW, Jackson JE, Knox FE (1959) The influence of light reduction upon the production, persistence and chemical composition of Coastal bermudagrass, Cynodon dactylon. Agron J 51:537–542 Butterworth MH (1964) The digestible energy content of some tropical forages. J Agric Sci 63:319–322 Cates RG, Rhoades DF (1977) Patterns in the production of anti-herbivore chemical defenses in plant communities. Biochem Syst Ecol 5:185–194 Clark FE, Paul EA (1970) The microflora of grassland. Adv Agron 22:375–435 Clements FF (1916) Plant succession: an analysis of the development of vegetation. Carnegie Inst Wash Publ 54:53–60 Cole MM (1963) Vegetation nomenclature and classification with particular reference to the savannas. S Afr Geogr J 45:3–14 Conseil Scientifique pour I’Afrique (CSA) (1956) CSA specialist meeting on phytogeography: Yangambi, vol 22, London, 28th July – 8th Aug 1956 Cooper CF (1959) Cover vs. density. J Range Manage 12:215 Cordoba FJ, Wallace JD, Pieper RD (1978) Forage intake by grazing livestock: a review. J Range Manage 31:430–438 Crowder LV, Chheda HR (1982) Tropical grassland husbandry, Agricultural series. Longman, London Daubenmire RF (1958) A canopy-coverage method of vegetational analysis. Northwest Sci 53:43–64 Davidson JL, Milthorpe FL (1965a) Carbohydrate reserves in regrowth of cocksfoot (Dactylis glomerata L.). J Br Grassl Soc 20:15–18 Davidson JL, Milthorpe FL (1965b) The effect of temperature on the growth of cocksfoot (Dactylis glomerata L.). Ann Bot (Lond) NS 29:407–417 Davies W (1960) The grass crop- it’s development, use and maintenance, 2nd edn. Spon, London De Mongolfier C, Le Houerou HN (1980) Study on the economic viability of browse plantations in Africa. In: Le Houerou HN (ed) Browse in Africa. The current state of knowledge. ILCA, Addis Ababa, pp 449–464 Dix RL (1961) An application of the point-centered quarter method to the sampling of grass vegetation. J Range Manage 14:63–69 Dougall HW, Bogdan AV (1958) Browse plants of Kenya with special reference to those occurring in South Baringo. East Afr Agric For J 23:236–245 Duvall VL, Blair RM (1963) Terminology and definitions. In: Range research methods. US Department of Agriculture, Forest Service, Miscellaneous publication, vol 94. pp 33–35 Dyksterhuis EJ (1949) Condition and management of rangeland based on quantitative ecology. J Range Manage 2:104–115 Egler FE (1954) Vegetation science concepts. I. Initial floristic composition- a factor in old field vegetation development. Vegetation 4:412–417 Evans RA, Love RM (1957) The step point method of sampling- a practical tool in range research. J Range Manage 10:208–212 Everist SL (1958) Our best fodder trees. Queensland Agric J 84:581–582 Everist SL (1969) Use of fodder trees and shrubs, vol 44. Queensland Department of Primary Industries, Division of Plant Industry, Brisbane, pp 21–22

2008

A. Swenne

Everist SL (1972) Australia. In: McKell CM, Pershing J, Goodin JR (eds) Wildland shrubs, their biology and utilization. USDA Forest Service General Technical Report INT-1 FAO (1989) Production yearbook, vol 43. FAO, Rome FAO (2013) FAO statistical yearbook 2013. FAO, Rome FAO/UNEP (1983) EMASAR Phase II. Notes on trees and shrubs in arid and semi-arid regions. FAO, Rome Franclet A, Le Houerou HN (1971) Les Atriplex en Tunisie et en Afrique du Nord. FO: SF/TUN 11, Rapp Tech 7. FAO, Rome Garrison GA (1953) Effects of clipping on some range shrubs. J Range Manage 6:309–317 Gasto J, Contreras D (1972) Análisis del potencial pratense de Fanerofitas y camefitas en regiones mediterráneas de pluviometría limitada. Bull Tech 35. Fac de Agron, Univ de Chile, Santiago Gibbens RP et al (1986) Some impacts of 2,4,5-T on a mesquite duneland ecosystem in Southern New Mexico. A synthesis. J Range Manage 39:320–326 Gilles JL, Jamtgaard K (1982) The commons revisited. Rangelands 4:51–54 Glover PE, Stewart JT, Gwynne MD (1966) Masai and Kipsigis notes on East African plants. Part I. Grazing, browse, animal associated and poisonous plants. East Afr Agric For J 32:184–191 Goodloe S (1969) Short duration grazing in Rhodesia. J Range Manage 22:369–373 Gray SG (1970) The place of trees and shrubs as sources of forage in tropical and subtropical pastures. Trop Grassl 4(1):57–62 Hamrouni A, Sarson M (1974) Valeur alimentaire de certaines plantes spontanees ou introduites en Tunisie. Note de recherche 2. Inst Nat Rech Forest, Tunis Hardin G (1968) The tragedy of the commons. Science 162:1243–1248 Harrington GN (1981) Grazing arid and semi-arid pastures. In: Morley FHW (ed) Grazing animals, World animal science. Bl. Elsevier, Amsterdam, pp 181–201 Heady HF (1960) Range management in East Africa. Govt Printer, Nairobi Heady HF (1973) Structure and function of climax. In: Arid shrublands. Proceedings of the 3rd workshop, US/Australia Rangelands Panel. Society of Range Management, Denver, pp 142–143 Henzell EF, Norris DO (1962) Processes by which nitrogen is added to the soil/plant system. In: A review of nitrogen in the tropics with particular reference to pastures. Commonwealth Bureau of Pastures and Field Crops Bulletin, vol 46. pp 1–18 Heyligers PC (1965) Vegetation and ecology of the Port Moresby-Kairuku area. In: Lands of the Port Moresby, Kairuku area, Papua-New Guinea, vol 14, Land research series. CSIRO, Melbourne, pp 7–8 Hilder EJ (1964) The distribution of plant nutrients by sheep at pasture. Proc Aust Soc Anim Prod 5:241–248 Holechek JL, Pieper RD, Herbel CH (1989) Range management. Principles and practices. Prentice Hall, Englewood Cliffs Holmes CH (1946) Grasslands and their afforestation in Ceylon. Indian Forester 72:6–11 Howell LN (1978) Development of multi-camp grazing systems in the Southern Orange Free State, Republic of South Africa. J Range Manage 31:459–461 Humphrey RR (1949) Field comments on the range condition method of forage inventory. J Range Manage 2:1–10 Huschle G, Hironaka M (1980) Classification and ordination of plant communities. J Range Manage 33:179–182 Hyder DN, Conrad CE, Tueller PT, Calvin LD, Poulton CE, Sneva FA (1963) Frequency sampling in sagebrush-bunchgrass vegetation. Ecology 44:740–746 Ingalls LR, Thomas JW, Benne EJ, Tessar M (1965) Comparative response of wether lambs to several cutting of alfalfa, birdsfoot trefoil, bromegrass and reed canarygrass. J Anim Sci 24:1159–1164 Kadambi K (1963) Useful fodder trees and grasses for cultivation in Ghana. Ghana Farmer 7:75–80

Range Management

2009

Keay RWJ (1959) Vegetation map of Africa south of the Tropic of Cancer. Oxford University Press, London Klipple GE, Costello DF (1960) Vegetation and cattle responses to different intensities of grazing on short-grass ranges on the central Great Plains. U.S. Department of Agriculture Technical bulletin, vol 1216 Kramer PJ, Kozlowski TI (1960) Physiology of trees. McGraw-Hill, New York Krueger WC, Laycock WA, Price DA (1974) Relationship of taste, smell, sight and touch to forage selection. J Range Manage 27:258–262 Kydd DD (1966) The effect of intensive sheep stocking over a five-year period on the development and production of the sward. Sward structure and botanical composition. J Br Grassl Soc 21:284–288 Lawton RM (1968) The value of browse in the dry tropics. East Afr Agric For J 33:227–230 Lay DW (1965) Effects of periodic clipping on yield of some common browse species. J Range Manage 18:181–184 Le Houerou HN (1975) Report on a consultation mission to the Range Organization of Iran. AGPC, Mise FAO, Rome Le Houerou HN (ed) (1980) Browse in Africa- the current state of knowledge. ILCA, Addis Ababa Lewis JK (1969) Range management viewed in the ecosystem framework. In: Van Dyne GM (ed) The ecosystem concept in natural resource management. Academic, New York, pp 88–91 Mannetje L’t (1965) The effect of photoperiod on flowering, growth habit and dry matter production in four species of the genus Stylosanthes. Aust J Agric Res 16:767–771 Mannetje L’t (1978) Measurement of grassland vegetation and animal production. CAB, Farnham Royal McKell CM, Whalley RD, Brown V (1966) Yield, survival, and carbohydrate reserve of Harding grass in relation to herbage removal. J Range Manage 19:86–89 Minson DJ, Milford R (1966) The energy value and nutritive value indices of Digitaria decumbens, Sorghum almum and Phaseolus atropurpureus. Aust J Agric Res 17:411–423 Monjauze A, Le Houerou HN (1965) Le role des Opuntia dans l’economie agricole Nord Africaine. Bull Ec Nat Super Agric Tunis 8–9:85–164 Moore CWE (1964) Distribution of grasslands. In: Barnard C (ed) Grasses and grasslands. Macmillan, London, pp 182–203 Naveh Z (1966) Range research and development in the dry tropics with special reference to East Africa. Herb Abstr 36:77–85 Odberg FO, Francis-Smith K (1977) Studies on the formation of ungrazed elimination areas in fields used by horses. Appl Anim Ethol 3:27–34 Ojima K, Isawa T (1968) The variation of carbohydrates in various species of grasses and legumes. Can J Bot 46:1507–1511 Owensby CE (1973) Modified step-point system for botanical composition and basal cover estimates. J Range Manage 26:302–303 Paris OH (1969) The function of soil fauna in grassland ecosystems. In: Dix RL, Beidleman RG (eds) The grassland ecosystem: a preliminary synthesis, Range science series, 2. Colorado State University, Ft Collins Parker KW, Woodhead PV (1944) What’s your range condition? Am Cattle Producer 26(6) Pechanec JF, Pickford GD (1937) A weight-estimate method for determination of range or pasture production. J Am Soc Agron 29:894–904 Perry RA (1970) Productivity of arid Australia. In: Dregne H (ed) Arid lands in transition. American Association for the Advancement of Science Publication, vol 90. pp 303–316 Phillips J (1959) Agriculture and ecology in Africa. Faber and Faber, London Pieper RD (1978) Measurement techniques for herbaceous and shrubby vegetation. New Mexico State University, Las Cruces Pieper RD, Dwyer DD, Banner RE (1974) Primary shoot production of blue grama grassland in southcentral New Mexico under two soil nitrogen levels. Southwest Nat 20:293–302

2010

A. Swenne

Piot J (1980) Management and utilization methods for lignous forages: natural stands and artificial plantations. In: Le Houerou HN (ed) Browse in Africa. The current state of knowledge. ILCA, Addis Ababa, pp 339–350 Poppi DP, Minson DJ, Ternouth JH (1981) Studies of cattle and sheep eating leaf and stem fractions of grass. Aust J Agric Res 32:99–137 Pratt DJ, Greenway PJ, Gwynne MD (1966) A classification of East African rangeland, with an appendix on terminology. J Appl Ecol 3:369–382 Rains AB (1963) Grassland research in Northern Nigeria, Samaru miscellaneous paper, 1. IAR, Zaria Rattray JM (1960) The grass cover of Africa, vol 49, FAO, Agricultural studies. FAO, Rome Raymond WF (1969) The nutritive value of forage crops. Adv Agron 21:1 Risser PG (1984) Methods for inventory and monitoring of vegetation, litter, and soil surface condition. In: National Research Council/National Academy of Sciences (ed) Developing strategies for rangeland management. Westview Press, Boulder Rose Innes R, Mabey GL (1964) Studies on browse plants in Ghana. I. Chemical composition. Emp J Exp Agric 32:115–124 Roseveare GM (1948) The grasslands of Latin America. Imp Bur Pasture Field Crops Bull 36:56–69 Rossiter RC (1966) Ecology of the Mediterranean annual pasture type. Adv Agron 18:1–56 Runge CF (1981) Common property externalities: isolation, assurance, and resource depletion in a traditional grazing context. Am J Agric Econ 63:595–606 Sandford S (1983) Management of pastoral development in the third world. Wiley, Chichester Shaw NH, Bryan WW (eds) (1976) Tropical pasture research: principles and methods, Commonwealth Bureau of Pastures and Field Crops Bulletin, 51. Commonwealth Agricultural Bureaux (CAB), Farnham Royal Shoop MC, McIlvain EH (1963) The micro-unit forage inventory method. J Range Manage 16:172–179 Simpson JR, Evangelou P (eds) (1984) Livestock development in subsaharan Africa. Constraints, prospects, policy. Westview Press, Boulder Simpson JR, Stobbs TH (1981) Nitrogen supply and animal production from pastures. In: Morley FHW (ed) Grazing animals, World animal science. Bl. Elsevier, Amsterdam, pp 261–287 Sims PL, Singh JS (1978) The structure and function of ten western North American grasslands. Abiotic and vegetational characteristics. J Ecol 66:251–258 Smith D (1972) Carbohydrate reserves in grasses. In: Younger VB, McKell CM (eds) The biology and utilization of grasses. Academic, New York, pp 318–333 Snaydon RW (1981) The ecology of grazed pastures. In: Morley FHW (ed) Grazing animals, World animal science. Elsevier, Amsterdam, pp 13–31 Soriano A (1972) South America. In: McKell et al (eds) Wildland shrubs, their biology and utilization. USDA Forest Service General Technical Report INT-1 Stamp LD (1961) A history of land use in arid regions, vol XVII, Arid zone research. UNESCO, Paris Stechman JV, Laude HM (1962) Reproductive potential of four annual range grasses as influenced by season of clipping or grazing. J Range Manage 15:98–103 Stem WR, Donald CM (1962) Light relationships in grass-clover swards. Aust J Agric Res 13:599–614 Stewart G, Hutchings SS (1936) The point-observation-plot (square-foot-density) method of vegetation survey. J Am Soc Agron 28:714–722 Stoddart LA (1946) Some physical and chemical responses of Agropyron spicatum to herbage removal at various seasons. Utah Agricultural Experiment Station Bulletin Stoddart LA, Smith AD, Box TW (1975) Range management, 3rd edn. McGraw-Hill, New York Talbot LM (1962) Food preferences of some East African ungulates. East Afr Agric For J 27:131–138

Range Management

2011

Talbot LM, Talbot MH (1963) The wildebeest in Western Masailand, East Africa. Wildl Monogr 12:27–45 Taylor MS (1980) Initial performance of Leucaena at a sub-humid mid-altitude location in Ethiopia. In: Le Houerou HN (ed) Browse in Africa. The current state of knowledge. ILCA, Addis Ababa, pp 415–418 Theron EP, Booysen PV (1966) Palatability in grasses. Proc Grassl Soc South Afr 1:111–120 UNESCO (1973) International classification and mapping of vegetation. Ecology and conservation series, 6 UNESCO (1979) Tropical grazing land ecosystems. Natural resources research. UNESCO/UNEP/ FAO, Paris USAID (1974) An approach to the recovery and stabilization of the Sahelian/Sudanian range and livestock industry. Technical Staff Paper AID/AFR/CWR Vallis I, Henzell EF, Marin AE, Ross PJ (1968) Isotopic studies on the uptake of nitrogen by pasture plants. IV. Uptake of nitrogen from labelled plant material by Rhodes grass and Siratro. Aust J Agric Res 19:65–77 Van Dyne GM (1966) Ecosystems, systems ecology, and systems ecologists. ORNL-3957. Oak Ridge National Laboratory, Oak Ridge Van Rensburg HJ (1948) Notes on some browse plants. East Afr Agric J 13:164–166 Vickery PJ (1981) Pasture growth under grazing. In: Morley FHW (ed) Grazing animals, World animal science. Elsevier, Amsterdam, pp 55–77 West O (1950) Indigenous tree crops for southern Rhodesia. Rhod Agric J 47:214–217 West O (1955) Veld management in the dry, summer-rainfall bushveld. In: Meredith D (ed) The grasses and pastures of South Africa, Pretoria, South Africa. pp 624–636 Whyte RO (1947) The use and misuse of shrubs and trees as fodder. Joint publication Commonwealth Agricultural Bureaux, 10. Hurley Widstrand GG (1975) The rationale of nomad economy. Ambio 4:146–153 Williams RE et al (1968) Conservation, development and use of the world’s rangelands. J Range Manage 21:355–360 Wilson AD (1969) A review of browse in the nutrition of grazing animals. J Range Manage 22:23–28 Wilson JG, Brendon RM (1963) Nutritional value of some common cattle browse and fodder plants of Karamoja, Uganda. East Afr Agric For J 28:204–208 Wilson RT, Clarke SE (1976) Studies in the livestock of southern Darfur, Sudan. II. Production traits in cattle. Trop Anim Health Prod 8:47–57 Wolfe EC, Lazenby A (1973a) Grass-white clover relationships during pasture development. I. Effects of superphosphate. Aust J Exp Agric Anim Husb 13:567–574 Wolfe EC, Lazenby A (1973b) Grass-white clover relationships during pasture development. II. Effects of nitrogen fertilizer and superphosphate. Aust J Exp Agric Anim Husb 13:575–580 Wood JG, Williams RJ (1960) Vegetation. In: The Australian environment. CSIRO, Melbourne, pp 67–84 Young SA (1980) Phenological development and impact of season and intensity of Defoliation on Sporobolus flexuosus (Thurb.) and Bouteloua eriopoda (Torr.) Torr. PhD thesis, New Mexico State University, Las Cruces

Agroforestry: Essential for Sustainable and Climate-Smart Land Use? Reinhold G. Muschler

Contents Setting the Stage for Agroforestry: Lessons from Monocultures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agroforestry: Evolution, Definition, Practices, and Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evolution of Agroforestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Definition and Classification: What Is Agroforestry? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agroforestry Practices and Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Principal Agroforestry Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Improved Fallows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alley Cropping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Linear Tree Plantings: Windbreaks, Shelterbelts, and Living Fences . . . . . . . . . . . . . . . . . . . . Silvopastoral Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Taungya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plantation-Crop Combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Homegardens, Multistrata Systems, and Tree Gardens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practices for Soil Conservation and Watershed Rehabilitation . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Agroforestry Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Distribution of Agroforestry Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Roles and Potential of Agroforestry for Sustainable Land and Landscape Management . . . Requirements for Sustainable Land Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Benefits of Trees for Microclimate Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Benefits of Trees for Soil Fertility and Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . How and Where Does Agroforestry Work? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Constraints to the Success of Agroforestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plant Selection for Agroforestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tree Domestication in Agroforestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Crop Species for Agroforestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selection of Animal Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2015 2018 2018 2019 2021 2021 2025 2025 2028 2031 2033 2034 2036 2038 2040 2042 2044 2044 2045 2046 2049 2055 2056 2057 2060 2065

R.G. Muschler (*) Agroecology and Agrobiodiversity, Agroforestry and Sustainable Agriculture Program, Centro Agrono´mico Tropical de Investigacio´n y Ensen˜anza (CATIE), Turrialba, Costa Rica e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_300

2013

2014

R.G. Muschler

Selection and Use of Microbial Symbionts and Other Beneficial Soil Organisms . . . . . . . Lessons Learnt from Tree, Crop, and Animal Domestication: Widening the Search . . . . Information to be Added to Crop and Tree Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multifunctionality of Agroforestry: Climate-Smart Production, Protection, and Ecosystem Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Agroforestry for Biodiversity Conservation and Ecosystem Services . . . . . . . . . . . . . . . . . . . . Contributions of Agroforestry to Climate-Smart and Multifunctional Agriculture . . . . . . Design and Modeling of Agroforestry Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Designing for Agroecological Intensification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Factors for the Tree/Shade Decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Designing Agroforestry Systems for Ecological Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . Designing Agroforestry Systems for Nutrition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Training Materials for Promoting and Designing Agroforestry Systems . . . . . . . . . . . . . . . . . Modeling Agroforestry Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions and Recommendations: Priorities for Research and Development . . . . . . . . . . . . . Characterization of Crop and Tree Species: Expanding the Passport Information . . . . . . . Increasing the Scope for the Selection of Crops, Trees, and Animals . . . . . . . . . . . . . . . . . . . . Management of Beneficial Soil Fauna and Microorganisms for Soil Health and Fertility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Optimizing the System Design and Management for Maximum Resource Use Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Creating Climate-Smart and Pest-Suppressive Landscapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biophysical Research Methods and Improved Experimental Design . . . . . . . . . . . . . . . . . . . . . Socioeconomic Aspects: Paying for Externalities and Services . . . . . . . . . . . . . . . . . . . . . . . . . . Integrative and Cross-Disciplinary Work: Stitching it all Together at the Landscape Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2066 2068 2069 2070 2072 2075 2080 2080 2082 2082 2087 2087 2088 2090 2091 2092 2093 2094 2094 2095 2097 2099 2102

Abstract

Over the past four decades, a solid body of research has revealed the potential of agroforestry for increasing or maintaining system productivity while protecting natural resources and providing environmental services, including pollination, pest control/prevention, carbon sequestration, and the conservation of soil health, water quality, and biodiversity. Thus, agroforestry is well suited as a central tool for “sustainable intensification” within a land use paradigm that should be based, in alignment with a recent call by FAO, much more on biology and agroecology, rather than on chemistry and fossil fuels. With success stories from around the world and new methodological tools for valuing also environmental services, we can now apply these tools to design practices and systems that match the outputs of sustainable crop, tree, and animal agroforestry systems to the local needs. To custom-tailor the systems to the respective environmental and socioeconomic conditions, and rise to the challenge of sustainably producing more food that is less contaminated and less contaminating, we should advance in the following directions: (i) expand the species characterizations, (ii) widen the scope of plants and animals used and include “neglected and underutilized species” (NUS), (iii) intensify work on “using” beneficial soil organisms for soil and plant health, (iv) optimize the system design and management to maximize resource use efficiency and minimize pest incidence,

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2015

(v) create climate-smart and pest-suppressive landscapes, and, finally, (vi) advance toward more holistic socioeconomic assessments including an improved valuation of environmental services. A call is made to apply also relevant experiences from other fields such as biointensive or organic production, urban agriculture, and permaculture. Keywords

Adaptation • Agrobiodiversity • Agroecology • Agroforestry • Alley cropping • Animal husbandry • Biochar • Biodiversity conservation • Biological control • Biointensive agriculture • Bird-friendly coffee • Cacao • Carbon sequestration • Climate-smart landscape • Certification • Coffee • Design • Diversification • Ecosystem service • Environmental service • Food security • Homegarden • Improved fallow • Living fences • Microclimate management • Mitigation • Modeling • Monocultures • Multifunctional agriculture • Multipurpose trees • Mycorrhizae • Neglected and underutilized crops (NUS) • N-fixation • Nutrition • Nutrient cycling • Organic production • PES (payment for environmental services) • Pest suppressiveness • Plantation crops • Polycultures • Rehabilitation • Shade management • Shelterbelt • Silvopastoral system • Soil conservation • Soil degradation • Soil fertility • Soil health • Sustainability • Sustainable intensification • Symbiosis • Taungya • Training materials • Tree domestication • Urban agriculture • Windbreaks

Setting the Stage for Agroforestry: Lessons from Monocultures Since the dawn of agriculture, humans have used trees in combination with crops and livestock. Most early land use systems evolved under trees or in savannas, with more or less drastic modification of the tree component. With the evolution of agriculture and forestry as professional sciences since the 1800s, the production of crops and trees became increasingly separated, driven by the observation that their individual productivity could be maximized for most crops, at least in the short term, by excluding any other competing or harmful organism. As a consequence, monocultures of crops or trees began to expand and, ultimately, rise to predominance in many environments. Driven by the development of specialized machinery for planting and harvesting, plus the increasing use of agrochemicals to improve plant growth, this tendency accelerated greatly since the 1940s. For some time, impressive productivity gains often multiplied former yields, fueling the massive expansion of monocultures of corn, wheat, and soy in the vast plains of North America, substituting deep-rooted grasslands and forests. Elsewhere, particularly pronounced in Latin America, the Caribbean, and Asia, expanding monocultures of bananas, coffee, sugarcane, rubber, oil palm, rice, and soy began to engulf formerly biodiverse biomes (Kimbrell 2002). However, after only a few decades, first cracks started to appear in the “get big or get out” paradigm of ever larger monocultures able to provide apparently unlimited productivity gains. One of the earliest indicators was that soil erosion by water and

2016

R.G. Muschler

Fig. 1 Soil degradation as a global problem of the past and present. Top: a dust storm in Texas in 1935 and the result of the “Dust Bowl” in South Dakota, USA, in 1936. The elimination of vegetation, particularly perennial grasses, but also trees and shrubs, and years of extreme drought have left the soil unprotected and vulnerable to wind erosion. Bottom: plowing and lack of protection disposes volcanic soils for vegetable production in Costa Rica to massive soil erosion even in the present. The long-term result can be seen as degraded slopes stripped of all fertile topsoil in Haiti (Photo credits: top, Wikimedia Commons; bottom, R. Muschler)

wind became ever more prevalent, illustrated in its extreme by the US “Dust Bowl” in the 1930s (Fig. 1) followed by similar events. The response gave rise to the establishment of the US Soil Conservation Service and its homologues in many other countries. As an immediate measure in the USA, shelterbelts and windbreaks were established to curb the effects of wind and protect the soil. This was an early widespread application of agroforestry. Lesson learnt: unprotected soil can neither be stable nor productive in the long run. Up to 10 cm and more of unprotected soil can be lost in just a single year (Montgomery 2012). And losses of more than 1 cm per year, a multiple of the rates of soil formation, are common in many regions around the world today (Amundson et al. 2015), including carelessly plowed fertile volcanic lands in Costa Rica or poor calcareous soils in Haiti (Fig. 1, bottom). Unfortunately, replicas of these examples are almost ubiquitous in many topographically similar situations around the tropics. Some years later, on the chemical front, Rachel Carson’s Silent Spring provided 1962 an eye-opening account of the negative and long-term impact of depending on agrochemicals for pest and disease control. Although these products had been

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2017

hailed as the solution to almost all problems of plant nutrition and protection, the painful lessons about the bioaccumulation and long-term toxicity of persistent pesticides, plus their inevitable side effects like groundwater contamination and the loss of beneficial organisms such as insects, birds, or symbiotic soil microbes (Bardgett and van der Putten 2014; Stamets 2005; Kimbrell 2002), have shown that an overly dependence on agrochemicals and fossil-fuel-driven machinery is a deadend road (IASS 2015). Recent information about their effects on human health raise concerns only more (Leu 2014; Samsel and Seneff 2013). Clearly, even though this approach has been effective in some cases (ignoring its externalities) and for a short time, it cannot be sustained in the long run in the face of increasing prices of the external inputs and mounting evidence of their massive side effects on natural resources and human health (Colburn et al. 1997; Conway and Pretty 1991). The accelerating loss of biodiversity; the drastic decline of bee populations (Pilatic 2012); the contamination with fertilizers and pesticides of most water bodies, including the Gulf of Mexico and the great lakes of the USA; and the groundwater in so many regions of Europe, Asia, and Latin America provide ample evidence of this (IASS 2015; Kimbrell 2002). Obviously, the business-as-usual model cannot be sustained. Learning from past mistakes, acceptance is growing that the only truly sustainable long-term options depend on using an effective agroecological design for linking productive units embedded in a biodiverse and functional landscape (FAO 2013; Trumper et al. 2009). The judicious distribution and integration of appropriately sized patches of agricultural, agroforestry, and forestry systems allows maximizing agroecological benefits, both in terms of products and services (Fig. 2). The functional outputs of such a system include (1) the mutual stimulation of growth rates of compatible components, (2) the prevention and control of pests and diseases, (3) the protection and enrichment of soils, (4) the maintenance of water flow and quality, and (5) the support of (functional) biodiversity to provide essential products and services including pollination, biological pest control, and adaptation/ mitigation to climate change (Nicholls et al. 2013). Figure 2 illustrates how the spatial dimensions and distribution of agricultural, agroforestry, and forestry systems modify a landscape and, therefore, determine, to a large extent, the products and environmental services that can be provided. A treeless barren landscape cannot protect the soil or water resources. Beyond these geometric factors, obviously, also the density and arrangement of trees and crops; their characteristics, as well as their temporal sequence; and the interactions among all components contribute to mold the overall benefits that can be derived from such a landscape. Reversely, if the goal is rehabilitation of a degraded landscape like in Fig. 1, it is imperative to include a substantial number of trees. Ideally, the local site conditions will define the species and their best arrangements and management. In all cases, agroforestry, with its multiple products and services, is at the heart of sustainable landscape design (Nicholls et al. 2013; Trumper et al. 2009; Muschler and Bonnemann 1997). This chapter presents the potential of agroforestry as an essential component of truly sustainable climate-smart land use.

2018

R.G. Muschler

Design

Function

Agroforestry Products

AgroCulture, control, food security profitability

Crop fields, pasture

vs

Natural forest

vs

vs

Product value chains

Tree crops

Tree

ForesNatural vegetation, biodiversity, wilderness

Timber Food Fuelwood Fodder others

Services Soil/water conservation Biodiversity Adaptation/mitigation Scenic beauty others

Fig. 2 Multifunctionality of agroforestry. Agroforestry allows combining agricultural and forestry production on the same plot of land. Furthermore, it serves as a biological corridor to link agricultural and forest lands in the landscape. As indicated by the two-directional arrow on top, there is a two-way relationship between the “design” of the landscape mosaic and its “functions” in the form of products and services. While a given landscape design determines largely the functions that can be met, the opposite also holds true: the definition of expected functions determines largely the most appropriate design (Source: the mosaic graph is from van Nordwijk 2014)

Agroforestry: Evolution, Definition, Practices, and Systems Evolution of Agroforestry Although its practice dates back to the beginning of human land use, agroforestry, the purposeful combination of trees with crops and/or animals, emerged only in the 1970s as a professional field of its own. This new discipline emerged when it was realized that it held great potential for combining production with protection purposes. The official start for the evolution of agroforestry as a stand-alone discipline is commonly cited with the establishment in 1977 of the “International Council for Research in Agroforestry” (ICRAF), headquartered in Nairobi, as the global lead institution under the umbrella of the “Consultative Group for International Agricultural Research” (CGIAR). Interestingly, in Latin America, the beginning was marked at the same time by a series of German-funded projects at the Center for Research and Higher Education in Tropical Agriculture (CATIE), a spinoff institution of the “Inter-American Institute for Cooperation on Agriculture” (IICA), headquartered in Costa Rica. With German core funding from 1976 to 2003 for the study of Central American silvopastoral systems with cows, goats, and, later, agroforestry systems with cacao and coffee, CATIE converted itself into the pioneer institution that led the way of agroforestry development in Latin America (CATIE 2001, 1999; Beer 2000; Beer et al. 1987). In Africa and Asia, ICRAF has led the way. Today, ICRAF (now under the name “World Agroforestry Centre”),

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2019

with its global mandate, has a regional office at CATIE, and there is increasing cooperation between these two lead organizations. The initial geographic focus for the evolution of agroforestry was on the tropics, since the issues to be addressed there are particularly urgent. After the first two decades dedicated largely to definitions, descriptions, and characterizations of the different types of agroforestry (Nair 1993; MacDicken and Vergara 1990; Steppler and Nair 1987), the 1990s saw an increasing interest in studying functional aspects to improve the systems’ productivity (tree-crop interactions, tree selection, pruning regimes, shade management, input substitution, etc.; see Huxley 1999; Rao et al. 1998; Ong and Huxley 1996) and to explore the potential of agroforestry for addressing the increasingly more urgent issues linked to natural resource degradation and contamination, biodiversity loss, and the deterioration of family farms. Today, the solid research about the production and protection functions of agroforestry1 has established this field as a central pillar for truly sustainable land use systems, that is, systems, which can be sustained indefinitely. The outburst of scientific publications over the past 40 years on the many ecological, economic, and social benefits that agroforestry can provide in many environments around the world provide a solid justification for the well-chosen title of a recent global review “Agroforestry – the Future of Global Land Use” (Nair and Garrity 2012). In the USA, in recognition of the potential of agroforestry for enhancing agricultural landscapes, watersheds, and rural communities, the USDA released in 2011 its “Agroforestry Strategic Framework.” While agroforestry is not a magic bullet for solving all problems, it clearly provides many of the central tenets for developing productive and climate-smart agricultural systems which, at the same time, can provide essential environmental services for maintaining a healthy and sustainable landscape (FAO 2013). This takes us to the definition of agroforestry.

Definition and Classification: What Is Agroforestry? One of the shorter yet most comprehensive definitions of agroforestry states that it is “the intentional integration of trees and shrubs into crop and animal farming systems to create environmental, economic, and social benefits” (USDA). As such, it encompasses a wealth of different land use systems at the interface of agricultural, livestock, and forestry systems. A widely accepted structural classification of agroforestry systems is based on the type of components present in a system (Fig. 3; Nair 1993). According to this classification, there are four main groups that combine trees or woody shrubs with: • Annual or perennial crops (agrisilvicultural systems) • Pastures and grazing animals (silvopastoral systems)

1

For the roles of agroforestry for biodiversity conservation, see Schroth et al. (2004), and for the provision of ecosystem services, see Rapidel et al. (2011).

2020

R.G. Muschler

Fig. 3 Classification of agroforestry systems based on the type of components. The three main groups combine trees or woody shrubs, as the essential perennial component, either with (i) pastures and animals (silvopastoral systems; left photo, an example from the Dominican Republic), (ii) with annual or perennial crops (agrisilvicultural systems; right, coffee with timber and service trees), or (iii) all three (agrosilvopastoral systems). The forth segment (“other systems”) collects other systems such as combinations of fish or shrimp farming or bee keeping with trees or shrubs providing habitat or fodder. For details see text (Graph from Nair 1993; Photo credits: R. Muschler)

• Crops and animals (agrosilvopastoral systems) • Non-pastoral animal species (“other systems”) such as combinations of fish, shrimp, or bees with trees or shrubs providing habitat or fodder Besides this structural classification according to components and their products, other classifications are based on: • The tree cover in agricultural lands with a minimum of 10 % for some systems and higher values for others (Zomer et al. 2014) • The intended functions or roles of the systems, e.g., for soil protection, fodder production, or for providing microclimatic benefits as is the case of windbreaks • Socioeconomic aspects such as the intensity of management, the type and level of inputs, or whether the production is for commercial or subsistence purposes • Biogeographic conditions for which the systems are selected. For example, while a particular mix, arrangement, and management of species may be appropriate for dryland systems, these attributes would likely not be appropriate or sufficient for systems in the humid tropics. Similarly, the species mix and arrangement will differ depending on the desired products and biophysical attributes such as rainfall distribution and intensity, temperatures, topography, soil fertility, windiness, and even native pest and disease organisms. Details on these other classification systems can be found in Nair (1993) and in MacDicken and Vergara (1990). The latter authors have assembled the central aspects of agroforestry practices and systems in most of the major agroecological zones of the world, including the humid tropics, the semiarid tropics, tropical highlands, and temperate zones.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2021

Agroforestry Practices and Systems According to Nair (1993), an agroforestry “practice” refers to a “distinctive arrangement of components in space and time,” while an agroforestry “system” is a specific local example of a practice. As such, a system is characterized by the sitespecific selection of plant species, their arrangement, management, and socioeconomic performance. While there are only some 20 practices that can be found even in widely different settings around the world (see section “Principal Agroforestry Practices”), there is a multiple of site-specific systems. Table 1 summarizes the major agroforestry practices and their main characteristics. For detailed information about the diversity of agroforestry systems, the reader is referred to the agroforestry system inventory spearheaded by the World Agroforestry Centre (ICRAF) headquartered in Kenya. Its first comprehensive inventory was edited by Nair (1993), which still serves as a basic reference. This work has since then been complemented by ICRAF and agroforesters around the world, much of it published in “Agroforestry Systems,” the principal scientific journal in the field. Further aspects on the definitions of agroforestry practices have been provided by Sinclair (1999). The following sections provide a brief characterization of the principal agroforestry practices.

Principal Agroforestry Practices The combination of trees, crops, and animals predates the evolution of monocrop forestry or agriculture, which focuses largely on producing just a single or a few primary commodities. Until the Middle Ages, it was common even in Europe, as it still is in many parts of the world today, to thin or clear-cut forest patches, burn the slash, and, then, plant selected food crops and trees. As a result, more or less stable long-term arrangements of crops, trees, and, often, animals such as sheep, pigs, cattle, or fowl evolved with different emphases according to the local sociocultural context and markets. A particular arrangement of components, independent of the specific choice of locally adapted species, can be described under the term “agroforestry practice” (Nair 1993). Despite the evolution of monocultures of trees or crops as the ultimate simplification for ease of management and harvest, a trend that started in industrialized Europe and North America, followed by relatively sparsely populated developing countries with vast areas (e.g., Brazil, Argentina, Paraguay), a wealth of diverse and more complex systems have persisted in many places of the tropics (see section “Distribution of Agroforestry Systems”). The following sections characterize the most important and distinctive agroforestry practices. Due to the limitations of the present document, the list is not exhaustive and the accounts are as short and succinct as possible. For more details, the reader is referred to Nair (1993) and MacDicken and Vergara (1990) and to the journal “Agroforestry Systems.” Following the arguments of Leakey (1996) to view agroforestry practices as different stages along the continuum of succession, the listing starts with practices that correspond to “early-stage” (i.e., systems with

2022

R.G. Muschler

Table 1 Major agroforestry practices and their main characteristics (Modified from Nair 1993) Description/ Major components Agroforestry arrangement of (T = tree; C = crop; Agroecological practice components A = animal) adaptability Agrisilvicultural systems combining trees with crops (including shrubs and vines) 1. Improved fallow Woody species are T: fast-growing In shifting cultivation purposely planted and preferably areas left to grow during the leguminous “fallow phase” C: common agricultural crops 2. Taungya Temporary T: spp. for plantation No restrictions association of forestry agricultural crops C: common with young tree agricultural crops, plantations (usually preferably some until tree canopy shade tolerance closes) 3. Alley cropping Agricultural spp. are T: fast-growing, Subhumid to humid (hedgerow grown in the alleys leguminous, areas with high intercropping) between hedges of vigorous regrowth/ human population woody spp; coppicing pressure and fragile microzonal or strip (productive but easily C: common arrangement degradable) soils agricultural crops 4. Multipurpose trees on crop lands

5. Plantation crop combinations

6. Homegardens

Trees maintained with or without a systematic pattern on bunds, terraces, or plot/field boundaries (i) Integrated multistory (mixed, dense) mixtures of plantation crops (ii) Mixtures of plantation crops in alternate or other regular arrangement (iii) Shade trees over plantation crops (iv) Intercropping with agricultural crops Intimate, multistory combination of many tree and crop species around homesteads

Trees on bunds, terraces, and raisers,

T: multipurpose and fruit trees C: common agricultural crops

All ecozones, esp. in subsistence farming; often with animals

T: timber, fruit, and multipurpose service trees; plantation crops like coffee, cacao, coconut, rubber, black pepper, etc. C: usually shadetolerant spp., more preponderant in (iv) and (i)

Humid lowlands or humid/subhumid highlands; often smallholder subsistence system

T: often dominated by fruit trees, woody vines C: shade-tolerant vegetables and medicinal plants T: multipurpose and/or fruit trees,

All zones, particularly in areas of high population density

Sloping lands, reclamation of (continued)

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2023

Table 1 (continued) Agroforestry practice 7. Trees for soil conservation and reclamation

Description/ arrangement of components

Major components (T = tree; C = crop; A = animal)

sometimes with grass strips: trees on degraded or marginal lands and wastelands

often leguminous and pioneer species C: deep-rooted perennial crops and grasses T: combination of tall, erect, and short dense species; wind tolerant C: common agricultural crops T: firewood species C: common agricultural crops

8. Windbreaks and shelterbelts, living fences or hedges between fields

Trees and shrubs around farmland and plots

9. Fuelwood production

Interplanting firewood species on or around agricultural lands Multispecies, multilayer dense plant associations with no organized planting arrangement

10. Multilayer tree gardens and Forest farming

T: woody components of varying form and growth habits C: sometimes shadetolerant spp. 11. Forest farming Production of shadeT: typically species adapted crops under of commercial tree overstory interest C: shade-tolerant spp. such as ginseng or mushrooms Silvopastoral systems (trees + pasture and animals) 12. Trees on Trees scattered or in T: multipurpose, rangeland or systematic preferably legume or pastures arrangement on high fodder value; rangeland and occasionally highpastures value timber C: improved pasture species A: grazing and browsing animals 13. Fodder/protein Production of protein- T: leguminous fodder banks rich tree fodder on trees/shrubs farm and rangelands C: improved fodder for cut-and-carry species and/or fodder production legumes A: stabulated

Agroecological adaptability degraded, acid, alkali soils, and sand-dune stabilization

Wind-prone areas

All ecozones

Areas with fertile soils, labor availability, and high human pressure

Various

Extensive grazing areas

No limitations; particularly on steep lands

(continued)

2024

R.G. Muschler

Table 1 (continued) Major components (T = tree; C = crop; Agroecological A = animal) adaptability T: plantation crops Areas with relatively low population C: preferably improved fodder spp. pressure A: grazing animals Agrosilvopastoral (and other) systems (trees + crops + pasture and animals) 15. Homegardens Intimate, multistory T: often dominated All zones, particularly with animal combination of many by fruit trees, woody in areas of high production tree and crop species, vines population density plus animals around C: shade-tolerant homesteads vegetables and medicinal plants A: minor species, specialty markets 16. Multipurpose Woody hedges for T: fast-growing and Humid to subhumid woody hedgerows browsing and coppicing fodder areas with hilly and and riparian provision of mulch, shrubs and trees sloping terrain buffers green manure, soil C: agricultural crops conservation, etc. A: grazing and browsing animals 17. Apiculture with Melliferous trees and T: melliferous tree Various regions, trees on farmlands bees on farmlands species of particularly semiarid complementary to subhumid phenology rangelands remote from conventional C: bee-pollinated agriculture crops, e.g., cucurbits

Agroforestry practice 14. Plantation crops with pastures and animals

18. Aquaagroforestry

19. Multipurpose woodlots

Description/ arrangement of components Example: cattle under coconuts in Asia or the Caribbean

Trees lining fishponds (leaves and fruits serve as fish fodder). Example: chinampa system in Mexico Mix of species for various uses (fuelwood, fodder, soil protection, etc.)

A: bees T: trees and shrubs producing leaves and fruits for fish A: fish, shrimps, etc. T: multipurpose, leguminous, timber, fruit, etc. C: shade-tolerant crops A: various

Various

Various

young and relatively small or sparse woody components) and evolves toward practices of “late-stage” agroforestry succession with dominant and permanent woody components. For information on appropriate species, please see section “Plant Selection for Agroforestry” on plant selection for agroforestry.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2025

Improved Fallows The traditional use of the word “fallow” refers to a temporal sequence of crop or tree species on a given piece of land. In shifting cultivation, after a period of agricultural use of several crop cycles until soil fertility has declined or weeds, pests, and diseases have built up to a point of unacceptably low yields, the land is left “fallow” for an extended period of time. During this time, usually several years, the successional vegetation of noncultivated crops and trees rebuilds the ecological webs and the stocks of organic matter and nutrients in the soil, reconstructing, at least partly, the former soil fertility which allows for another few cycles of crop production. The term “improved” fallow refers to a system in which the fallow vegetation includes crop or tree species that have been selected and planted, or favored, by man in order to speed up the desired soil fertility recuperation (Buresh and Cooper 1999), a requirement due to the ever-increasing human population pressure on the limited land resources. Often, the “improvement” may also include management interventions such as the inoculation with microbial symbionts such as N-fixers or mycorrhizae in order to favor the plant establishment and growth (Stamets 2005). Depending on the degree of this human-lead biological intensification, the ultimate result is the transformation of a shifting cultivation system to one of permanent cultivation. Besides the ecological benefits mentioned, the selection criteria of the most appropriate plant species for improved fallows may also consider economic benefits derived from using products generated during the fallow period. These include tree products such as poles or firewood, medicinal plants, and even beans and other edible or useful products from annual plants established for soil improvement. Examples of improved fallow system include the enrichment planting of rattan by the Luangan Dayak people of Borneo, Casuarina species in Papua New Guinea, gum arabic in the Sahel (von Maydell 1986), or multipurpose fallow woodlots in the Philippines. From different parts of the world, examples of tree and crop species, many of which are N-fixers, include Acioa barteri, Anthonotha macrophylla, Alchornea cordifolia, Gliricidia sepium, Leucaena leucocephala, as well as annual or perennial climbers or scramblers in the genera of Mucuna, Pachyrrhizus, and others (Fig. 4). Together with taungya systems and alley cropping, improved fallows correspond to early successional stages in which the plant components are relatively young and have restricted dimensions of root and shoot systems. Details on these systems from Latin America were given by Kass and Somarriba (1999) and worldwide reviews were published by Buresh and Cooper (1999) and by Thurston (1997).

Alley Cropping Alley cropping, often also known as “hedgerow intercropping,” refers to the practice of growing crops between hedgerows of planted shrubs or trees, preferably

2026

R.G. Muschler

Fig. 4 Left: improved fallow in Talamanca, Costa Rica, with the leguminous yam bean (Pachyrhizus erosus) for soil enrichment with N and occasional volunteer plants of sesame and manioc. Right: intercropping system of manioc with perennial pigeon pea (Cajanus cajan) for soil improvement and bean production (Photo credits: R. Muschler; photos taken at “Finca Loroco,” Bribri, Costa Rica).

Fig. 5 Left: typical alley cropping system with hedges of leguminous trees or shrubs that are pruned to provide biomass to the crops grown in the alleys. This example from Cameroon shows maize growing between hedges of Leucaena leucocephala. Right: alley cropping is more successful in subhumid and humid environments. In arid and subarid environments with less than 200–300 mm rain, water competition by the trees reduces crop yields (Photo credit: pixgood. com; Graph redrawn from Nair 1993)

leguminous species to provide biologically fixed N to the cropping alleys (Kang et al. 1989; Kang and Wilson 1987; Fig. 5). According to Nair (2012), alley cropping may be practiced on up to 100 Mio ha globally. To reduce shading and root competition for the associated crops, while at the same time providing nutrients and soil protection to the alleys, the hedges are periodically pruned and the biomass is spread onto the alleys. Typically, the pruning period ranges from 2 to 6 months, depending on the intensity of pruning and the speed of growth of the hedges. Partial pruning allows for more frequent harvesting than pollarding, i.e., the pruning of all branches at the same time. In order to synchronize the availability of the nutrients as

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2027

they are being released during decomposition with the demand of the developing crop, it is essential to choose the right time of pruning, considering the speed of growth of the hedges and of the nutrient demands of the crops (Crews and Peoples 2005). For example, for growing maize, the pruned biomass should not release most of its nutrients until the time of maximum demand of the maize plants after 2–8 weeks of growth (Witt et al. 2009; Cadisch and Giller 1997). Alley cropping is one of the most widely researched agroforestry practices, largely due to (i) the linear arrangement of woody and annual components, (ii) the same age and size of the individual crop plants or trees, and (iii) the relatively small number of species involved. Having only two (one crop + one tree) species for the simplest system (like in Fig. 5), up to a handful of species in others, the “tree-crop interface,” i.e., the region where crops and trees interact, is relatively easy to define and study. In this setting, research on above- and belowground interactions becomes much more tractable than in systems with complex and irregular distribution patterns of a multitude of different-sized and differentaged plants, as is the case with species-rich systems such as homegardens. Where does alley cropping work best? In many parts of humid and subhumid Africa, where water limitations are not extremely strong, trials have shown a positive effect of the tree component on soil fertility and nutrient supply for the crops. For example, in an 8-year alley cropping trial on sandy soils in southern Nigeria, maize yield could be raised from 0.66 t/ha in control treatments to 2 t when Leucaena prunings were applied and even to over 3 t/ha, when an additional 80 kg N/ha was applied as supplement (Kang et al. 1989). Similar results were obtained from many other humid or subhumid areas in the tropics. For example, Kang and Duguma (1985) demonstrated that maize yields in 4 m wide alleys of Leucaena were the same as when 40 kg/ha of N were applied to the crop. In Malawi, after 11 years of intercropping Gliricidia sepium with maize, Makumba et al. (2006) concluded that the 4–5 Mg DM ha1 of Gliricidia prunings, when applied to the crop, increased maize yield threefold (3.9 Mg ha1) over sole maize cropping (1.1 Mg ha1). A fertilizer complement of 46 kg N ha1 increased maize yield by another 29 %. These results show that a significant portion of crop nutrient demand can be met by adding pruning residues of the mostly leguminous hedges (Kang et al. 1990). Often, a combination of biomass from the hedges and a fertilizer supplement produces the best results. Of course, the effects vary greatly with the site conditions, particularly soil fertility and water availability (Witt et al. 2009), as well as with the plant species used and the pruning regime applied. Besides the species attributes, these factors determine largely the quality of mulch in terms of its rate of decomposition and its nutrient content (Cadisch and Giller 1997; Budelman 1988, 1989). However, while these results are encouraging, the effects of alley cropping are not universally positive (Nair 1993; Szott et al. 1991; Kass 1987). For example, on acid soils in Yurimaguas, Peru, the yields of all crops in alleys, except for cowpea, were extremely low, often below that of the control plots. The main reasons for this were attributed to root competition and shading from the hedges, but also to reduced seedling emergence under the mulch applied to the alleys and, possibly, also to the

2028

R.G. Muschler

extremely low fertility of the soils which may have impeded the nutrient recycling expected of the woody hedges (Szott et al. 1991). Another well-studied aspect in alley cropping is competition for water between the crops and the trees. It turns out that the degree of water competition is key for determining the success of alley cropping systems (Fig. 5, right). As a result, under conditions of water limitations in semiarid environments in India, NW Nigeria, and Kenya, the hedgerow species actually reduced the crop yields due to competition for water (Corlett et al. 1989; Mittal and Singh 1989; Nair 1993). In some cases, even the biomass yields from the hedgerows were lower under alley cropping than from non-alley-cropped hedgerows. So, under severe moisture limitations, the competition for water between the trees and the crops can become too severe to allow alley cropping to work, at least on the score of production. The unfavorable performance of alley cropping under nutrient- or moisturelimited conditions demonstrates that this agroforestry practice, like others, requires certain minimal conditions that allow the tree-crop interactions to unfold their positive potential. On extremely acid soils in the Peruvian Amazon, some fertilizer may have to be supplied to allow cropping to work, and there is a great need to identify the most appropriate tree species for such conditions. Some examples are Inga edulis, I. felulei, Erythrina spp., Cassia reticulata, or Gliricidia sepium (Szott et al. 1991). In order to evaluate the potential of alley cropping under particular site conditions, it is important to not only consider the agronomic performance but also other benefits such as long-term contributions to soil fertility, biological interactions with pests and diseases, as well as limitations such as the costs of establishment and labor requirements for maintaining the systems (Kang et al. 1989). When animals are integrated into alley cropping, the system has been called alley farming (Kang et al. 1990), and the pruned biomass can be used as animal fodder. In these cases, the economic value of the hedgerow products is likely to augment, and the recycling of animal residues as soil amendments can improve the biophysical interactions in the systems. A short practical guide on alley cropping has been provided by Elevitch and Wilkinson (1999).

Linear Tree Plantings: Windbreaks, Shelterbelts, and Living Fences Around the world, linear tree plantings are important components in the landscape. Usually they are established to confine animals to patches of pastures, to separate plots, or to protect cropland from strong winds and resultant erosion. The latter was the case for the government-promoted establishment of windbreaks in the USA after the massive dust storms and soil displacement during the Dust Bowl. Windbreaks and shelterbelts, two expressions often used interchangeably, are widely used, protecting around 300 Mio ha of farmland (Nair 2012). In many tropical silvopastoral and agricultural systems, living fences are used to separate pasture areas and to protect agricultural plots from animals. These agroforestry practices share the two-dimensional arrangement of trees or shrubs planted in single or

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2029

Fig. 6 Left: windbreaks in crop fields in Oklahoma. Right: living fence of Gliricidia sepium to separate pastures in the north of the Dominican Republic (Photo credits: Oklahoma Forestry Service and R. Muschler)

narrow multiple row arrangements on farmland. Depending on their function and the environment, they differ in the width, the spacing between trees, and the species composition of the tree rows. With increasing fragmentation of tropical landscapes, windbreaks, living fences, and isolated trees become more and more important as “biological corridors” for the displacement and survival of many animal species (Harvey et al. 2004). Windbreaks/shelterbelts. The central function of windbreaks is to protect crops, animals, soils, water bodies, and infrastructure from strong winds and their negative effects (Mendez et al. 2000; Reifsnyder and Darnhofer 1989; Fig. 6). For example, in many semiarid savannas of Africa, hedgerows of Euphorbia tirucalli protect crop fields and settlements, and multispecies shelterbelts with drought-resistant species protect cropland from drying winds and, hence, desertification. Around the world, particularly in wind-exposed highlands, but also in coffee fields in the southern plains of Brazil, windbreaks of different-sized shrubs and trees protect vegetables, coffee, and other crops. In addition to the beneficial effects from reducing wind speed, these rows of trees and shrubs also generate important benefits for the protection and enrichment of soils and provide habitat for the local fauna and flora (Harvey et al. 2004). For maximum effectiveness of a windbreak, the different-sized plants should form a two-dimensional mosaic, which, ideally, should have a permeability of 20–50 % through the windbreak to avoid the formation of potentially negative turbulence on the leeward side of the windbreak (Stigter et al. 1989). The higher values should be in the bottom third. As a rule of thumb, the protective effect of a windbreak is commonly estimated to extend to a leeward distance of eight to ten times the height of the windbreak (Mendez et al. 2000; Geilfus 1994). Details on the biophysical effects of windbreaks and shelterbelts, and recommendations for their design, were given by Cleugh et al. (2002), Stigter et al. (1989), Brandle et al. (1988), and Dronen (1988). On the ecological side, it is interesting to note

2030

R.G. Muschler

that windbreaks, even when they consist of only a few species of trees, can be important for the (re-)establishment of forest trees. The visiting birds and other animals can bring in substantial amounts of seeds, which can find favorable conditions for germination in the understory of the windbreak (Harvey et al. 2004). For example, in a study at Monteverde, Costa Rica, the windbreaks (consisting of Montanoa guatemalensis, Cupressus lusitanica, Casuarina equisetifolia, and Croton niveus) received 2 times more species and 40 times more seeds than adjacent pastures (Harvey et al. 2004). When windbreaks were connected with adjacent forest, or less than 20–50 m from forests, the numbers of tree seedlings and species increased even further than for windbreaks at greater distances. In many places, the species used for windbreaks are also selected to provide additional products such as firewood, fodder (e.g., Prosopis), fruits (e.g., cashews), nectar, honey, and others. Occasionally, where grazing animals may be a problem, some unpalatable species such as neem (Azadirachta indica) can be included on the outer sides to protect the windbreak from animal damage. However, care must be taken that the species on the outside of the windbreak are compatible with the adjacent crops. This can be a problem, for example, for Azadirachta indica or species of Eucalyptus and Juglans that are known to reduce growth of nearby crops due to competition or allelopathy. The effects of windbreaks on crop yields vary greatly, from significant increases in the productivity of the protected crops to reductions due to competition by the trees, shading, and “loss” of agricultural land occupied by the shelterbelts. However, if one considers the integral benefits of these structures, including their long-term benefits for soil fertility, water protection, and habitat for biodiversity (including seed dispersers, pollinators, and agents of biological control), then the sum of all effects is likely to be positive in most cases. Furthermore, management interventions in the form of pruning and selective harvesting, as well as the inclusion of species that are more compatible with the crops, will help to tip the scale toward an overall positive balance. Living fences. In many agricultural and pastoral landscapes, individual plots of land are separated from each other by living fences, which, typically, consist of closely spaced trees in a single row that serve as supports for horizontal sticks, vines, or barbed wire (Fig. 6). In some cases, the trees are planted so closely together or are so tightly interplanted with spiny unpalatable plants that there is no need for wires. Budowski (1987) provided an overview of the practice of using living fences in the Neotropics. Ideally, the species used for living fences are trees which can be readily pruned or pollarded, usually two or three times a year, and which can be established easily from stakes. Furthermore, in order to not strain the wires, it is desirable that the trees do not have a particularly strong secondary growth of the stems. By using stakes of 2–2.5 m length that will be planted 20–30 cm deep, the emerging shoots at the top of the poles will be out of reach of most grazing animals, and the trees will be able to establish themselves in 2–3 months. In the Neotropics, some of the most widely used tree species include Erythrina berteroana, Gliricidia sepium, Diphysa robinioides, and Bursera simaruba. All of them can be easily reproduced and planted from stakes.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2031

An interesting side effect is the possibility of using living fences for the reintroduction of tree seedlings into pastures. Love et al. (2009) showed that the survival of timber tree seedlings (Cedrela odorata, Samanea saman, Tabebuia rosea, Pachira quinata) was much higher when these trees were planted within living fences rather than in the open pasture. Further information is available through the online resources of the “World Agroforestry Centre.”

Silvopastoral Systems According to FAOSTAT, the global area under pasture and fodder crops was about 3.5 billion ha in 2000, representing 26 % of the total land surface and about 70 % of the global agricultural land. This area supports some 360 million cattle and over 600 million sheep and goats. Globally, grazing animals are the principal source of livelihood for some 100 million people in arid areas, and probably a similar number in other zones, supplying about 10 % of beef and about 30 % of sheep and goat meat. Unfortunately, in many places, extensive cattle production has caused widespread environmental degradation due to the loss of soil plant cover from tree cutting, overgrazing, and soil compaction, causing massive erosion. In addition, the residues of the animals may contaminate waters, and ruminants contribute around 30 % of all agricultural emissions of greenhouse gases (equivalent to about 5 % of all GHG emissions), particularly in the form of CH4 (Houghton et al. 2001). Consequently, there is a great need for developing more sustainable and climatefriendly animal husbandry systems. One of the most promising options is the development of intense silvopastoral systems, which combine trees with pastures and livestock (Ibrahim et al. 2010; Pezo and Ibrahim 2001). In addition to the pasture grasses, silvopastoral systems include dispersed trees for fodder and shelter, as well as closely spaced trees for living fences, windbreaks, and fodder banks or alleys (Fig. 7). When possible, highly nutritious, often N-fixing, species are preferred (Pezo and Ibrahim 2001). If well chosen and appropriately managed, these additional tree components enhance nutrient cycling; benefit the pastures; provide complementary tree products in the form of fodder, timber, firewood, and other tree products; and improve animal productivity (Yamamoto et al. 2007). Of particular interest is the use of high-quality fodder shrubs, mostly legumes, planted at high densities (more than 10,000 plants ha1) in so-called protein banks, and the introduction of trees, palms, and improved pastures. Controlled rotational grazing, feed supplements, and a permanent water supply for the animals allow for higher stocking densities and increased production of milk and meat. Supplementing low-quality fodder with foliage of leguminous trees or shrubs, such as Erythrina poeppigiana or Gliricidia sepium, improves ruminant digestion and can increase milk yield relative to standard urea supplements (Camero et al. 2001), while reducing CH4 emissions by up to 75 % (Reid et al. 2004). Due to their higher structural and biological complexity than simple pasture systems, silvopastoral practices also benefit biodiversity. In fact, systems with a

2032

R.G. Muschler

Fig. 7 Top: trees in silvopastoral systems provide shade to pastures and animals and are often used as living fences, which also provide high-quality fodder, e.g., from Erythrina berteroana in Costa Rica (bottom left) or from Gliricidia sepium in the Dominican Republic (bottom center). Although Trichanthera gigantea (bottom right) is not a legume, its highly nutritious leaves and fast growth make it a highly valued species for fodder banks in humid climates (Photo credits: R. Muschler)

high number of trees in pastures and multistrata living fences can match the animal species richness of early secondary forest, and networks of living fences in pastures are key for landscape connectivity (Francesconi et al. 2011; Harvey et al. 2006). Furthermore, these systems can also fix significant amounts of carbon (Ibrahim et al. 2007; see also section “Multifunctionality of Agroforestry: Climate-Smart Production, Protection and Ecosystem Services”), while augmenting water infiltration and improving water quality compared to traditional pastures (Rı´os et al. 2007). Given their local adaptation and ecological values for wildlife, it is often preferable to use native trees and shrubs for rehabilitating overgrazed and degraded lands (Murgueitio et al. 2011). In order to motivate farmers to plant or retain a broad range of tree species in pasturelands in Panama, Garen et al. (2010) pointed to the importance of matching appropriate tree species to the individual site conditions and the farmers’ production goals. The main limitations for a more widespread adoption of these systems are difficulties to establish trees or improved pasture species in or around pasture areas, insufficient access to water, as well as lack of capital and high labor costs for establishing and managing the systems to maintain appropriate light levels and fencing (Dagang and Nair 2003). However, as demonstrated in Nicaragua, Costa Rica, and Colombia, when farmers received “payments for environmental services” for including more trees in their systems (from 82 to 92 USD ha1), many of them

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2033

were able to transform their systems into improved silvopastoral systems with economic and environmental benefits (Ibrahim et al. 2010; see Box 3). Similarly, in dry African savannas, even modest payments of carbon credits could contribute to retaining the carbon in undisturbed savannas (Reid et al. 2004). Considering the societal benefits generated by improved silvopastoral systems (soil retention, water quality, biodiversity, etc.), a more widespread application of such compensation schemes should help promote the climate-smart transformation of this sector. Another promising approach is the development and application of environmentfriendly certification programs such as the “standard for sustainable beef production” developed by the “Sustainable Agriculture Network” (www.san.ag) in 2010. This standard promotes practices with ecological, social, and economic benefits. As points of departure on silvopastoral systems, practitioners may want to consult the following training materials: for Latin America, Pezo and Ibrahim (2001) produced a comprehensive training material on silvopastoral systems, and the work of Benavides (1994) on trees and shrubs as animal fodder continues to be a reference. For Africa, Wambugo et al. (2006) assembled a practical extension manual on fodder shrubs for dairy farmers. Further materials are available through ICRAF and partners.

Taungya Taungya systems, derived from the Burmese words taung (hill) and ya (cultivation), refer to the practice of intercropping young forest stands with agricultural crops as long as the trees permit adequate crop growth (Jordan et al. 1992). Once the trees have grown enough to establish crown closure and larger root systems, usually after 3–5 years depending on the tree species, planting arrangement, and growth conditions, the conditions for the agricultural crops become marginal. The integration of the crops allows deriving early income at a time when the trees are still too young to yield products, and the crop management, particularly weeding and fertilization, will also benefit the trees. This system, originally developed in British colonial India in the 1850s, is now widely used in many parts of the tropics to establish forest plantations. Some of the most widely used timber species include Eucalyptus spp., Gmelina arborea, Pinus spp., Shorea robusta, Tectona grandis, Terminalia, and others (Schlo¨nvoigt 1998; Nair 1993). In some regions, this system is used to promote the establishment of forest plantations by providing the benefits of temporary land use to the farmers who care for the annual crops. One successful example is the “forest village” scheme in Thailand, which combines the land use benefits during the early phases of crop planting with a permanent allocation of agricultural plots to former shifting cultivators (Boonkird et al. 1984). In the Caribbean lowlands of Costa Rica, the taungya system has allowed to establish plantations of Cordia alliodora, Eucalyptus deglupta, Acacia mangium, and other timber species with Arazá (Eugenia stipitata), a perennial fruit shrub introduced from the Amazon, which starts to produce already after 2–3 years (Schlo¨nvoigt 1998). However, when introducing

2034

R.G. Muschler

new crops (as was the case with Arazá in Talamanca), care should be taken that the introduction of the crop into the planting system responds to a market demand and is accompanied by parallel activities to process and promote the new product. Based on a systematic spacing experiment in Costa Rica, Schlo¨nvoigt and Beer (2001) concluded that maize was compatible with Cordia alliodora and Eucalyptus deglupta during the first year of growth, while cassava (Manihot esculenta) significantly reduced the growth of these trees. Therefore, care must be taken to choose compatible tree-crop combinations and appropriate spacing in order to not compromise the trees’ growth. In systems with wide-enough spacing between the trees, or when the tree stand gets thinned with time, it may be possible to maintain crops for longer or even permanently, at which point the taungya system may transform into a forest farming or shaded plantation system. Further details about taungya systems can be found in Schlo¨nvoigt (1998) and Nair (1993), as well as in “Agroforestry Systems.”

Plantation-Crop Combinations In the tropics, perennial plantation crops occupy a substantial portion of the agricultural lands. The most important plantation crops include oil palm, rubber, coconut, coffee, cacao, sugarcane, tea, fruit trees (particularly citrus, avocado, bananas, mangos, cashews), pineapple, and black pepper. Depending on the crops’ light requirements, their size and architecture, and the local growing conditions, they may be grown as the overstory of the systems or as part of the understory. In many regions, most of these crops have been grown extensively as monocultures in full sun exposure. However, due to the increasing pressure from pests and diseases in long-standing large monocultures (very pronounced in banana plantations, but increasingly important also in coffee, cacao, and pineapple) and environmental degradation, there is now a growing need to diversify monocultural landscapes by adding trees and shrubs in hedges, windbreaks, living fences, riparian buffer zones, or as living supports for climbing crops like black pepper or vanilla. This landscape diversification adds “islands” for organisms of biological control and generates other ecological benefits such as the protection of soils, waters, and fauna (Vazquez 2014). Several of these crops, particularly rubber, coffee, cacao, tea, and black pepper, are often grown under some level of shade (Fig. 8), depending on the site conditions and the level of inputs. In general, the more limiting the conditions are for a specific crop, the more it will benefit from shade as such and also from the other benefits of shade trees, such as biologically fixed N, or organic matter provided to feed the soil (Muschler 2001a; see section “How and Where Does Agroforestry Work?” for a discussion of the sun-shade issue). The tree architecture, phenology, and vigor must be considered for adequate spacing, and thinning, pruning, or pollarding can be used to adjust the shading pattern and the biomass production of the trees to the needs of the associated crops. For example, for sun-loving black pepper plants, the lighter shade under Gliricidia sepium trees used as living supports is usually

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2035

Fig. 8 Plantation crop combinations with coffee and vanilla. Top left: the dominant tree species in this coffee plantation in Costa Rica is the red-flowering legume Erythrina poeppigiana used to provide shade and biomass. When pollarded, the trees form small dense crowns over the coffee as seen in the foreground and in the coffee fields behind the tall trees. Top right: a mix of many species form three shade strata for coffee in Caranavi, Bolivia. The species richness and the structural diversity of this coffee plantation make it eligible as “bird friendly” (see text for further details). Bottom left: coffee under Inga shade, with bananas and root crops in the foreground. Bottom right: biodynamic vanilla in Costa Rica on living stakes of Gliricidia sepium (Photo credits: R. Muschler)

preferable over the heavier shade of Erythrina spp., unless the latter species is pruned or pollarded more often (Muschler et al. 1993). In any case, the microclimatic benefits need to be balanced against those from the production of biomass for soil improvement and other benefits and, of course, against the management costs. Similarly, the spacing and management of shade trees for coffee, cacao, tea, and other crops need to be custom-tailored to the needs of the crops as a function of tree attributes and environmental parameters (Muschler 2001a; Muschler et al. 2006). In various regions, studies have shown that some particularly shade-tolerant plantation crops, such as coffee and cacao, can sustain very high levels of biodiversity. In fact, the structurally and botanically most complex systems under agroecological management (e.g., Fig. 8, right) can sustain a major portion of the biodiversity of tropical rainforests (Perfecto et al. 1996; Schroth et al. 2004; Vaast and Somarriba 2014). Interestingly, for some types of organisms, the species diversity in such agroforestry systems can even exceed that of the forest.

2036

R.G. Muschler

For example, Hoehn et al. (2010) reported from Indonesia that the highest species richness of bees was actually found in cacao agroforestry systems, exceeding the values of closed primary forest and also those of open agricultural land. Clearly, these examples illustrate the great potential to “use” species-rich plantation crop systems for combining production with protection. Considering that Central America alone is home to more than 2,000 tree species, it becomes clear that there is a vast potential to incorporate many more species, ideally favoring those of highest ecological importance (“keystone species”) as we learn more about their roles for productivity and the conservation of the associated flora and fauna (Ye´pez et al. 2003; Muschler et al. 2006). Similar to the case of sustainable production of livestock, systems of environmental, social, and economic certification can help drive the transition of plantation crop systems toward higher biodiversity and sustainability. Good examples are the standards established by the “Sustainable Agriculture Network” for the production of crops like coffee or cocoa (www.san.ag) and other standards that are even more rigorous, such as for organic production or, still more complex, the criteria for “bird-friendly coffee” production established by the Smithsonian Migratory Bird Center (http:// nationalzoo.si.edu/SCBI/MigratoryBirds/Coffee/default.cfm; see section “Agroforestry for Biodiversity Conservation and Ecosystem Services”; Fig. 25).

Homegardens, Multistrata Systems, and Tree Gardens In the continuum of plant successional stages, homegardens, multistrata systems, or tree gardens, expressions sometimes used synonymously, correspond to the final stages sharing similar architectural characteristics. The intense mix of a wide range of edible, medicinal, and utilitarian plants, including annuals and perennials of all dimensions, all the way up to possibly decades-old emergent overstory trees, establishes a complex multistrata system (Fig. 9). In many regions, the expression “homegarden” is primarily associated with relatively small units around the farmers’ homes that provide a wide range of products. In contrast, multistrata systems, sometimes also called multistrata tree gardens, are not necessarily linked to a homestead. Often, they include one or a few components such as coffee, cacao, coconut, or some spice, which may provide a dominant share of the family income. However, as stated by Nair (2001), there is no clear-cut distinction between multistrata systems and homegardens. The large structural, functional, and taxonomic diversity of homegardens, as their central defining feature, has been characterized in the classic paper by Fernandez and Nair (1986), followed by a wealth of other descriptive studies. In a study of 80 Mayan homegardens in Mexico, DeClerck and Negreros-Castillo (2000) have reported more than 150 species of plants and their specific roles as analogs of plants in different successional stages. In the species-rich homegarden shown in Fig. 9, an example from the humid tropical lowlands of SE Costa Rica (“Finca Loroco,” Bribri, Talamanca), the most prominent plants include emergent timber trees (Cordia alliodora, Cedrela odorata), fruit trees (mango, Spondias spp., Chrysophyllum caimito, Citrus spp.),

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

Timber

Water apple

2037

Sapote

Banana Sesame

Okra Sesame

Squash Amaranth

FOOD CROPS

E

UR

AN

EN

VEGETABLES

M

FOOD

E

GR

GREEN MANURE

LABOUR

TREE CROPS

OUR

FIR EW &T OO IMB D ER F F RUIT

OD

LAB

S

S

BLE

ETA

VEG

IN

URE

LA BO UR CO ME

FOO

RE

U MAN

MAN

R

URE MAN C ET D S, GG ,E UR LK MI BO LA

EN GRE

DE

MANURE

CASH CROPS

LIVESTOCK

Fig. 9 Top: a typical homegarden assembly of annual and perennial species. In this example from the lowlands of Costa Rica, more than 50 species are mixed, including annual and perennial plants of different sizes and animals. For details please refer to the text. Center: the presence of many species generates an ecologically more stable system and prevents the buildup of pests or diseases. Bottom: interactions among the different components of a homegarden from Kerala in India (Photo credits: R. Muschler at “Finca Loroco”; drawing from Nair and Sreedharan 1986).

2038

R.G. Muschler

bananas/plantains, tubers (Colocasia esculenta, Ipomoea batatas, Pachyrhizus erosus, Manihot esculenta), annual vegetables (squash, amaranth, sesame, okra), spices (aromatic herbs, Fernaldia pandurata, turmeric), and medicinal plants. Furthermore, various species are used for soil improvement (e.g., Canavalia, Mucuna, Pachyrhizus erosus), live fences (Erythrina berteroana, Diphysa robinioides, Gliricidia sepium), and the production of fodder (Erythrina poeppigiana, Trichanthera gigantea) for the goats, pigs, chicken, and fish that are also being raised. In addition, the system contains hedges along plot borders with melliferous and small-fruit-bearing species to support the undomesticated yet beneficial diversity of insects (bees, wasps, and many others), mammals, birds, and microorganisms, which provide the essential services of pollination (Nicholls and Altieri 2013), biological control, soil suppressivity, and, one of the most important yet often ignored services, the prevention of pest and disease problems to start with (Vazquez 2014; Daily 1997). The high diversity of species, and the structural and functional richness of the system, creates a wealth of ecological interactions, which make these systems resemble forest ecosystems (DeClerck and Negreros-Castillo 2000). It is this richness that generates higher levels of resistance and resilience to extreme climatic events and to price volatility of individual crops. As elaborated by Nair (2001), there is a great need for developing further tools to better understand the proven long-term ecological and socioeconomic stability of these systems, which have perdured more environmental and economic shocks than systems with fewer species. The classic global compilation of homegarden studies by Landauer and Brazil (1990) was complemented by a collection for Central America by Lok (1998a) and the global review by Kumar and Nair (2006).

Practices for Soil Conservation and Watershed Rehabilitation For soil conservation in hilly terrains and for watershed rehabilitation, agroforestry practices such as the establishment of vegetative barriers of woody perennials along contour lines or across erosion gullies can stabilize the soil (Fig. 10). For higher effectiveness, these rows can be interplanted with, or bordered by, parallel rows of perennial grasses such as vetiver or lemongrass and dense-rooted groundcovers like Arachis pintoi. In contrast with physical measures to retain soil, such as stone or concrete walls or terraces, living barriers tend to be more effective due to their permanent root system. In addition, once they are well established, they can provide sticks, stakes, fodder, and ecological services such as habitat for wildlife. For the protection of riverbanks and lakeshores, it is necessary to retain or establish riverine forests or vegetative buffers along watercourses (Fig. 10). Often these linear vegetative strips are the remnants of former forest or brush lands that have persisted along water courses or that were put back in order to stabilize the river banks from being eroded and undercut by floods (Fig. 10, bottom). Furthermore, in high-input agricultural landscapes, riverine forests are essential to protect the water from runoff of fertilizers and agrochemicals. Finally, these elongated structures are

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2039

Fig. 10 Top left: gully formation during massive soil erosion can be slowed, or even reverted, by establishing fast-growing perennial plants with dense root systems. This example comes from the Lake Bogoria watershed in Kenya. Top right: rehabilitation of gully with the establishment and protection of fast-growing legumes in Haiti. Bottom left: a riverine protection forest along a river in Iowa. Bottom center and right: tree roots protect riverbank in Haiti while unprotected portions are being undercut and eroded by floods (Photo credits: Amundsen et al. (2015) (top left); USDA, public domain (bottom left); R. Muschler (others))

important ecological corridors for wildlife and constitute important reservoirs for beneficial soil organisms (Harvey et al. 2006; Stamets 2005). The great ecological importance of these perennial vegetative strips is also reflected in certification guidelines like the ones of the “Sustainable Agriculture Network” (www.san.ag). Another important avenue of using agroforestry is through the establishment of artificial lakes in drylands, which serve as “oasis” that become the starting points for “regreening” the landscape. For example, during the past 20 years, more than 120 artificial lakes were established with great success in the dry highlands of Haiti (Gantheret 2010; Nicolas 2010). There, the artificial lakes retain rainwater as high up as possible in the watershed in order to maximize its potential use in gravity-fed irrigation systems downhill from the lakes (Mollison 1996; Fig. 11). As the lakes get established, flash floods diminish and the moisture permits to plant multipurpose trees in the agroforestry buffers around the lakes. As the trees grow, goats can be kept in enclosures and fed with a cut-and-carry system of annual and perennial fodder. With time, ecosystem services are reestablished and the retained water feeds a growing system of irrigated agriculture. Furthermore, fish and other animals can be produced in and around the lakes, providing additional income and high-quality food. The success of these efforts to rehabilitate degraded watersheds with community support, even more so under the extreme conditions of Haiti, illustrates the power inherent in applying agroforestry and permacultural knowledge and tools. In the

2040

R.G. Muschler

Fig. 11 Top left: the establishment of artificial lakes in the upper watershed allows retaining water in the highlands and diverting it for many uses along its downhill course (Source: Mollison 1996). Others: examples of artificial lakes established in a largely deforested semiarid landscape in the entral Highlands of Haiti. The shorelines of newly established lakes can be protected with strips of Vetiver or other grasses. In the course of some years, the lakes permit to “regreen” the environment (Photo credits: R. Muschler)

medium term, it is desirable to continue enriching the system with further ingredients until a patchwork of permaculture is established that maximizes production and protection functions at the same time (Mollison 1996). Undoubtedly, this experience merits being studied and replicated in many other regions with appropriate geological and social conditions. For further details on agroforestry for soil protection, see section “Benefits of Trees for Soil Fertility and Protection.”

Other Agroforestry Practices Besides the most common agroforestry practices discussed so far, others are restricted to local importance, often for specialty products, such as honey, fish, waterfowl, and others. These systems can include fish farming in natural or artificial ponds established for retaining water. Ideally, these ponds should be protected by a border of multipurpose trees and shrubs. In the example in Fig. 12, the protective ring of plants around the pond protects the banks and provides food for birds and fish. The water hyacinths (Eichhornia crassipes), a highly invasive species, are fed to pigs and chicken, and leftovers are composted. Another attractive agroforestry practice is bee farming, preferably in regions with limited exposure to pesticides.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2041

Fig. 12 Aquaculture and bee keeping as part of agroforestry systems in Costa Rica and Cuba. Top Left and Right: the pond serves as water reservoir for the dry season and is used to produce Tilapia fish. Notice the densely vegetated border with fruiting shrubs and trees to protect the banks, shade the pond, and feed the fish. The highly nutritious fodder species Trichanthera gigantea (in the foreground of the left picture) is used for feeding goats, pigs, and the fish in the pond. When the water hyacinths overgrow the pond, they are removed and fed to pigs and chicken, and the residues are composted. Bottom: honey production can provide an important source of income. In this example from Cuba, the trees in the system are selected to provide forage for the bees (Photo credits: R. Muschler at “Finca Loroco,” Costa Rica, and “Finca La Marta,” Cuba)

When the system is designed in such a way that the trees, shrubs, and annual crops provide nectar and pollen during most of the year, the production of honey can become a major source of income, as is the case in the “Finca La Marta” in Cuba (Fig. 12). Numerous sources on bee keeping are online through ICRAF and other centers. The Spanish-speaking reader may want to consult the recommendations on melliferous plants for bee keeping in the tropics by May and Rodriguez (2012) or by Espina and Ordetx (1983), as well as the compendium on tropical bee keeping by Medina-Solı´s (1990).

2042

R.G. Muschler

Distribution of Agroforestry Systems Except for deserts, trees are an integral part of most agricultural landscapes. Using the criterion of a minimum of 10 % tree cover on agricultural lands, Zomer et al. (2014) estimated that some form of agroforestry is being practiced on more than 43 % of all agricultural land globally. This area covers over 1 billion ha and is home to 30 % of the rural populations, i.e., more than 900 million people. According to Nair (2012), some 60 % of this land may be under some form of tree intercropping, including improved fallows and tree intercropping in African parklands (Fig. 13). Using the same criterion, agroforestry is particularly prevalent in Central America where 95 % of agricultural land has >10 % tree cover, in Southeast Asia with 77 %, and in South America with 53 %. For other regions, the figures are Europe 49 %, East Asia 45 %, North America 40 %, sub-Saharan Africa 27 %, Southern Asia 19 %, and, finally, N and W Africa 10 %. Two key findings of the comparison between 2000 and 2010 were that (1) tree cover continued to increase slightly on agricultural lands in most regions around the world (except for Northern and Central Asia), and (2) tree cover tends to increase with humidity (Zomer et al. 2014). These figures and tendencies demonstrate the tremendous ecological and social relevance of agroforestry. Similar to the tendency of increasing species diversity toward the tropics, also the diversity of farming systems is particularly high in the tropics. While scattered trees in the landscape can be found in most tropical landscapes, the most complex and biodiverse systems, such as homegardens, are found primarily in humid or

Fig. 13 Trees outside of forests: in many regions, the most important tree resource is not found in forests, but as scattered trees in the landscape. Wide spacing allows intercropping of many plants including maize with various timber and fruit trees in Haiti (left; photo by R. Muschler) or with Faidherbia albida and Borassus akeassii in parkland of Burkina Faso (right; photo by Marco Schmidt)

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2043

Fig. 14 Distribution of agroforestry systems in the different agroecological zones of the tropics (Source: Nair 1993)

subhumid regions, often with high population densities. Examples are homegardens in Java (Marten and Abdoellah 1988), India, El Salvador, Haiti, Nicaragua (Me´ndez et al. 2001), and many other countries in Africa and Latin America (Landauer and Brazil 1990). Figure 14 gives a geographic overview of the distribution of the major agroforestry practices, and Fig. 15 indicates the prevalence of different practices according to humidity and elevation. The availability of moisture, in general correlated with higher population density, is possibly the most important factor for the distribution of agroforestry systems. Humid and subhumid zones show many systems of improved tree fallows, taungya systems, alley cropping, plantation crop combinations (cacao, rubber, bananas), and homegardens. In contrast, semiarid and arid lands tend to be dominated by multipurpose and fuelwood lots, scattered trees on pasture and rangelands, and shelterbelts and windbreaks. At intermediate moisture levels, we find fodder banks and an overlap of the mentioned systems. In the highlands, the predominant agroforestry practices are directed at the production of fuelwood, plantation crops (particularly coffee and tea), and animal fodder, with a more prominent role for shelterbelts and windbreaks to protect soils and crops.

2044

R.G. Muschler LEGEND:

Altitude (m) 1,000

500

1,500

A

Improved tree fallows

B

Improved taungya

0

B

4

6

8

Average annual precipitation (mm)

Number of dry months per year

2

C

B

A

1,500

C 1,000

L

F

F K E

A

E B

A

L

J

F

B

J G K

I

B

H

G C

F

D

G

E A

A

E

L

F

I D

F

K

H

K

500

J

H

10

J L

F

K

12

Humid lowlands

Semiarid and Arid lands

Highlands

POPULATION DENSITY

F

C

Alley cropping

D

MPTS on crop lands

E

Plantation-crop combinations Home gardens, multilayer tree gardens

F

G

Trees in soil conservation

H

Shelterbelts/ windbreaks

I

Fodder banks

J

Fuelwood lots

K

Trees on pasture/ rangelands

L

MPT woodlots

Fig. 15 Schematic distribution of agroforestry systems according to a matrix of rainfall and elevation (Source: Nair 1993)

Roles and Potential of Agroforestry for Sustainable Land and Landscape Management This section reviews the role of trees for sustainability and sketches the environmental conditions for the successful incorporation of trees into climate-smart landscapes.

Requirements for Sustainable Land Use For systems to be attractive and ecologically sustainable in the long run, land use systems should (Benyus 2002; Daily 1997; Ewel 1986, 1999): 1. Be adapted to the natural environment and mimic natural successional communities in order to make efficient use of sunlight and recycle natural resources without degrading them (e.g., soils, water) or losing them (nutrients, functional biodiversity) 2. Foster biological activity at all levels (from soil microbiology, e.g., mycorrhizae and N-fixers, all the way to the macroflora and macrofauna) in order to maximize (i) nutrient capture mechanisms through symbiotic associates,

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

3.

4. 5.

6.

2045

(ii) ecosystem suppressiveness to pests and diseases, (iii) pollination, and (iv) resistance and resilience to extreme climatic events Minimize the liberation of greenhouse gases and the dependency on external inputs (particularly agrochemicals) to reduce the C footprint and negative externalities Correspond to an early successional stage to be productive enough to be economically interesting Be diverse enough to resist ecological and economic stress to individual products (e.g., coffee affected by leaf rust, bananas by Sigatoka, cacao by Monilia etc.) Be socially acceptable to be adopted and maintained

Ideally, these systems should be simple enough to be horticulturally manageable, yet diverse enough, and have sufficient active biomass (deep root system, high leaf area index, flower and fruit production, etc.), to sustain the abovementioned essential ecosystem services. Without these services, the systems cannot function in the long run as demonstrated by many societies that have disintegrated after their agroecosystems collapsed (Diamond 2011). To a large extent, these services are provided by the high above- and belowground biomass of mid- to late-successional ecosystems with perennial components. Unfortunately, the ecosystems’ investment for maintaining this high biomass limits their harvestable output compared to annual crops (Ewel 1999). Consequently, the challenge for the ecosystem designer is to strike the right balance between high productivity of early successional stages and the protection services of mid- or late-successional stages (Mollison 1996). It is the capacity of agroforestry to combine production with protection that establishes it as a pivotal tool for the long-term sustainability of productive systems (Smith 2010). As mentioned in Fig. 2, striking the right balance involves not only finding the right components and their best arrangement but also finding the most appropriate dimension and location for each type of land use within the landscape mosaic. For this, we need to recognize the effects of trees on their environment. The following sections summarize the main effects of trees.

Benefits of Trees for Microclimate Improvement The most obvious microclimatic benefit of trees is to provide shade to associated crops, animals, and, of central importance yet often ignored, the soil. Shade, and protection of the soil by mulch or living plants, is beneficial for most soil organisms (Martius et al. 2004) and can be essential for the symbionts in the topsoil (Bardgett and van der Putten 2014; Stamets 2005). The other effects of trees include reducing wind speed and the variability of air humidity and protecting plants in the understory from direct rain (Geilfus 1994; Fig. 16). Depending on the particular climatic setting, these modifications can have positive or negative effects. In general, where crops or animals are exposed to growth-limiting factors, such as scorching sun, chilly winds in the highlands, or desiccating winds in drylands, the beneficial

2046

PROTECTION from sun, rain and wind

R.G. Muschler

SHADE

HUMIDITY - higher and - more constant TEMPERATURE - more constant - cooler days, warmer nights

Fig. 16 Left: principal microclimatic benefits provided by trees (Source: Geilfus 1994). Right: notice the vigor of the coffee plants under the shade of this Inga tree in Northern Mexico. In contrast, unprotected coffee bushes have died back where neither shade nor agrochemicals alleviate the stress of full sun exposure. It must be noted that the microclimatic benefits are confounded with belowground effects of the trees (Photo credit: R. Muschler)

effects of trees become obvious when the plants protected by trees perform better than plants away from the trees (Fig. 16) or when animals seek out the trees. Concepts and applications were compiled by Reifsnyder and Darnhofer (1989) for a wide range of agroforestry systems, complemented by the recent sourcebook on agrometeorology by Stigter (2010), an excerpt of which was given by Stigter et al. (2011). For coffee systems, shade effects are amply documented, including the works by Barradas and Fanjul (1986) and by Muschler (1998), who reported detailed PAR measurements under different levels of shade. Caramori et al. (1996) reported on the use of Mimosa scabrella for frost protection of coffee plantations in Brazil. Martius et al. (2004) described positive effects of canopy closure on the soil fauna.

Benefits of Trees for Soil Fertility and Protection Among the many roles of trees for ecosystem sustainability, probably the most important one is the capacity of trees to conserve soil and, sometimes, even improve it under appropriate conditions (Mutua et al. 2014; Magdoff and Van Es 2009; Young 1989). The main mechanisms are the physical anchoring of soil by the presence of tree roots (see Fig. 10), the formation of a mulch layer to cover the soil (Fig. 17), and the inputs of soil organic matter (SOM) from the decomposition of leaves, branches, roots, and wood. The main functions of SOM include the following: 1. Increasing nutrient cycling and retention by increasing the soils’ capacity to retain nutrients in plant-available forms (Sanchez 1995) 2. Feeding the soil organisms responsible for a healthy rhizosphere (Bardgett et al. 2014; Nardi 2007) and disease suppressivity (Lowenfels and Lewis 2010; Stamets 2005)

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2047

Fig. 17 Schematic representation of the benefits of woody shrubs and trees for nutrient cycling and the conservation of soil fertility (Source: Geilfus 1994)

3. Improving aggregation and soil structure which, in turn, augment water infiltration and water holding capacity 4. Reduced evaporation and soil crusting 5. Better root development (Huxley 1999) The central effects of including trees and woody shrubs in agroforestry systems are illustrated in Figs. 17 and 18. Since SOM is “the warehouse of most of the N, P, and S potentially available to plants, is the main energy source for microorganisms, and is a key determinant of soil structure” (Ewel 1986), long-term sustainability of production hinges on continuous inputs of biomass to protect and nourish the soil and its fauna and flora. Perennial trees and shrubs with extensive root systems and high biomass productivity are ideally suited for this. As much as it is necessary to maintain SOM inputs, it is also important to minimize soil disturbance, and consequent

2048

R.G. Muschler

Fig. 18 Some of the most widely used tree-crop arrangements on slopes for the conservation of soils in different land use systems. For details, the reader is referred to Young 1989 (Illustrations redrawn from Young 1989)

decomposition of SOM, by using practices such as “conservation agriculture” with minimum or no tillage (Mutua et al. 2014; Derpsch et al. 2010). The soil fauna provides a range of substances that facilitate the formation of organo-mineral complexes that are essential for the stability of soil aggregates. Fungal hyphae and their products contribute further to aggregate stability (Cardoso and Kuyper 2006: Jime´nez and Thomas 2001). The burrowing and mixing actions of soil fauna, a widely recognized service of earthworms in humid climates (Lavelle et al. 1999) and of ants and termites in dry environments (Evans et al. 2011), increase the number and dimensions of macropores and, therefore, the infiltration capacity of soils, a determinant for the retention and slow release of water in watersheds (Jime´nez and Thomas 2001). In many tropical agroecosystems, earthworms can increase agricultural productivity by 40 % and even more, an effect that is particularly pronounced for grain crops (Brown et al. 1999). For dryland grain production in Australia, Evans et al. (2011) reported a 36 % increase due to the presence of ants and termites. However, to sustain earthworms and other beneficial organisms, the inputs of crop and tree residues from permanent agroforestry systems are

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2049

indispensible since they represent the principal source of energy for the heterotrophic soil organisms. In coffee systems, shade and organic crop management increased earthworm abundance substantially (Sánchez de Leo´n et al. 2006). Another beneficial effect of some perennials is linked to the unique capacity of certain perennial mycorrhizal species like Cajanus cajan (Shibata and Yano 2003) and Tithonia diversifolia (Jama et al. 2000; Phiri et al. 2003) to solubilize P from insoluble pools that are usually not available to other plant roots. Using the capacity of plants like these, and their microbial associates, allows to “mine” nutrients and incorporates them into the mobile P cycle of the system (Cardoso and Kuyper 2006). Given the increasing scarcity of P supplies (IASS 2015), this is a topic that merits further research. Over the past decades, a massive body of quantitative information has been amassed on nutrient stocks and flows in agroforestry systems. The summary of the first decade of research by Fassbender et al. (1991) and Fassbender (1993) was complemented by a series of studies on the transformation and flow of individual elements, particularly N and P, using nutrient balance studies and methods with tracers (Schroth and Sinclair 2003; Cadisch and Giller 1997; Sanchez 1995; Haggar et al. 1993, 1991). Hartemink (2005) presented a review on nutrient stocks and cycling in cocoa systems. For detailed information on methods to assess aspects like the effects of trees on nutrient cycling, SOM, soil properties, carbon sequestration, and related topics, the reader is referred to Schroth and Sinclair (2003). Young (1989) provided a detailed review on the use of agroforestry for soil conservation. The effectiveness of trees for soil conservation and, given appropriate conditions and enough time, even recuperation is illustrated by the capacity of residual tree “islands” to retain soil and water and from reforesting degraded landscapes by community-supported tree planting and protection trees in Haiti (Fig. 19).

How and Where Does Agroforestry Work? Trees are perennials, which means that they will usually be longer present in the system than animals, crops, and man. Consequently, their effects will be felt for decades or longer. Due to their larger size (in most cases), their foliage and branches can permanently, and predictably (often modified by pruning), moderate the microclimate for the crops underneath by reducing wind speed and extremes of temperature and moisture which otherwise might be growth limiting. At the same time, and in the absence of rooting barriers (not always the case) or chemical or water limitations, their roots can potentially exploit a larger soil volume than those of the relatively short-lived and smaller crop plants. This can, then, translate into an uptake of nutrients which otherwise would not be accessible to the crops and the incorporation of these nutrients into the crop environment once the tree sheds leaves, branches, and roots. This is the most basic premise for the benefits of plant associations in agroforestry (Ong and Huxley 1996). Put in other words, agroforestry is beneficial whenever trees can protect crops, of particular importance in stressful environments, or help crops to access more nutrients and water than they would be able to access or use if they were alone.

2050

R.G. Muschler

Fig. 19 Signs of hope: trees to protect and recuperate soils. In the south of Haiti, the determined community of “La Vallue” has started a slow but steady recovery of their natural resources by judicious use and protection of trees on hilltops and exposed ledges. Of particular importance are multipurpose trees that provide biologically fixed nitrogen for soil improvement and fodder for the stabulated animals as well as fruits for human consumption. Note how the presence of trees allows protecting the soil in the three agroforestry “islands” (top left) and on the ridge of the steep slopes (bottom). Without the protection of the trees, soil degradation leads quickly to complete loss of fertile topsoil. This results not only in the loss of productivity of the land but also of its capacity to store rainwater for slow release during the dry season (Photos: R. Muschler)

An illustrative example comes from a study of nutrient cycling in coffee ecosystems in Costa Rica (Table 2; Muschler 1998), which showed that the inputs of N, P, K, Ca, and Mg in the form of pruning residues of Erythrina poeppigiana shade trees were higher (for many nutrients by a multiple) than the nutrient extraction in the form of harvested coffee beans, even at a high productivity or 7.5 t of coffee beans per ha. The differences were particularly high when the trees were subject to “selective pruning” (i.e., selective removal of some large branches to create an “open shade” pattern) compared to complete pruning of all juvenile branches (“pollarding”), because the trees were able to provide more biomass. With “selective pruning,” about half of the large branches are retained, which allows the trees to produce biomass year-round. In contrast, “pollarding” sets back the biomass production drastically due to the loss of all branches until they resprout 3–6 weeks after pruning. As a result, the C supply to the N-fixing bacteria is interrupted causing N-fixation to decline drastically until 3 months after the pollarding (Nygren

3.6 6.6 31.5 1.5 33.6 36.6

58 106 219 27 250 298

1,241 2,277 – 1,380 – –

P

27 85.5 100

18 33 94.5

K

3.6 114.4 119.4

18 33 100

Ca

1.5 40.3 43.6

4.0 7.3 37.8

Mg

Organic matter inputs from tree pruning (dry weight of leaves and green non-lignified shoots); does not include litter fall, branches, stems, roots, or coffee residues b Calculated for 156 trees per hectare (8  8 m spacing) c Standard recommendation per hectare and year: 2  315 kg of complete fertilizer (18-5-15-6-2; N-P-K-Mg-B) + 315 kg NH4NO3 + 250 kg CaCO3

a

Input/output via Inputs 1. Pollardingb 2. Open shade (selective pruning)b 3. Commercial fertilizationc Outputs 4. Harvested coffee (7.5 Mg fresh weight ffi 30 sacks green coffee of 46 kg each) Balance for pollarding (1 + 3  4) Balance for open shade (2 + 3  4)

N OMa kg ha1 year1

Table 2 Nutrient balance for agroforestry systems with coffee under the leguminous Erythrina poeppigiana shade trees subjected to different pruning regimes. The inputs stem from fertilization and the nutrient inputs of pruning residuesa from Erythrina poeppigiana shade trees. The nutrient export was calculated for extracting 7.5 Mg of fresh ripe coffee berries per ha (Muschler 1998)

Agroforestry: Essential for Sustainable and Climate-Smart Land Use? 2051

2052

R.G. Muschler

and Ramirez 1995). However, even when the shade trees were pollarded twice per year, the most drastic treatment possible, the nutrients in the pruned biomass still exceeded the extraction, except for K. When the inputs from the commercial fertilization were included in the nutrient budget, the overall balance for N indicated an excess of 250 kg of N per ha and year. Besides economic implications, this excessive fertilization also has strong ecological implications due to N leaching into water systems and N losses to the atmosphere. As this example shows, the integration and management of the tree component can, under appropriate conditions, supply a substantial portion of needed nutrients for the long-term sustainability of the production systems (Muschler 2001a, 2004; Beer et al. 1998). At the same time, the nutrient contributions from the trees contribute to mitigate climate change by reducing the needs for synthetic fertilizers (see section “Multifunctionality of Agroforestry: Climate-Smart Production, Protection and Ecosystem Services”). Another central benefit of agroforestry is its capacity to harvest and use more sunlight than simpler agricultural systems. A more complex system with more plants, distributed in different strata and with a higher leaf area index, can channel more sunlight into photosynthetic products compared to simpler systems. This explains the generally higher land equivalent ratio of agroforestry compared to annual cropping systems. The challenge is to match the most compatible crops and trees and manage them in such a way that the complementary phenology and optimum display of leaves throughout the year and the diurnal course of the sun will maximize light interception (Ong et al. 1996). When animals are added, the biggest benefit is obtained when they can make use of plant material which otherwise would not be used and transform them into useful products such as meat, milk, and organic fertilizer. Examples are the use of savanna grasses or the foliage of fodder shrubs and trees in dryland silvopastoral systems. When the trees in a given agroforestry system are able to capture more additional resources with only minimal effects on the resource capture of the crops, the interactions are positive and complementary (Fig. 20; Ong and Leakey 1999). This is the case, for example, when trees are able to access nutrients or water below the rooting depth of the crops or to make better use of off-season rainfall, which could not be used by crops. In contrast, when the capture of a given resource by trees causes a disproportionately large reduction of the same resource for the crop, the interaction is competitive. This occurs, for example, under water stress when the water use by trees may cause a drastic reduction of crop yield (Fig. 20), as shown for maize growing at less than 250 mm of rain in Kenya (Ong and Leakey 1999). With higher humidity (above 650 mm), this competition did not occur. It is this same relationship of competition as a function of resource availability that allows alley cropping to work only as long as there is enough water and a minimum of nutrients for both trees and crops (see section on “Alley Cropping”). Since agroforestry systems can transform more sunlight into photosynthates than simpler agricultural systems, there will be more living biomass and, therefore, also more dead biomass. In turn, this contribution of transformed sunlight (much like the fossil fuel, a reminder of million-year-old transformed sunlight, that modern society depends on so desperately) feeds the soil fauna and its heterotrophic microflora.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2053

1000 y = –9.6962x + 980.81 R2 = 0.9336

Capture by Crop

Complementary Neutral 2

Competitive 3

Maize yield (kg/ha)

1 800 600 400 200 0 Capture by Tree

0

20

40

60

80

Tree water use (mm)

Fig. 20 Left: depending on the attributes of the species used and the environment in which they grow, the interactions between trees and crops can be complementary, neutral, or competitive, depending on the overall sum of resource capture by the trees and crops. Right: under moisture-limited conditions, tree water use determines the productivity of maize (Source: Ong and Leakey 1999)

As described in section “Benefits of Trees for Soil Fertility and Protection,” “feeding the soil” and its biota is essential for moving closer toward the goal of ecological intensification and true sustainability (Lowenfels and Lewis 2010; Stamets 2005). After all, SOM and the associated soil organisms are key to protect soil fertility (Bardgett and van der Putten 2014; Young 1989), to access more nutrients and water, and to prevent the buildup of pests and diseases (Altieri and Nicholls 2003). Thus, the trees contribute both to the improvement of the crop environment2 and to the maintenance of soil productivity as the fundamental prerequisite for productive and ecologically sustainable agroecosystems (Gliessman 2015). The overall benefits of the system depend on the components and their interactions which, in turn, depend on the species mix, the planting arrangements, the management and pruning, as well as on the biophysical and socioeconomic environments which define growth rates and the availability of natural resources, labor, and inputs (Muschler 1993). Just as the availability of water determines whether an alley cropping arrangement can be beneficial or not (see section on “Alley Cropping”), so are temperature/elevation and soil fertility two key factors that determine the overall benefits of shade for Coffea arabica (Fig. 21). Figure 21 illustrates generalized shade responses of long-term coffee production across the elevational range for coffee. These graphs reflect the quantitative data of many studies (for details see Muschler 2004). Irrespective of soil condition, the highest production of unshaded coffee typically occurs at intermediate altitudes since they provide the ideal climate for coffee. In Central America, this is often the case between altitudes of about 900 and 1,300 masl. At lower elevations, unshaded

2

In the ideal agroforestry system, also the reverse is true: crops will improve the environment for the trees, e.g., through contributions of biologically fixed atmospheric N from N-fixing crops which can benefit the trees.

2054

R.G. Muschler

Fig. 21 Idealized coffee production in full sun (dotted lines) and under trees giving 50 % shade (thick lines) as a function of elevation for soils without (a) and with (b) limitations of rooting depth, nutrients, or moisture. Notice that the benefit from shade (“shade contribution”) is largest at elevations below or above the optimum elevation for coffee. For further explanation see text

coffee production decreases in response to increasing heat stress, while at higher elevations it decreases due to low temperatures and possibly wind damage. Under such suboptimal conditions, trees can reduce the microclimatic stress to the coffee plants through shading at low elevations and through reducing winds at high elevations. Consequently, trees tend to increase coffee production over that of unshaded plots. This benefit is marked as the dappled area in Fig. 21 referred to as “shade contribution.” In contrast, within the optimum elevational range, because the microclimate is already ideal for coffee, trees cannot exert such a beneficial effect via microclimate improvement. Under these conditions, shading may even reduce coffee production. This is marked as the hatched area labeled as “excessive shade” in the “good soil” scenario. On “bad” soils, the productivity of both coffee systems is relatively lower as a result of the nutrient or moisture limitations. However, the productivity of unshaded coffee drops off more under such conditions due to the absence of the beneficial effects of trees on nutrient cycling and water retention. Consequently, the “shade contribution” becomes larger and probably extends across the whole elevational gradient. Considering these different environments, the seeming contradiction between studies that report benefits of shade and others reporting shade-induced yield reductions disappear: both positions can be right, but each one for a different environment. The benefits of trees and shade under suboptimal conditions for coffee have been amply characterized by Muschler (1998), who found that intermediate levels of shade (40–60 % shade) were the best option because they permitted combining high productivity with improved coffee quality (Muschler 2001b), while generating environmental benefits through weed control, increased nutrient cycling, and better plant health. The work on coffee quality of Salazar et al. (2000) and recent work of Pinard et al. (2014) in East Africa has confirmed these conclusions. Furthermore, it is interesting to note that tree benefits can even

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2055

be detected for some environments under optimal conditions (Siles et al. 2010). Having reviewed these positive effects of trees, the question arises why trees are not used in all land use systems. What are the constraints that limit or prohibit their association with crops under certain conditions?

Constraints to the Success of Agroforestry The main limitations for agroforestry are due to historic, socioeconomic, or biophysical factors. The historical limitation is that industrial agricultural development of the last 50 years or so has focused on yield maximization of high-input monocrops, often hybrids selected for high productivity in full sun. This is usually achieved through the complete separation of trees and crops. When the objective is to maximize the yield of one single product, there is little room for other plants, let alone high-diversity systems. This tendency remains strong even today as illustrated by the strong push for monocultures of rubber in Asia (van Nordwijk et al. 2012), of cocoa in Ghana (Ruf 2011), and of coffee in Latin America (Jha et al. 2014). The reasons for limited interest to plant trees include labor costs for planting and tending the trees, reduced productivity of the main crop, insecure land holding rights, mobility of farmers, limited rights over the use of the planted trees, little information about compatible trees, and lack of financial recompensation for generating ecological benefits from a diversified production. In different mixes, these limitations are commonplace in the tropics. The main socioeconomic constraints to agroforestry success are higher labor demand for, e.g., pruning and biomass recycling (this argument is often mentioned against alley cropping) and delayed returns from the trees, which may need more than a decade to grow to commercial dimensions. The main biophysical limitations for tree-crop associations fall into three groups. First, the production of sun-demanding crops may be strongly reduced by excessive shade of the trees. Adequate selection and management of the trees to reduce shading and of the crop species and varieties to tolerate more shade can reduce or even eliminate this problem. Second, in nutrient- or water-limited situations, trees can affect crops negatively via competition for water and/or nutrients. An example of this is the failure of alley cropping in arid environments where water competition by the tree reduces crop production strongly (see the section on “Alley Cropping”). Furthermore, some tree species, for example, certain species of Juglans and Eucalyptus, have been shown to suppress associated plants with chemicals liberated into the soil. This effect of allelopathy is particularly pronounced in dry climates where the allelochemicals are not leached from the soil before they can act. In both cases, competition and allelopathy, it is, again, the environment that determines the degree to which the tree may negatively affect associated crops. But also the expression and magnitude of positive effects depend on the environment. And this is the third biophysical limitation. In optimum environments, with minimal or no environmental stress for the crops, the beneficial effect of trees by alleviating a stressful condition may simply not be important. Examples may be the nutrient-rich and

2056

R.G. Muschler

deep volcanic (some coffee areas of the Central Highlands of Costa Rica) or alluvial soils in moist climates where crops are not stressed for nutrients or water. Under such conditions, the benefits from the trees, ignoring their universal long-term contribution to maintain soil fertility, may not be strong enough to balance their negative effects. However, such sites are the exception. Sites with some soil or climatic constraints are much more common, and it is on such sites where the positive contributions of trees can be exploited for the benefit of the associated crop. With recent concerns about environmental degradation from high-input monocultures and the volatility of commodity prices, the production objectives are starting to become wider to include income diversification, hence increased stability, plus increased ecological resistance and resilience of the systems to pests, diseases, and climatic extremes. Increasingly, low-input technologies are being investigated where biological inputs and services substitute chemical inputs. This work assumes special relevance in tropical countries with limited financial resources. In order to generate the most effective systems, the selection of the most appropriate plants is key.

Plant Selection for Agroforestry The experiences with agroforestry over the past four decades have shown great promise for associating trees, crops, and animals for mutual benefit and for generating essential ecosystem services for the people who depend on these systems. The use of ecologically appropriate and economically attractive trees and crops is possibly the main tool in the fight against the “monoculturization” of many systems, including rubber (van Nordwijk et al. 2012), cacao, or coffee (Jha et al. 2014; Klein et al. 2008). Finding the right trees and compatible productive crops, and promoting their use in appropriate environments, should be a central goal of future efforts. In order to allow for the systematic screening of tree species compatible with particular crops, various authors generated lists of selection criteria of tree attributes. The criteria for identifying compatible shade trees for coffee, cacao, and tea given in the classic papers by Willey (1975) and Beer (1987) were taken up, and complemented, for the training manuals written by Geilfus (1994) and Muschler (2001a). With a focus on optimizing belowground interactions, Schroth (1995) generated a list of tree root characteristics for selecting appropriate species. As mentioned in the section on agroforestry for soil improvement, the selection of plants based on root attributes should also consider the particular capacity of mycorrhizal plants like Cajanus cajan and Tithonia diversifolia to extract P from insoluble pools in the soil. These criteria and species lists were further expanded by ICRAF and other organizations. Today, they are largely incorporated into the databases with a wealth of plant descriptors and the selection tools provided by ICRAF (see Box 1). For the selection of overstory plants, the most desirable attributes, besides their products, include their capacity to quickly provide shade, shelter, and environmental benefits to the associated trees, shrubs, and crops. Ideally, these services should be synchronized with the needs of the associated plants (for N synchrony, see Crews and Peoples 2005) along with their phenological cycle, and of increasing

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2057

importance under the scenarios of climate change, the overstory trees should be able to provide these services even under extreme climatic events. Therefore, the most desirable attributes for overstory plants include the capacity to sprout and grow easily, even on poor or shallow soils, the development of a strong and deep root system for good anchoring of the plant, and the resistance to drought and wind (Muschler 2001a; Geilfus 1994; Nair 1993; Beer 1987; von Maydell 1986). Unfortunately, as the experiences over the past decades with the widespread introduction of the leguminous shade tree Erythrina poeppigiana in Costa Rica and other regions have shown, these requirements cannot always be met simultaneously. This species, native to Colombia, has been widely promoted and planted in most of the coffee regions of Costa Rica since the 1950s due to its ability to sprout easily from branch cuttings, to grow very fast so as to provide substantial shade within 6 months, and to produce much biomass. However, when trees that have been established from cuttings grow to more than 10 m in height, they may be easily toppled by strong winds; their root systems are not as strong as those from trees planted from seed. Another limitation of this species is its aggressive reproduction, which, together with its fast growth, makes it a species that is potentially invasive. Today, in many coffee farms in Costa Rica, this species has become so dominant that there is little space for other shade species and their products and services. With the interest in diversifying systems also for sustaining ecological services, care must be taken that the selection of trees and crops avoids the promotion of invasive species which can take over large areas as was the case with Leucaena leucocephala, neem, Casuarina, and many other species (Richardson et al. 2004). Lists of appropriate crop and tree species, and their attributes, have been assembled for different regions. Notable examples are the classic “Crop Species Manual” by Nair (1980), lists of multipurpose and other trees by Nair (1993), as well as the compilation of von Maydell (1986) on trees of the Sahel. Authoritative publications on the selection, improvement, and management of trees are Árboles de Centroamerica (Cordero and Boshier 2003), Specialty Crops for Pacific Islands (Elevitch 2011), Brazilian Fruits & Cultivated Exotics covering 827 types of fruits by Lorenzi et al. (2006), and the classics on Tropical Forests and their Crops by Smith et al. (1992) and Fruits of Warm Climates by Morton (1987). The following sections cover central aspects of selection of trees, crops, animals, and microorganisms for creating productive and ecologically stable agroforestry systems.

Tree Domestication in Agroforestry With respect to the domestication of trees for agroforestry, Leakey (1999) provided a review on the potential for novel food products from agroforestry trees. Recently, Leakey et al. (2012) summarized the development of this important field since its kickoff at the 1992 conference in Edinburgh, UK, on “Tropical trees: The Potential for Domestication and the Rebuilding of Forest Resources.” It was at this conference that the attention focused on the need to work on these overlooked and underutilized “Cinderella” species holding a considerable potential for developing

2058

R.G. Muschler

specialty products for niche markets and for improving the nutrition and income of their cultivators. The first decade of tree domestication for agroforestry in the 1990s was dedicated primarily to assess the potential of more than 50 different tree species, mostly from Africa. The species include Irvingia gabonensis and Dacryodes edulis from Cameroon and Nigeria, two of the most widely studied indigenous fruit trees from Africa. With time, farmers added other tree species to the list of promising species for providing timber, fodder, medicines, and fuelwood. The research topics included techniques for the production of improved germplasm, the characterization of morphological and genetic variation, the promotion and marketing of such species, and farmers’ rights. Successful vegetative propagation (marcotting) of trees allows to obtain fruits as early as 2 or 3 years after planting. This aspect is essential for farmers who need to receive a fast return on their investment in future tree crops. A key lesson learned is the importance of differentiating among the quality, size, and taste of fruits from individual trees. Only when this is done from the start, and maintained during the promotion of a new crop, can the market provide financial incentives for the speedy evolution of new high-quality fruits. Looking at the topics covered by 424 scientific publications on tree domestication between 1992 and 2012 (Leakey et al. 2012), the vast majority of papers refer to the domestication concept and strategy (61 papers), propagation and germplasm (69), species potential (69), genetic and morphological characterization (89), followed by much less emphasis on nutritional benefits (23), agroforestry enrichment (22), commercial issues (22), and only a few handful of papers on all other topics. In particular it calls attention that only 14 papers addressed ecology, with a meek 5 papers on adoption and impact (all from Africa) and 6 on participatory implementation on farms (5 from Africa). Assuming that such an analysis (despite its limitations due to thematic overlaps) reveals a real trend, it shows that more work needs to be done on the upper portion of the value chains and also on strengthening the work in Asia and Latin America. Although, compared to Africa, more work may already have been done in Asia by the PROSEA network which collected detailed information on medicinal crops, fruits, and vegetables (see online resources) and Latin America (ChizmarFernández et al. 2009; Padulosi et al. 2002; BOSTID 1989; Morton 1987), the realm of underutilized tree species from all continents exceeds the number of currently (widely) used species by one and possibly up to two orders of magnitude. Globally, estimates indicate that more than 50,000 species of plants are edible, including many trees. From the Brazilian Amazon, possibly the largest remaining pool of NUS, some of the salient accounts of underutilized species include more than 800 species and varieties (Lorenzi et al. 2006; Smith et al. 1992, 2007). From Africa, Nyambo et al. (2005) have summarized information about fruits and nuts for Tanzania, and Msyua et al. (2009) have reported the contents of Fe, Zn, and β-carotene for noncultivated indigenous vegetables. For Latin America, one of the most useful resources for the agroforester, available for free download at www.arbolesdecentroamerica.info, is an authoritative compendium on the ecology, growth, and uses of more than 180 tree species, including many native species of Central America (Cordero and Boshier 2003).

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2059

With increasing uploading of essential passport information of tree and crop species to databases on the Internet, mostly through organizations of the CGIAR and UN systems, universities, and NGOs, the amount of information on tree species for the tropics is now overwhelming, although some aspects are largely missing (see section “Lessons Learnt from Tree, Crop, and Animal Domestication: Widening the Search”). Key information on more than 600 tree species for agroforestry and on over 22 000 tree and crop species can be accessed through the “Agroforestree” Database and the “Agroforestry Species Switchboard” of ICRAF (Box 1). Box 1. Online Resources on Tree and Crop Species of the “World Agroforestry Centre” (ICRAF)

On the page http://intranet.icraf.org/treesnmarkets/sd3/decision_support_ tools.php, ICRAF provides access to a series of support tools for selecting and managing tree species for agroforestry and forestry. The two most widely used systems are the “Agroforestree Database” and the “Agroforestry Species Switchboard.” Other tools on tree domestication, nursery practices, species selection for different environments, genomics, and other relevant topics are found under this same link.

1. The Agroforestree Database This database (Orwa et al. 2009) is a species reference and selection guide for currently 600 tree species that are deliberately grown and managed in agroforestry systems to provide multiple outputs. The database provides key information on native and exotic trees globally and allows users also (continued)

2060

R.G. Muschler

to search by country, whether a species is native or exotic, or by products and/or services provided. The characterization of each species includes a botanic description plus details on climate, distribution range, ecology, propagation, management, and uses. URL: http://www.worldagroforestry.org/resources/databases/agroforestree 2. The Agroforestry Species Switchboard The Agroforestry Species Switchboard (Kindt et al. 2013) currently includes more than 22,000 plant species in 13 web-based databases. When possible, hyperlinks are provided to facilitate access of information to the linked databases. URL: http://www.worldagroforestry.org/products/switchboard/index.php

Selection of Crop Species for Agroforestry Besides its desired products and functions, whether a particular plant species is appropriate or not for agroforestry depends on the attributes of this species that make it fit into the specific agronomic niche in an agroforestry arrangement. For example, for understory plants, this includes the capacity to tolerate a certain level of shade and tolerance of the associated plants. As summarized by Cannell (1983), the response of a crop plant to shading depends on the type of plant and whether its final product will be a fruit, a leaf, or a storage organ (Fig. 22, left). Similarly, the response of crop plants to soil fertility will vary between different types of crops depending on the size of their root systems and symbiotic partners. While nonleguminous vegetables and fruit trees tend to suffer heavily on less fertile soils, N-fixing legumes and many tree species are able to cope better with lower soil fertility (Fig. 22, right), being helped by N-fixing symbionts, mycorrhizae, and larger root systems. This generalized behavior has obvious implications for the choice of appropriate crops under particular conditions of shading and soil fertility. Of course, it must be stressed that, as always, generalizations must be treated with caution and that there are exceptions. The response of crop plants to shade has been studied in great detail for some tropical crops, particularly coffee and cacao (Theobroma cacao) that have been produced very successfully as an understory crop under the shade of multiple tree strata (Vaast and Somarriba 2014; Somarriba and Beer 2011). Another example is coffee, which can, depending on species and variety, cope with a wide range of light conditions ranging from full sun exposure in the lowlands for Robusta or Liberica coffee (Coffea canephora or C. liberica) all the way to an intermediate shade level of 40–60 % or even more under suboptimal conditions for C. arabica (Muschler 2004). While the agronomic research since the 1970s has generated a lot of research on the shade-sun requirements or preferences of these two crops as a function of environmental and management factors, this detail of information is, unfortunately, still only partially available for many other crops. For the agroforestry practitioner, the classic summary of environmental requirements of many crops (Nair 1980) is still a good starting point, complemented by additional information from online resources.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use? Leaf yielding

Stor

root

crops

crop

s

Fru

it-y

Non-legumes, fruit trees, many herbaceous crops

Legu

iel

mes

din

gc

rop

, ma

ny tr

s

ee s

pp.

s

lm Pa

Yield per ha

age

2061

Increasing shade

Decreasing soil fertility

Fig. 22 Generalized response of different types of crops to increasing levels of shade and decreasing soil fertility (Modified from Cannell 1983)

For the selection of underutilized crops, a good starting point is the recent review by Ebert (2014). The range of promising crops and trees includes more than 100 edible species from Central America (Chizmar-Fernández et al. 2009) and a multiple from many other tropical regions. While some of them provide relatively small contributions to the overall productivity because only their flowers, select leaves, or a small portion of their stems are commonly used, some of them offer an interesting potential as complementary crops, particularly when they substitute plants of less use. With regard to the conservation and better use of agrobiodiversity, it should be a high priority to collect locally adapted crop species and their seeds or other reproductive tissues and to document the tacit information about their uses. One interesting example comes from El Salvador, where more than 20 species have been documented recently as promising components in climatesmart agroforestry systems designed for improving food security (SánchezSalmero´n et al. 2015; Box 2). Box 2. Identification of Underutilized Promising Species for Agroforestry Systems in El Salvador (Sa´nchez-Salmero´n et al. 2015)

The leaves of many plants can provide high levels of vitamins A and C, as well as proteins and micronutrients such as Fe and Zn, which are often deficient in the diets of many poor communities in the tropics (FAO 2011). Based on interviews with farmers of different ages, Sánchez-Salmero´n et al. (2015) obtained a list of 23 promising species that combine favorable agronomic attributes (Geilfus 2002) with adaptation to stress of extreme climate change events and high micronutrient content. Of these 23 species, 18 are high in Fe, 11 high in Zn, and 8 have medium or high levels of vitamin A. In addition, 12 of these species are reported as drought resistant, 11 resist (continued)

2062

R.G. Muschler

strong winds, 4 resist inundations, and 8 withstand high temperatures. Finally, two of the top species (Moringa oleifera and Cnidoscolus chayamansa) provide nutritious leaves for human nutrition throughout the year; both are woody agroforestry species. The list includes the species presented in Table 3.

Similar information from Africa, Asia, and elsewhere should be increasingly considered for the search for new food crops that allow adding nutrition and agronomic resilience to the food systems of the future. Good starting points are the publications on Lost Crops of the Incas and Lost Crops of Africa by BOSTID (1989, 1996a, b) as well as the encyclopedic information of the Plant Resources of South-East Asia Network (PROSEA), as well as Smith et al. (1992). Although many promising crop materials are already available in the extensive germplasm collections of organizations like CATIE (Ebert et al. 2007) in Costa Rica and “The World Vegetable Center” (AVRDC) in Taiwan, they are yet to be studied in more detail in order to realize their agronomic and nutritional potential. Although still greatly underfunded and in its infancy, an increasing amount of information is coming online on underutilized crops, also called “orphan crops” or “foods of the future” (Jaenicke and Ho¨schle-Zeledon 2006), promoted by international nongovernmental initiatives such as “Slow Food” (www.slowfood.org) or “Grain” (www.grain.com), which support small farmers and social movements in their struggles for community-controlled and biodiversity-based food systems. Undoubtedly, in the years to come, the search for new crops or trees for agroforestry systems will also have to address the potential and limitations of genetically transformed crops. The range includes not only first- or secondgeneration GMOs but also plants modified for nutritional benefits or for higher photosynthetic capacities such as a C4-type rice that is currently in development. The recent reviews of Antoniou et al. (2012) and by Funes-Monzote and Freyre (2009) are good starting points to analyze the potential implications for human and ecosystem health. For the selection and best arrangement of crops and trees, it would be good to build on the experiences of “companion planting,” i.e., the matching of particularly compatible species such as carrots and tomatoes (Riotte 1998; see also the section in Wikipedia). While companion planting is a time-proven practice in horticulture and gardening, the systematic application of this practice is yet pending for agroforestry. The positive examples of species matches such as certain legume and timber species with coffee and cacao (Somarriba et al. 2014; Somarriba and Beer 2011), or the tree – pasture – animal matches in silvopastoral systems, illustrate that much information exists, but the building of a structured systematic tool for identifying the best matches is still pending. The development of a matrix of crop-crop, crop-tree, and tree-tree compatibility may be a useful step. Finally, considering the large intraspecific variability of crop responses and quality (e.g., hundreds of varieties of squash, tomatoes, etc.), also this level of variability should be considered to match the most appropriate varieties to the specific system and the

Ipomoea batatas Matelea sp. Moringa oleifera Portulaca oleracea

Brosimum alicastrum Calathea macrosepala Cnidoscolus chayamansa Crescentia alata Crotalaria longirostrata Gliricidia sepium

Bromelia alsodes

Scientific name (bold letters indicate woody spp.) Artocarpus altilis

x

Flower

Madre cacao Camote Chuchulaya Moringa Verdolaga x

x

Seed Leaf

Morro Chipilı´n

Chaya

Root Leaf Leaf Leaf

x

Flower, root Leaf

Ojushteb

Macuz

x x

Plant part used Seed

Drought

Young shoots Seed

Árbol de pan Pin˜uela

Common name (in El Salvador)

Strong winds x

x x

x x

x

x

Inundation x

x

High temperature x

x

x

x

Fe l h h h

m

h h

h

h

h

m

h

Zn l l m nd

nd

a nd

h

h

h

l

h

a nd m m

nd

nd m

a

nd

m

nd

l

Jan x

x

x

Feb x

x

X

x

x

Mar x

x

x

x

x

x

x

x

x

x

x

x

x x

x

x

Harvest season in El Salvador

Micronutrient levela

VitA

Apr

Adaptation to CC Tolerant to:

June

May

Plant name

July x x

x

x

x x

x

x

x

x

Aug

Table 3 Promising underutilized plant species for agricultural and agroforestry systems in El Salvador (Sánchez-Salmero´n et al. 2015)

Sep x x

x

x

x x

x

x x

x

x

x

x

(continued)

Oct x

Nov x

Dec x

Agroforestry: Essential for Sustainable and Climate-Smart Land Use? 2063

Maı´z amarillob Maı´z negrob Maı´z tizateb

Izote

x

x x x x

Flower, shoot Seed

Seed Seed

b

x

x x

x

Drought

Seed Root

Seed Fruit, leaves Seed

Maicillob Jocote

Frijol chilipucasb Frijol arrozb Malanga

Plant part used Seed Leaf

Common name (in El Salvador) Ajonjolı´b Mora

h high, m medium, l low, nd not determined Plant products can be stored for more than a year

a

Zea mays Zea mays

Zea mays

Vigna umbellata Xanthosoma sagittifolium Yucca elephantipes

Phaseolus lunatus

Scientific name (bold letters indicate woody spp.) Sesamum indicum Solanum americanum Sorghum sp. Spondias purpurea

Strong winds x

x

x

x

Inundation x

x

High temperature x

x

x x

Fe h h

h

h

h h

h

h h

h h

Zn h h

h

l

h h

h

H nd

h m

nd nd

b

nd

b

nd

nd m

l m

Jan x x

x

x

Feb x

Mar x

x

x

x

x

x

x

x

x

x

Harvest season in El Salvador

Micronutrient levela

Apr

Adaptation to CC Tolerant to:

VitA

May

Plant name

June

Table 3 (continued)

July x x

x

x

x

x

Aug x

x

x

Sep x

x

x

Oct x x

x

x

x

x

Nov x x

x

x x

x x

Dec x x

x

x

2064 R.G. Muschler

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2065

season. This is already a practice in urban gardening in Cuba by using different varieties for different seasons (INIFAT 2011).

Selection of Animal Species While most of the mainstream agricultural and agroforestry work with animals has concentrated largely on systems with cattle (Pezo and Ibrahim 2001), goats (Benavides and Arias 1995), and chicken, there is also great potential in exploring other animal species for the design of locally adapted agroforestry systems. For example, in some places, sheep are being used for the production of meat and milk or for providing the service of weeding as, for example, in coffee fields in Central America (Leupolz 2000). In other cases, pigs can be integrated to produce meat and biogas for cooking. Furthermore, there is a wide field to be explored with the production of “exotic” species and microbreeds for special animal production systems. A global review published by BOSTID (1991) revealed a significant potential for microbreeds of cattle, goats, sheep and pigs, but also more than 30 species of poultry, rodents, lizards, and others. In many places, rabbits, guinea pigs, and other small animals can be raised profitably with tree fodder from agroforestry systems, but so far, relatively little attention is being paid to these options. In Central America, a remarkable example is the native green iguana, which can easily be raised for meat production on a leafy diet from agroforestry systems (BOSTID 1991). Undoubtedly, with increasing pressure on natural resources and the evolution of specialty markets, there is a significant potential to be explored. Another field that is yet to be developed is the production of edible insects. Compared to the production of beef, pork, chicken, or milk, the production of an equivalent amount of animal protein from raising insects such as mealworms requires only a fraction of the food and water and liberates much less greenhouse gases, up to 100 times less (van Huis et al. 2013). Considering the lower feed requirements of cold-blooded insects and the higher edible proportion of insect biomass (up to 80 %), the feed efficiency for crickets, for example, is twice as high as that of chicken, at least four times higher than that for pigs, and even 12 times higher than for cattle. Clearly, this makes edible insects an interesting alternative to the conventional production of meat, either for direct human consumption or for indirect use as feedstock for animals (van Huis et al. 2013). However, much work needs to be done to boost awareness about the contributions of insects for sustaining, complementing, or even enhancing agroforestry systems and for contributing valuable and nutritious food and feed. The synergistic effects of having a higher insect diversity were shown by Brittain et al. (2013) for the improved pollination of almond trees when honeybees are accompanied by stingless bees. Overall, there is a great need to make better use of the potential benefits that can be generated by diversifying the insect communities in agroforestry systems. Again, also this dimension underlines the pivotal importance of diversifying the production systems with more plant species, not

2066

R.G. Muschler

only for providing a broader palette of products but also to foster the survival and services of an effective and synergistic associated and auxiliary biodiversity (Vazquez 2014; Nicholls and Altieri 2013).

Selection and Use of Microbial Symbionts and Other Beneficial Soil Organisms As mentioned in the section on agroforestry benefits for soil fertility, increasing attention should be given to the active selection and more widespread use of freeliving and symbiotic microbes and soil fauna, particularly earthworms. Of the more than 6000 species of earthworms, only about 10 have been studied in some detail (Lavelle et al. 1999). There is a great need to advance our knowledge about their ecology and management in order to maximize their benefits for the overall productivity of the agroforestry systems of the future. The impressive developments in composting and vermiculture, i.e., the use of earthworms to transform organic materials into high-quality fertilizer, demonstrate the vast potential (Benzing 2001; Aranda et al. 1999; Beck 1997). One illustrative example was reported by Castello´n et al. (2000), who demonstrated that adding 25 % of composted or vermicomposted coffee pulp to the substrate of organically grown coffee seedlings stimulated growth and plant health just as much as the conventional fertilizer and fungicide treatment. Therefore, the conventional treatment with synthetic chemical inputs can be substituted with no harm by biological treatments. This is but one example for mitigating climate change, as discussed in section “Multifunctionality of Agroforestry: Climate-Smart Production, Protection and Ecosystem Services.” At least 90 % of all plant species have evolved in association with beneficial soil microorganisms such as bacteria, actinomycetes, and fungi as symbiotic partners (Lowenfels and Lewis 2010). Some of them, particularly rhizobial bacteria, convert atmospheric N to N forms that can be used by plants, and others, particularly fungi, form symbiotic relationships with plant roots to access soil nutrients which would often not be available to just bare roots (Hauggaard-Nielsen and Jensen 2005; Margulis 1998). Just as important as the N-fixers are for providing nitrogen to plant associations (Jalonen et al. 2009; notice that natural forest ecosystems tend to have an adequate supply of nitrogen as evidenced by the color of their canopy), so are mycorrhizal fungi for improving physical, chemical, and biological soil quality (Bardgett and van der Putten 2014; Jordan 1985). The association of plant roots with mycorrhizal fungi increases the active surface of the now greatly extended root system of the plant-fungus complex by several orders of magnitude, multiplying the capacity of the associated plant to access poorly available nutrients, particularly P and some micronutrients like Zn (Lowenfels and Lewis 2010; Stamets 2005; Sieverding 1991). Furthermore, the mycorrhizae can reduce Al and Mn toxicity, benefit N-fixing rhizobia and other beneficial soil organisms, and help protect against pathogens (Cardoso and Kuyper 2006). Finally, recent research has demonstrated how “common mycorrhizal networks” can benefit plants, even of different species and photosynthetic pathways, when they are connected by joint

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2067

mycorrhizal networks. Walder et al. (2012) demonstrated an increase of 11 % for plants of sorghum and flax when they were connected by mycorrhizae. For flax, productivity even increased by 46 % when the plants were linked to hyphal networks. These examples illustrate the great potential to be explored by optimizing the matching plants and microbial partners. Doubtlessly, for the much needed “biological intensification” of agroforestry and other systems (FAO 2013), it will become of increasing importance to identify and make wider use of free-living and symbiotic N-fixers and their most effective mycorrhizal partners which can help to greatly boost both the water and nutrient use efficiency of the inoculated crops. To reduce costs of inoculation and to avoid time lags until the symbionts are well established, increasing attention should also be given to foster agroforestry interventions, reduced soil disturbance, minimum tillage, and appropriate crop rotations so as to assure the survival and activity of these beneficial microorganisms in the soils (Mutua et al. 2014; Magdoff and van Es 2009; Cardoso and Kuyper 2006). Furthermore, given the essential functions of these symbiotic, and other freeliving, microorganisms for soil health and its suppressiveness, i.e., the biological capacity of a given soil to use its beneficial microorganisms and the soil fauna to hold in check potential pathogens and nematodes, it becomes clear that much more attention should be given to study the effects of fertilizers and pesticides on the soil microbiota (Stamets 2005; Sieverding 1991). While synthetic inputs often affect the soil flora and fauna negatively (Kimbrell 2002), the opposite is true for organic compounds from tree and crop residues produced on site or from biological inputs in the form of animal manures, composts, and waste materials. This constitutes the base for the obligate positive feedback loop of “feeding the soil to feed the crops,” which is not only the key message of concerned soil scientists (Montgomery 2012; Magdoff and Van Es 2009) but is also at the heart of permaculture (Mollison 1996), biointensive agriculture (Jeavons 2014), and biologically intensified urban agriculture in Cuba and elsewhere (Altieri 2002, 1999). Clearly, being able to produce (sustainably) more than 10 kg/m2/year of fresh organic vegetables on poor lateritic soils under stressful climatic conditions in Cuba (INIFAT 2011) gives convincing evidence of the central importance of incorporating organic matter into the soils and adding, as well as caring, for the beneficial soil microbes (Martinez-Viera and Dibut-Álvarez 2012; Lowenfels and Lewis 2010) for increased nutrient access and disease suppressiveness (Thuerig et al. 2009; Weller et al. 2002). For agroforestry practitioners, the Internet resource www.mycorrhizae.com provides easily readable materials on the “whys” and “how to” aspects of using mycorrhizal inoculants (Amaranthus et al. 2012). The textbook Mycelium Running by Paul Stamets (2005) is a comprehensive reference. For Spanish-speaking practitioners, Martinez-Viera and Dibut-Alvarez (2012) published a detailed review on bacterial biofertilizers, based on the extensive advances in this field in Cuban organic urban agriculture. For the future, the conscientious selection and more widespread use of microbial symbionts will have to play a much more important role for the agroecological optimization of agroforestry systems.

2068

R.G. Muschler

Lessons Learnt from Tree, Crop, and Animal Domestication: Widening the Search The selection of trees, crops, and animals for climate-smart agroforestry systems of the future should start from a wider scope of potential candidate species (FAO 2015; Kahane et al. 2013; Padulosi et al. 2002), including a wide range of un- or semidomesticated fruit trees and underutilized, sometimes forgotten, crops from the Amazon (Lorenzi et al. 2006; Villachica 1996), Africa (Abukutsa-Onyango 2014), and Asia (Ebert 2014; Keatinge et al. 2010) which may offer substantial nutritional benefits (Leterme et al. 1996). The breeding and selection efforts should give special attention to striking a balance between productivity and quality traits on the one hand and the plants’ capacity to resist pests and diseases as much as possible (Robinson 2007). The selection of crops to withstand stressful conditions, such as increasingly frequent and severe droughts that are likely to occur under the scenarios of climate change, must respond to a wider range of attributes beyond the standard agronomic characteristics and should also make wider use of the experiences on other continents. For example, the extensive experiences on breeding of drought-resistant varieties of sorghum and millets from Africa (Reynolds et al. 2010) are likely to become of increasing importance also for the Caribbean and for increasingly drier areas on the Pacific side of Central America and elsewhere. Another example comes from the breeding and promotion efforts on making better use of highly nutritious plants like the African nightshade, spider plant, jute mallow, vegetable cowpea, slenderleaf, African kale, and vine spinach (AbukutsaOnyango 2014). The experiences with the promotion of NUS should be shared more widely without being subject to continental limits. In the Neotropics, ancestral or new uses of traditional crops such as Portulaca oleracea, Cnidoscolus chayamansa, Brosimum alicastrum (called “Maya Nut,” a tall, nutritious, and productive tree for subhumid environments), or species of Amaranthus (“food of the gods” for indigenous Neotropical populations, but vilified by Spanish colonizers) for the production and human consumption of highly nutritious edible leaves in the form of a tropical spinach should become more important as we diversify our diets with fruits and vegetables (Keatinge et al. 2010; Box 2 and Fig. 23). As Leterme et al. (1996) have shown for Colombian lowland species, the mineral content of leaves of edible woody plants can be much higher than that of fruits and tubers, especially in Ca (280–1,242 mg Ca/100 g edible portion) and Fe (0.7–8.4 mg Fe/100 g edible portion), two elements often deficient in human diets. As the outstandingly high concentrations of Ca, Fe, and other elements in the leaves of Trichanthera gigantea (Fig. 23) show (up to more than 40 times higher than corn, bread fruit, and fruits; this species is already being promoted for improved animal nutrition), there is great potential for improving also human nutrition by adding more leaves to the diets. Interestingly, the high concentration of Ca, Fe, and Zn in the tubers of Pachyrhizus erosus (on average, five to ten times higher than most “traditional” crops; Leterme et al. 1996; Fig. 23) illustrates that there is also great potential among tropical tubers. Pachyrhizus sp. is an example of a very promising multipurpose plant to diversify production systems with an edible, drought-resistant N-fixing scrambling crop for soil protection or recuperation. At a global level, more support should be

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2069

Fig. 23 Some examples of the wide range of promising underutilized crop species which can help to diversify diets and the production systems, adding micronutrients, vitamin A, and “climate smartness.” Top, from left to right: Trichanthera gigantea, Sesamum indicum, amaranth for leaf consumption. Bottom: yam bean (Pachyrhizus erosus) is a N-fixing legume for soil improvement and produces an edible tuber. The nutritious plate is loaded with vitamins and minerals from squash, tree spinach (Cnidoscolus chayamansa), peach palm fruits (Bactris gasipaes, bottom left), beans, and a mix of tomatoes with white pieces of fresh yam bean. The drink is made from roasted sesame seeds (Photo credits: R. Muschler at “Finca Loroco,” Costa Rica)

given to identify and promote neglected and underutilized crops (NUS = “orphan crops”), for human consumption and the (demand-driven) diversification of land use systems (Jaenicke and Ho¨schle-Zeledon 2006).

Information to be Added to Crop and Tree Databases While there is a lot of information in the abovementioned databases, there is yet little information on the nutritional composition of many of these plants3 (Leterme 3

A good starting point is the USDA National Nutrient Database at www.nal.usda.gov/fnic/ foodcomp/search/index.html

2070

R.G. Muschler

et al. 1996) and on their average productivity that could be expected given appropriate conditions for growth. This information is essential for estimating the total nutritional contributions that could be derived from individual plants (total nutritional output per plant = number or quantity of fruits or other edible plant part * weight/fruit * edible portion * nutritional concentration). In turn, such estimates are essential for calculating the number of plants needed to meet certain nutritional targets, particularly with respect to micronutrients and enzymes important to combat “hidden hunger.” Furthermore, there is often very limited, if any, information on (i) acceptable or ideal soil conditions (often the categories used are so broad to be of little practical value); (ii) the association with beneficial microorganisms such as N-fixers and mycorrhizae (species/strains, inoculation, etc.); (iii) the speed of growth that can be expected; (iv) the best management (pruning intensities, planting, etc.); (v) the importance for pollinators and other organisms; and (vi) indications about harvesting, post-harvest handling/transformation, and storage. Two other key aspects almost completely lacking despite their great importance are (vii) the ecological/agronomic compatibility with other crops and trees (in the sense of “companion planting”) and (viii) the capacity to resist extreme climatic events such as inundations, extreme droughts, and strong winds. Undoubtedly, this information, arranged in an expanded matrix approach, possibly like the one proposed by Sanchez-Salmero´n et al. (2015) will become more and more important for deciding on the best species mixes for the climate-smart agroforestry systems of the future.

Multifunctionality of Agroforestry: Climate-Smart Production, Protection, and Ecosystem Services One of the central attractives of agroforestry is its multifunctionality, which allows combining the provision of different products with effective environmental protection and resulting ecosystem services; the latter are a prerequisite for climate-smart production systems that are able to adapt to, and mitigate, climate change. Wherever the trees’ beneficial effects, including income from the sale of timber, fuelwood, and fruits, outweigh possible competitive effects, the agroforestry association will be a better land use option than a monocrop. This is likely true for large areas of the tropics where trees can alleviate climatic or edaphic constraints and supply the increasing demand for tree-derived products. Furthermore, the increasingly urgent quest for environmental sustainability and conservation of biodiversity (bees and migratory birds are two of the most widely discussed groups), the rising demand for timber, and the recent interest in C sequestration and a reduction of greenhouse gas emissions from agriculture and livestock also drive the growing interest in planting more trees in agroforestry (Nair and Garrity 2012). Still, decided efforts are needed on multiple fronts to raise the common awareness about the benefits of agroforestry systems for long-term sustainability.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2071

a) High productivity (favorable environment) System productivity (crop + tree products)

System with trees

Sys

tem

with

out

tree s b) Intermediate productivity (environment with some limitations)

c) Low productivity (harsh or degraded environment)

Time (decades)

Fig. 24 Conceptual graph to illustrate the tendencies of system productivity (i.e., crop plus tree products) of land use systems with trees (solid lines) compared to systems without trees (dotted lines). Scenarios a–c indicate that the tendencies vary according to the initial productivity as a function of environmental factors, particularly soil fertility and climatic patterns. Although trees may cause temporary reductions as the immature trees compete with the crops without providing products themselves, well-managed agroforestry systems tend to generate benefits in the long run. These are due to their production of multiple outputs and their contributions to conserve or even improve soil fertility, water retention capacity, and beneficial biodiversity. The long-term ecosystem benefits of agroforestry systems are now widely recognized as essential for designing climatesmart land use systems

One starting point for this is to recognize that short-term losses during the first years of establishing agroforestry systems, if they occur, are usually over-compensated by long-term benefits (Fig. 24). The four key messages from Fig. 24 are as follows: (i) agroforestry systems tend to maintain their productivity including ecosystem services over time, while treeless systems tend to decline in the long run due to the loss of soil, water, or biodiversity; (ii) there is a time lag of some years after establishing agroforestry systems during which the total productivity may be lower compared to a treeless system, because the juvenile trees start to compete with the crops without providing products yet. However, this short-term disadvantage is usually compensated by the long-term effects; (iii) the productivity decline in treeless systems is likely higher in environments that have initially high productivities, because there are more natural resources to be lost over time, and (iv) in harsh or limiting environments, the benefits from including trees can lead to a longterm net increase of system productivity due to slow improvements of soil fertility and the recuperation of other natural resources. The following sections present

2072

R.G. Muschler

some salient examples of how agroforestry contributes to the conservation of biodiversity and ecosystem services on the one hand and to climate-smart production systems on the other.

Agroforestry for Biodiversity Conservation and Ecosystem Services Agroforestry systems have more plant, animal, and microbial species than typical agricultural systems. In addition, the distribution of these species often is nonuniform, forming patches of different plant associations in the landscape. In addition, the higher species richness and patchiness of the planted components in agroforestry systems favor the presence of more associated biodiversity in the form of insects, birds, and other heterotrophic organisms. Consequently, these attributes permit that agroforestry contributes significantly to the conservation of cultivated and wild biodiversity. The authoritative multiauthor review of the contributions of agroforestry to biodiversity conservation in tropical landscapes (Schroth et al. 2004) provides a wealth of information on the linkages between conservation biology, landscape ecology, and agroforestry. It reviewed the contributions of different agroforestry practices to the conservation of wild biodiversity by establishing, or serving as, biological corridors, buffer zones, or surrogate forests. As natural ecosystems continue to be transformed and, hence, loose much of their original biodiversity, agroecosystems are becoming more and more important for the conservation of biodiversity. Agroforestry systems and agroecological practices are at the heart of this. Compared to biologically impoverish sun-grown systems, shaded coffee, cacao, or tea plantations are among the most biodiverse agroforestry systems (Philpott et al. 2008; Perfecto et al. 1996). Other examples of high-biodiversity systems include rubber agroforests in Asia, homegardens in India, Chagga homegardens in Africa, and many other complex systems elsewhere (Nair 1989, 1993). These agroforestry systems are examples of ecologically rich, and therefore more stable, systems (Rapidel et al. 2011; McCann 2000). The web of interactions among their many components reduce the probabilities of devastating pest and disease outbreaks (Staver et al. 2001; Ewel 1986) and enable these systems to sustain themselves over time, while generating income and work through the varied outputs such as coffee, cacao, timber, fruits, and fuelwood (Beer et al. 1998). When such systems are replaced by simpler ones, it is often due to economic incentives or distortions that respond to short-term objectives or ignore the real costs of negative externalities such as soil erosion or water contamination. Unfortunately, this is the case for the current substitution of many biodiverse rubber agroforests in Indonesia by rubber monocultures (van Nordwijk et al. 2012) or for the ongoing abandonment or transformation of many shade-coffee systems in Central and South America by more productive high-input unshaded systems in countries with lower production costs in Asia and Brazil (Jha et al. 2014; Philpott et al. 2008). Of course, in these cases, “more productive” refers to rubber or coffee as the only products, while the tree products and services from the shaded systems are lost. These shifts, trading long-term benefits and multiple products for short-term gains and just one product,

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2073

pose great threats due to the loss of the associated biodiversity and environmental services essential for the long-term sustainability. In Latin America, the loss of biodiversity from coffee fields is of particular concern, because most coffee production areas coincide with biodiversity “hotspots” (Hardner and Rice 2002). As a response to Daily’s (1997) classic call of attention regarding the modern erosion of “Nature’s Services” and the “Millennium Ecosystem Assessment” in 2005, which provided further evidence of a global decline of ecosystem services, there has been an increasing interest in understanding the factors responsible for this decline. Undoubtedly, as mentioned in section “Agroforestry: Evolution, Definition, Practices and Systems,” the green-revolution practices oriented toward monocultures played (and continue to play) a major role (Kimbrell 2002) besides increasing human population pressure. During the past decade, research has revealed many ways for agroforestry to contribute to biodiversity conservation (Schroth et al. 2004) and for providing and paying for ecosystem services (see Rapidel et al. (2011) for a wealth of case studies from Latin America). The major ecosystem services of agroforestry can be arranged in six dimensions (Table 4): (1) biodiversity conservation (above- and belowground; from microorganisms to mammals and trees), (2) tree/crop/animal facilitation (positive effects among the components), (3) soil conservation and enrichment (biological and chemical), (4) conservation of air and water quality and quantity, (5) carbon sequestration and climate change mitigation (including through the deposition of slow-release C in the soil as SOM and biochar), and (6) aesthetic and cultural richness. These services have high relevance for the long-term sustainability of ecosystems, as well as for the adaptation to and mitigation of climate change. Just like the conservation of functional biodiversity is key to reducing the environmental footprint of chemical plant protection agents by substituting these by (ideally, self-reproducing) biological means, so is soil enrichment key to reducing the needs for synthetic fertilizers (cf. section “Selection and Use of Microbial Symbionts and Other Beneficial Soil Organisms”). It is interesting to note that while all functions have local relevance, some also have regional relevance within the local landscape context such as the services for the prevention and management of pests and diseases or pollination (Fig. 2), and a few have even global relevance, as is the case for migratory birds which depend on habitat and food along their annual migratory routes (NABCI Canada 2012). Although our understanding has greatly increased regarding the interactions between agroforestry systems and biodiversity conservation (Schroth and Harvey 2007; Schroth et al. 2004; Stolton et al. 2000; Rice and Greenberg 2000), much remains to be discovered. Of particular importance is the question of how to optimize symbiotic and synergistic species associations to increase water and nutrient use efficiencies as well as the pest and disease suppressivity of the system (Vazquez 2014), while maximizing the products and services of the system (Vaast and Somarriba 2014). Only when we learn more about the functions of the different organisms, including birds (Sekercioglu 2012) and their potential as predators of arthropods (e.g., van Bael et al. 2008); different organisms that pollinate crops (Klein et al. 2008); others like ants, wasps, and spiders that are key for pest and disease suppressiveness (Daghela et al. 2013); and decomposers such as dung

2074

R.G. Muschler

Table 4 Spatial scale of ecosystem services of agroforestry (Expanded from Jose 2009) Ecosystem service Dimension (mechanism) Biodiversity conservation

Facilitation among components Soil conservation & enrichment

Air and water

Spatial scale Function/effect

Local (farm)

Regional (landscape)

Global

Locally adapted crops, trees, animals and microorganisms Prevention and biological control of pests and diseases Pollination/seed dispersal Migratory species (birds, mammals…) Microclimatic benefits (e.g. effects of nurse/shade trees) Symbiotic interactions (N fixation, mycorrhizae) Deep-soil nutrient capture and cycling Improved soil cover Enrichment with organic matter Soil suppressiveness (diseases, nematodes, etc.) Soil stabilization/erosion control Clear air and water (filters dust and pollutants) Fosters water retention in watersheds

Flood mitigation C storage in growing biomass and long-term deposition (e.g. as biochar in soils) Biological synergies reduce needs for synthetic inputs (fertilizers, pesticides) Improved nutrition of ruminants reduces CH4 emissions Aesthetics & cultural values Scenic beauty Food diversity & nutrition benefits Local identity/traditions C sequestration & CC mitigation

beetles which affect soil fertility greatly (Nichols et al. 2008), will we be able to better value, and foster, their services by paying attention to each organisms’ needs for survival. Clearly, additional research is needed on the links between biodiversity, ecosystem functions, and ecological services in order to optimize the system design (Kremen 2005). Recognizing the great importance of multistrata agroforestry systems with coffee and other crops for migratory birds, including many whose numbers are rapidly dwindling (some by more than 60 %, NABCI Canada 2012), the “Smithsonian Migratory Bird Center” has established certification criteria4 for the production of “bird-friendly coffee” (Fig. 25). These criteria include using at least 11, preferably native, tree species per ha, arranged in 3 strata, and growing the coffee under organic management. Using such a certification approach can contribute greatly to promote the maintenance of species-rich multistrata systems. Unfortunately, the price differential that a producer can receive for bird-friendly coffee is not sufficient for most producers to compensate for the lower productivity of coffee in such a system under organic management (Lyngbaek et al. 2001). Much needs to be done to create more consumer awareness, more fairness in the value chain (increasing the benefits to the producers), and to promote this and other certification approaches effectively (Soto and Le Coq 2011). A very useful overview of different certification criteria, including biodiversity, social, and economic standards, is available at http://www.coffeehabitat.com/certification-guide/.

4

http://nationalzoo.si.edu/SCBI/MigratoryBirds/Coffee/default.cfm

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2075

Fig. 25 The three main strata required for the certification of “bird-friendly coffee” according to the criteria established by the “Smithsonian Migratory Bird Center”

In order to help map ecosystem services, Kareiva et al. (2011) recently assembled a methodological toolbox. Undoubtedly, the application of such tools, including modeling and valuation approaches, holds great promise for the design of climate-smart landscapes including agroforestry systems at the interface of production and conservation of ecosystem services.

Contributions of Agroforestry to Climate-Smart and Multifunctional Agriculture The potential of agroforestry for climate-smart agriculture, sometimes also called multifunctional agriculture, has been amply recognized by the research and development community. One central aspect is that a higher diversity of species tends to favor stability (McCann 2000), and another one is derived from the long-term benefits of trees on soil fertility, carbon stocks, and the reduction of the needs for chemical inputs (see section on “Soil Fertility”). In its sourcebook of climate-smart agriculture, FAO (2013) recognizes that agroforestry can contribute both to climate change mitigation and adaptation through different ways. For mitigation, the main mechanisms are (i) C sequestration and retention in biomass and the soil, (ii) the substitution of synthetic inputs by biological mechanisms (this is of particular importance for N fertilizers which may release significant amounts of N2O), and (iii) the reduction of enteric CH4 emissions from ruminants by receiving improved feed and fodder. For adaptation, agroforestry can increase the resistance and resilience of the system to climate variability because the trees buffer against extreme climatic events, protect soils and watercourses, and diversify the production (Matocha et al. 2012; Muschler 2001a): shade trees reduce heat stress on animals and crops; fruit, timber, and fuelwood species provide additional products which buffer against price fluctuations of individual products; and fodder trees supply high-quality forage to reduce grazing pressure, land degradation, and methane emissions (Thornton and Herrero 2010; Reid et al. 2004).

2076

R.G. Muschler

Table 5 Absolute and relative carbon content in agroforestry systems compared to other ecosystems

System Primary forest (>30 years) Secondary forest (25–30 years) Coffee agroforestry system

C in biomass (t C/ha) 138

C in soil (t C/ha) 194

Total C in ecosystem (t C/ha) 332  37

(%) 100

Sources Cifuentes-Jara (2008)

73

184

257  14

77

Cifuentes-Jara (2008)

38 14.3–43.5

127

164  30

50

66

51

117  47

36

Callo-Concha et al. (2002) Van Rikxoort et al. (2014) Somarriba et al. (2013)

Cacao Agroforestry system Extensive silvopastoral system Improved pasture

51

63

114  18

35

Hassan (2011) and Ibrahim et al. (2007)

28

81

109  12

33

Degraded pasture

2

56

58

17

Hassan (2011) and Ibrahim et al. (2007) Hassan (2011) and Ibrahim et al. (2007)

The importance of agroforestry systems for climate-smart production systems is linked to their ability to maintain relatively high levels of carbon in living biomass and soils (Table 4), as well as high levels of biodiversity; this places these systems between agricultural and forest systems (Fig. 26). The data in Table 5 locate representative agroforestry systems in the range of one third to one half or more of the total C stock in climax forest systems. Obviously, systems with very few trees such as open parklands, severely degraded sites (see the examples of Haiti or some African experiences), or extremely dry systems will have lower values. More data on mitigation of different agroforestry systems can be consulted in Nair 2012. Figure 26 demonstrates how the transition from forest systems to less complex agroforestry systems, to agricultural systems, to pastures and, finally, to degraded lands leads to drastic reductions of biodiversity and carbon stocks, following roughly the numbers of Table 4. Just like the arrows labeled “mismanagement” indicate a reduction of C and biodiversity stocks when the resources of soil and biodiversity are not adequately protected, the reverse process allows recuperating these stocks, at least partially. However, an essential difference is the speed of these processes: while the degradation can happen in just a few years or even less time given extreme climatic events, the recuperation tends to require many years to decades of decided management and effective protection, as demonstrated by the successful examples of recuperating upland systems in Haiti through the construction of artificial lakes and

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2077

Fig. 26 The principal land use systems can be arranged by their carbon stocks and biodiversity levels. This conceptual graph situates the land use systems relative to each other to visualize the changes in carbon and biodiversity as one system gets transformed to another. The width and height of the ovals indicate varying ranges for different land use systems. For the sake of simplicity, the graph does not account for special systems with exceptional values. Notice that the three principal groups of agroforestry systems, highlighted in bold letters, are found at intermediate to high levels of both parameters indicating the potential of agroforestry systems for combining climate-smart production with biodiversity conservation. In order to be truly sustainable, land use systems should aim for the top right quadrant and preserve and/or increase carbon levels and biodiversity as far as possible

the establishment of agroforestry systems. Today, these systems provide, again, both products and the protection of environmental resources and services. As long as degraded ecosystems retain a sufficiently high5 stock of natural resources (soil, water, and biodiversity) to allow recuperating the functions of the ecosystems, the losses of biodiversity and C due to land use changes and mismanagement are, at least partially, reversible. Plants and animals can be reestablished, or may even migrate back in as the habitat conditions improve, particularly when native species are being used and connectivity is given (Montagnini and Finney 2011), and C will accrue as plants grow.

5

The definition of what is “sufficiently high” depends on the ecological factors which determine or limit the capacity of “reconstruction.” Clearly, once all topsoil has been eroded, plant growth in the subsoil is greatly inhibited and the recuperation may be limited to the much slower processes of “primary succession” rather than “secondary succession,” requiring decades or even more time.

2078

R.G. Muschler

According to Montagnini and Nair (2004), smallholder agroforestry systems in the tropics can sequester around 1.5–3.5 t of C ha1 year1, leading to average aboveground C stocks of about 10 t ha1 in semiarid regions, 20 t ha1 in subhumid, and 50 t ha1 in humid regions. Considering both above- and belowground C, the C sequestration potential for agroforestry ranges from about 0.3 Mg ha1 year1 for Sahelian fodder banks to more than 15 Mg ha1 year1 in species-rich systems in the humid tropics such as Puerto Rico (Nair et al. 2009). In general, the C sequestration potential increases with rainfall, soil fertility, and temperature. Carbon stocks in soils tend to be one to two orders of magnitude larger than those of aboveground biomass (Table 6). The C sequestration potential (CSP) is an estimate of the quantity of C that can be added to recently planted systems until they reach their C saturation indicated by the highest values of C stocks. Considering that the systems that cover the largest areas are intercropping systems (650 M ha), followed by silvopastoral systems (450 M ha), and protective systems (300 M ha), it becomes evident that improvements in these systems represent the largest potential for C sequestration. However, even multistrata systems, although represented on less than 10 % of global agroforestry coverage (about 100 M ha), can also contribute substantially since the total amount that can be stored in these systems is higher than in any other agroforestry system. Adding, furthermore, the potential reductions of N2O emissions by substituting some of the synthetic N fertilization in high-input systems through biologically fixed N, the mitigation impact of agroforestry becomes even higher. These aspects are at the heart of “nationally appropriate mitigation actions” like the Costa Rican pilot “NAMA coffee” project starting in 2015. This project, funded under the “International Climate Initiative” by Germany and the UK, is one of the first, globally, to explore the full potential of agroforestry for climate change mitigation in the coffee sector. Assuming a median C sequestration potential in biomass and soil of 94 Mg ha1, and that 585–1,215 M ha of tropical lands are, or could be, under agroforestry, Dixon (1995) estimated the global C sequestration potential over 50 years as 1.1–2.2 Pg, a figure that was adjusted to 1.9 Pg C for 1,023 M ha by Nair et al. (2009). Obviously, improving the vast amount of degraded croplands and pasturelands with AF practices holds an enormous additional potential to sequester carbon. Recently, Kumar and Nair (2011) provided a global assessment of the carbon sequestration potential of agroforestry systems. Besides a comprehensive reporting of data, this volume also raises the issue about the large variability of data, and data reliability, and makes a call for a more rigorous reporting and standardization of research methodologies used for assessing C stocks and sequestration potential. According to Kuyah and Rosenstock (2015), appropriate allometric relationships based on the simple measurement of the DBH of trees may still represent the best and cheapest way to obtain reliable and accurate data on C stocks in agroforestry systems. For most purposes, the negligible improvement of predictive power (1.3 %) by including additional variables such as tree height and crown diameter may not justify the additional time investment for measuring the additional variables.

Arid and semiarid, primarily sub-Saharan (Africa, China, and N and S America) Grazing systems, predominantly in semiarid to subhumid lands in (Africa, India, and the Americas) Firewood and fodder-tree systems mostly in tropics; land reclamation plantings in special problem areas

Region (including potential) Humid and subhumid tropics Temperate regions (North America, Europe) Humid and subhumid tropics, mostly lowlands, but up to 2,000 m elevation

2–15

450

1–12

2–10

300a

50

2–18

Up to 140

Up to 250

Up to 100

Up to 300

Estimated C stock (Mg ha1) Aboveground Belowground Up to 15 Up to 150 Up to 10 Up to 200

100

Planted in linear rows; the area refers to the area protected by the protective plantings

a

Woodlots (firewood, fodder, land reclamation, etc.)

Multistrata systems (shaded perennials, homegardens) Protective systems (windbreaks, shelterbelts, riparian buffers) Silvopastoral systems

Group of AFS Intercropping systems (incl. alley cropping)

Area (Mio ha, incl. potential) 650 50

1–5

3–10

1–8

2–10

40–70

80–120

20–60

100–200

Estimated CSP on new plantings (Mg ha1) Aboveground Belowground 2–5 25–75 2–6 50–150

Table 6 Global estimates of area, carbon stocks, and carbon sequestration potential (CSP) for the major types of agroforestry systems (Modified from Nair 2012). CSP corresponds to the estimated difference in carbon stocks between a recently established and a mature system in each category

Agroforestry: Essential for Sustainable and Climate-Smart Land Use? 2079

2080

R.G. Muschler

Many agroforestry and agroecological interventions permit combining adaptation with mitigation objectives. Examples include the planting of trees and woody shrubs, the protection of soils and protected areas, and the fostering of beneficial interactions among the components of the systems (Table 7). Integrating these practices into land use mosaics increases the capacity of landscapes to resist extreme climatic events.

Design and Modeling of Agroforestry Systems This section provides recommendations and tools for the design and modeling of agroforestry systems oriented toward long-term sustainable production and the provision of environmental services. In their thematic spread, the recommendations in this section expand the work of Jose and Gordon (2008) who compiled a series of review articles and case studies on above- and belowground resource allocation and on modeling approaches for the design of agroforestry systems. One of the foremost objectives for agroforestry design is to maximize production in a sustainable manner, i.e., “ecological intensification.”

Designing for Agroecological Intensification As shown in previous sections, the services of trees (e.g., shade, deep nutrient extraction, or microbially mediated nutrient access) can often be used to alleviate or overcome microclimatic or edaphic limitations. At the same time, trees diversify farm income, feed animals, and increase the resistance and resilience to extreme climatic events, and the higher levels of biodiversity in species-rich agroforestry systems help prevent or control pest and disease outbreaks (Vázquez-Moreno 2014; Newton et al. 2011; Staver et al. 2001). Hence, trees can be essential for the “ecological intensification” and for the long-term sustainability of systems with coffee (Muschler 2001a, b, 1998), cacao, and other crops. As Somarriba et al. (2013) have shown, cacao systems can be designed to combine high yields of cacao with that of the associated trees (Somarriba and Beer 1987), which, in turn, are essential for the provision of ecosystem services, including carbon sequestration (Vaast and Somarriba 2014). The key tools for such “ecological intensification” are the use of selected cacao materials that combine high productivity with quality and resistance to pests and diseases (Phillips-Mora et al. 2013); optimum plant spacing for cacao and associated crops and trees (Somarriba and Beer 2011); the selection of appropriate tree species based on their architecture, phenology, and functional traits (Tscharntke et al. 2011); and an intensive phytosanitary and agronomic management. To maximize C stocks while minimizing excessive shading to cacao, Somarriba et al. (2013) suggested to use tree species with (1) tall, cylindrical, and thick stems (a “sequoia” type of tree); (2) small canopies and small, light foliage; (3) deep and thick roots; (4) rapid growth; (5) high-density timber; and

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2081

Table 7 Benefits of land use practices for climate change adaptation and mitigation. These practices can be part of, or should be associated with, agroforestry interventions (Modified and expanded from Matocha et al. 2012) Practice Tree planting in crop and pasture systems or on unproductive land (agroforestry systems such as shade trees, alley cropping, living fences, windbreaks; reforestation)

Adaptation benefits Microclimatic benefits by reducing the impact of extreme weather events on crops, pastures, and animals Soil protection and fertility improvement by perennial root systems Product diversification reduces vulnerability

Mitigation benefits Increased carbon storage in biomass and soils Nitrogen from biological N-fixation can reduce fertilizer requirements (reducing liberation of GHG)

Practices to maintain or increase long-term soil fertility, including conservation agriculture, biointensive production, etc. (e.g., cover crops, intercropping with perennial grasses/ shrubs, mulching, use of compost and manures, minimum/no tillage)

Moderate soil temperatures Improve water retention and soil protection Increase soil biological activity for increasing nutrient and water use efficiency (mycorrhizae, N-fixers, beneficial microorganisms for suppressiveness) Reduced dependency on external inputs Improved soil retention and protection against flooding and landslides Ecological services from pollination and from prevention and biological control of pests and diseases Conservation of terrestrial and aquatic biodiversity Regulation of water flows Improved access and use of nutrients and water Ecological services from prevention and biological control of pests and diseases

Increased carbon storage in biomass and soils

Restoration and conservation of “protection areas” such as buffer zones, forest corridors, riparian forests, mangroves, and wetlands

Protection and fostering of symbiotic or beneficial associations between plants, animals, and microorganisms

Nitrogen from biological N-fixation can reduce fertilizer requirements (reducing liberation of GHG)

Key references Akinnifesi et al. (2010), Douglas (2009), Hergualc’h et al. (2012), Kumar and Nair (2011), Montagnini and Nair (2004), Muschler (1998, 2001a), Nair (2012), Schroth et al. (2004), and Somarriba et al. (2013) Amundson et al (2015), Bardgett and van der Putten (2014), Lal (2015), Trumper et al. (2009), and Young (1989)

Increased carbon storage in biomass and soils

FAO (2013), Heller and Zavaleta (2008), Pyke and Andelman (2007), and Scherr and Sthapit (2009)

Reduced dependency on external inputs (reduced liberation of GHG)

Bardgett and van der Putten (2014), Gliessman (2015), and Vazquez (2014)

2082

R.G. Muschler

(6) an inverted phenology to reduce shading and phytosanitary problems during the rainy season. When these recommendations are complemented by adequate shade and phytosanitary management, as well as by the other agroecological practices mentioned in previous sections, the cacao agroforestry system can meet the challenges of “ecological intensification” (Fig. 27). This way, the loss of ecological services from eliminating trees in high-input systems (the “simplification” arrow on the path to highest productivity but also highest vulnerability in full-sun systems) can be reverted to create an ecologically enriched system, which combines high production with the provision of environmental services. Notice that in this case of “intensified cacao agroforests,” the species composition of the trees and the genetic composition of the cacao clones will be drastically different from that of low-productivity unimproved rustic cacao systems. While structurally and functionally similar, the improved systems are much more productive and, hence, economically viable than the rustic systems.

Factors for the Tree/Shade Decision The decision whether to use trees, which ones, and how many in a given field depends also on the environmental factors, particularly the microclimatic and soil conditions, the production goals, the conservation objectives, and the availability of labor, inputs, and financing. For coffee and cacao, and possibly most crops, these factors can be arranged by objectives, environmental factors, and inputs (Fig. 28). Whenever the objectives of production and protection include aspects of climate-smart production, the conservation of soils, water, and biodiversity, as well as the production for specialty markets such as organic or bird-friendly coffee, the system will typically require the presence of trees. The same is, of course, the case when the outputs of the trees play a major role and when their services are needed to alleviate environmental limitations such as low soil fertility, lack of water (as long as this is not extreme – see discussion on “Alley Cropping”), high temperatures (for Arabica coffee), or wind. In contrast, unshaded systems oriented toward maximizing the production of a single product tend to be favored under ideal biophysical conditions and when sufficient inputs are available to supply the higher needs of sun-exposed plants (Muschler 2004). For coffee systems, Muschler (1998, 2004) provided a detailed account of the effects of tree and shade management on coffee productivity, quality, and environmental factors.

Designing Agroforestry Systems for Ecological Sustainability To maximize the productive potential and long-term ecological resilience of agroforestry systems, the following recommendations should be considered:

Fig. 27 Transition between cacao systems of different complexity. While the simplification of the system eliminates competing plants, it also eliminates beneficial functional biodiversity. Therefore, species-poor systems suffer a loss of ecosystem services, become more susceptible to pests and diseases, and depend strongly on external inputs. The reverse path, i.e., the diversification of cacao systems with carefully selected improved cacao materials combined with compatible crops and trees, and accompanied by active agronomic and phytosanitary management, creates systems that are more resistant and resilient to pests and diseases as well as to climatic and price fluctuations. The agroecological intensification will lead to a productive intensified cacao agroforest that is botanically different, yet functionally similar, to the rustic low-productivity system. Its goal is to combine higher productivity of high-quality cacao with a maximum of environmental services (Drawings from Rice and Greenberg 2000)

Agroforestry: Essential for Sustainable and Climate-Smart Land Use? 2083

2084

R.G. Muschler

1. If production/protection goal is no no no no no

NO SHADE, simple system

Climate-smartness yes Soil & watercons. yes Biodiversity cons. yes Specialty markets yes Additional products yes

2. If environment has high much high no

Soil fertility Moisture Elevation Wind

low little low yes

SHADE, high diversity

3. If inputs are yes Financing no yes Agrochemicals no High-yielding varieties Traditional varieties

Fig. 28 The three principal groups of factors which determine the decision about growing coffee, cacao, or other plantations crops under shade or not. The combination of the site-specific factors with the production goals and the tree attributes determines the number and diversity of trees to be planted and the appropriate management (Modified from Muschler 2004). For details see text

1. Use plants of different ages, sizes, and attributes to generate the highest possible structural and functional diversity in the system. 2. Associate compatible crop and tree species (e.g., coffee and Erythrina spp. or Inga spp.) and apply the experiences of “companion planting”6 (e.g., Cunningham 2000; Riotte 1998). 3. Aim for a certain genetic diversity within each crop or tree species and avoid monocultures of clones or hybrids as much as possible7; the recommendation to plant “polyclones” in improved cacao plantations (Phillips-Mora et al. 2013) is probably of universal validity.

6

In agroecological crop production, the association of compatible crops, such as tomatoes intercropped with carrots, can prevent and suppress diseases and may increase production; see also “companion planting” and “list of companion plants” at Wikipedia. 7 The negative experiences with the monoclonal large-scale plantations of banana varieties illustrate the high susceptibility of such plantations to specialized diseases such as Mycosphaerella or to nematodes. When highly productive clones of cacao are planted, it may be best to establish “polyclones,” i.e., mixes of different clones, in the plantations to reduce the risks associated with disease and pest susceptibility (Phillips-Mora et al. 2013).

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2085

4. Incorporate aromatic repellent and trap crops where possible (VázquezMoreno 2014). 5. Rotate crops of different types such as grain crops with tubers, leafy crops, and N-fixing crops. 6. Aim for permanent soil cover/protection and shading (live and/or dead plant cover). 7. Favor prevention and biological control over chemical options. 8. Minimize the use of agrochemicals, particularly herbicides and fungicides that impoverish the soil biota (rhizobia, mycorrhiza). 9. Integrate patches of land use under agricultural, agroforestry, animal husbandry, or forestry uses into climate-smart landscapes (Perfecto and Vandermeer 2010).

To optimize the system performance and long-term ecological sustainability, mechanisms for efficient nutrient cycling and agroecological pest management need to be built into the systems. With regard to nutrient management, a massive body of information is available on nutrient budgets, which were emphasized particularly in early agroforestry research until the 1990s (e.g., Sanchez 1995; Fassbender 1993; Nair 1993; Fassbender et al. 1991; Beer 1988). Since then, more attention has been given to understand the processes and factors that determine the stocks and speed of transformation of nutrients in plants (Akinnifesi et al. 2010; Jalonen et al. 2009; Mafongoya et al. 2000; Cadisch and Giller 1997) and their roles for soils and animals (Mafongoya and Hove 2008). Work has been done on nutrient budgets, fractions, and decomposition patterns as a function of (a) climatic and edaphic factors, (b) tree spacing and management (pruning intensity, e.g., Muschler 2004), (c) the type and attributes of trees (often N-fixers versus non-fixers, e.g., Budelman 1988), (d) degradability of tree litter depending on their composition and the presence of polyphenols and tannins (e.g., Mafongoya et al. 2000) ranging from easily degradable materials (e.g., Erythrina spp., Gliricidia sepium, or Leucaena spp.) to more persistent litter (e.g., Inga spp., Cordia alliodora, or Eucalyptus), and (e) the composition of simple versus complex litter or compost mixes, some with animal manure or other microbial additions such as efficient microorganisms (EM). One illustration of the potential importance of the last point is the recent work by Barantal et al. (2014) on the speed of decomposition as a function of decomposing fauna, litter composition, and nutrient stoichiometry. The leaf litter of six tropical tree species decomposed faster when leaves of different species were mixed rather than when individual species decomposed alone; nutrient addition experiments demonstrated that the relatively slower decomposition of leaves of any single species was due to C, N, or P limitations for the decomposing organisms. The authors demonstrated that stoichiometric dissimilarity of litter mixtures (i.e., the divergence in C/N/P ratios among species) can speed up decomposition by fostering the activity of the decomposers. This study shows a mechanism for modifying decomposition patterns.

2086

R.G. Muschler

Undoubtedly, the information from these studies is key for maximizing the supply of organic matter and nutrients from trees and for synchronizing their availability with the needs of the crops. While much knowledge has been generated specifically from agroforestry systems (e.g., Akinnifesi et al. 2010; Beer 2000; Beer et al. 1998), often including animal components (Mafongoya and Hove 2008; Benavides 1994; Kang et al. 1990), much of it also comes from the fields of agronomy or forestry. The rapidly increasing research also in these fields offers the opportunity to make wider use of the relevant information from all three areas for optimizing agroforestry systems. Another central aspect for long-term sustainability is the creation of climatesmart and pest-suppressive landscapes, based on an improved understanding on how different land use types interact at the landscape level and how to better link production with the provision of environmental services. For agroforestry systems with coffee, the reviews of Staver et al. (2001) and Avelino et al. (2012) provide central recommendations for preventing or reducing pests and diseases. The two principal avenues are (a) the establishment of site-specific optimum shade conditions and (b) the creation of biodiverse landscapes with patches of different land use. Both approaches will reduce the spread of the pest complex and maximize the effects of beneficial microflora and fauna acting against it. Obviously, the appropriate selection of tree species, the best density and spatial arrangement, as well as the optimum shade management regime are critical decisions. These recommendations apply also, in principle, to other crops. In general, higher structural and botanic diversity tends to generate higher pest and disease suppressiveness (Vázques-Moreno 2014; Altieri et al. 2005), hence reducing the need for pesticides, while simultaneously increasing the resistance and resilience the extreme climatic events. Besides these factors essential to reduce pest and disease incidence, there is also increasing evidence that the resistance of plants to pests is significantly affected by their nutrition. Since the use of high doses of inorganic N tends to lower pest and disease resistance of plants, more emphasis should be placed on keeping soil fertility high enough to provide N primarily from organic sources (Altieri and Nicholls 2003). This is also in the interest of climate-smart production aiming at mitigating greenhouse gas emissions. However, more studies are needed to understand the interactions between pest populations and plants treated with synthetic versus organic fertilizers. Another research priority should be to learn more about how to maximize the synergies when different species collaborate to provide ecological services such as pollination (Brittain et al. 2013), pest prevention or control (Vázques-Moreno 2014), or sustaining soil fertility and health (Amundsen et al. 2015; Bardgett and van Putten 2014; Akinnifesi et al. 2010), including the transformation of organic wastes by surprisingly important dung beetles that turn out to provide a range of important ecological functions (Nichols et al. 2008). Besides these agroecological factors, also aspects of human nutrition should be considered for the design of sustainable agroforestry systems so as to address the challenges of food and nutrition security.

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2087

Designing Agroforestry Systems for Nutrition As the global diet is getting increasingly more homogeneous (Khoury et al. 2014) and unhealthy, with an oversupply of cheap carbohydrates and fats, we should increase the amount of vegetables and fruits in our diets to reduce nutritional deficiencies (Ebert 2014; Frison et al. 2011; Keatinge et al. 2010). Such changes require marked adjustments to our systems of production, promotion, and distribution of food. Healthier diets with more green, red, and orange vegetables on our plates should go hand in hand with the diversification of our production systems with such crops. Species-rich agroforestry systems are a good starting point. The fastest way would be by including highly nutritious annual crops to quickly improve the nutritional output of our production systems (e.g., Marten and Abdoellah 1988). Obviously, changes in fruit and nut production from trees need more time. Work is needed on both fronts to make optimum use of all available resources. In many places, this means also reviving nutritious traditions (Fallon 1999) such as the consumption of edible “weeds” and other plants that have been lost over the past decades. Fortunately, today, there is a wealth of online information on edible “weeds” and leaves of crops such as manioc, tree spinach or “Chaya” (Cnidoscolus chayamansa), Moringa oleifera, sweet leaf or “katuk” (Sauropus androgynus), Amaranthus spp., lamb’s quarters (Chenopodium album), purslane (Portulaca oleracea), chickweed (Stellaria media), and many others (Ebert 2014). Ideally, the selection of such crops should be based on desirable agronomic attributes and high nutritional value. For a humid tropical environment, a simple example of a species mix for a complete diet was given by Thornton (2009). Clearly, more efforts will have to be dedicated to custom-tailor the species mix to the nutritional needs and the agroecological environment of each specific region, with an increased attention to the potential of NUS (Sánchez-Salmeron et al. 2015). The USDA “National Nutrient Database” (USDA 2015) can be used as a key reference for the nutrient contents of more than 8,000 types of food. Considering the multiple essential functions of trees besides nutrient provision, such as soil and water protection, biodiversity conservation, microclimate moderation, etc., it becomes clear that the judicious design of agroforestry systems must be based on a holistic balance of the multiple functions of its perennial and annual components. Of particular importance are the capacities of the components to provide useful products and large quantities of biomass for maintaining or improving soil fertility (one of the central elements of “biointensive” production; Jeavons 2014) and for improving the ecological balance in the system (Vázquez-Moreno 2014). The following training materials give practical advice for the design of such systems.

Training Materials for Promoting and Designing Agroforestry Systems One of the indications of the evolution of agroforestry is the publication of practical training materials for practitioners, extension services, and promoters of

2088

R.G. Muschler

agroforestry for development. Some of the most notable materials were assembled over the past two decades in the context of impact-oriented agroforestry projects. From Latin America, they include the basic reference by Geilfus (1994) on the multiple roles of trees and the series of training materials on agroforestry developed at CATIE (in Spanish) on “functions and applications of agroforestry systems” (Jime´nez et al. 2001), “trees in line plantings” (Mendez et al. 2000), “silvopastoral systems” (Pezo and Ibrahim 2001), “traditional tropical homegardens” (Lok 1998b), “trees in coffee systems” (Muschler 2001a), and “taungya systems” (Schlo¨nvoigt 1998). Other examples are the “Manual Keba Sula” (PAF-Ngo¨beBugle´ 2003) developed for work with indigenous groups in Panama on sustainable management of natural resources; the manuals on organic coffee production by Figueroa et al. (1998), Castan˜eda and Castan˜eda (2000), and Christiansen (2004); as well as the detailed review by Benzing (2001) on the mechanisms for sustaining organic production systems in the tropics. These illustrated and didactic manuals continue to be key references for Spanish-speaking practitioners. An extensive body of similar materials is available also from other parts of the world. However, unfortunately, their circulation is often limited when projects end and their reproduction stops. Some are available as (bulky) pdf files, but individual topics and figures cannot be searched easily. To facilitate access to these materials, it would be desirable to create a universally accessible electronic training platform where the individual figures and topics of these manuals are searchable by keywords and in multiple languages, possibly adopting the approach of Wikipedia. To facilitate the access to agronomic information for agroforestry, we should make more use of information available through initiatives or organizations like OISAT (2015), “Crops for the Future,” Bioversity International, and “Practical Action” (2015) and incorporate experiences from related fields such as biointensive agriculture (Jeavons 2014), permaculture (Mollison 1996; Permaculture Design n.d.), and urban agriculture (INIFAT 2011). Since agroforestry systems require a relatively long time to deliver all of their products and services, modeling of these systems has become an important tool for their design.

Modeling Agroforestry Systems The increasing climatic variability and resultant stresses for agroecosystems heightens the need to predict likely changes at different regional and temporal scales. Such predictions are essential for designing appropriate measures of adaptation and mitigation of climate change. Ideally, the models should allow to generate predictions at different regional and temporal scales, ranging from changes and recommendations at a national or regional level all the way down to the level of individual growers. However, due to their great vertical, horizontal, and temporal variability, agroforestry systems represent a formidable challenge for modeling. This is further complicated by the complexity and management of the generally nonlinear and often unknown interactions among the many species, which may be more than 50 in homegardens. Nevertheless, there is a growing field of emerging approaches to

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2089

generate useful and accurate predictions. For an overview of the most widely used approaches and models, the reader may wish to consult the online tools provided by the “World Agroforestry Centre,” the compilation of modeling approaches by Rapidel et al. (2009), and the section on modeling in Jose and Gordon (2008). Some of the models under development include the following fields and approaches: (A) Biophysical models to simulate the flow of light, water, nutrients, carbon, and other substances: – “WaNuLCAS,” a model to simulate the flows of water, nutrients, light, and carbon in agroforestry systems. This model, based on tree-crop architecture, physiology, and soil science, can be used for exploring positive and negative interactions for different combinations of trees, crops, soil, climate, and plant management (http://www.worldagroforestry.org/sea/Products/ AFModels/wanulcas/index.htm). – The “CASTANEA” model (Le Maire et al. in Rapidel et al. 2009) to simulate the carbon and water balances in homogeneous forests and, potentially, agroforests. – “Shade Motion 2.0” (Quesada and Somarriba in Rapidel et al. 2009) simulates tree shade patterns and generates a graphical representation of the shade patterns in a plot over the course of a day as a function of the characteristics and dimensions of the individual trees. – The “Agricultural Production and Externalities Simulator” (APES, www. apesimulator.org) to model the biophysical performance of agricultural and agroforestry production systems (Casellas in Rapidel et al. 2009). The outputs include plant growth and phenology, water and N soil dynamics, water, N and light competition, root profiles, the fate of pesticides, and a management module. – A biophysical conceptual model for coffee agroforestry systems making use of the “Agroecological Tool Kit” (Rebolledo et al. in Rapidel et al. 2009). – Plot-scale modeling of agroforestry systems with coffee (van Oijen et al. in Rapidel et al. 2009) to predict productivity, N leaching, N losses to the atmosphere, as well as the loss of organic C and N in surface runoff. This model separates the agroforestry system into sun-exposed and shaded regions. (B) Ecological models to simulate the behavior of biodiversity, pests, and diseases: – The “Agricultural Production Systems Simulator” (APSIM; Huth and Carberry in Rapidel et al. 2009), a model developed in Australia that evolved from simulating biophysical processes in farming systems into a decision-making tool for land managers. It has been used for modeling crop production and economic performance of agroforestry systems such as crops with windbreaks. Web-based tools allow incorporating climatic information and biodiversity benefits. – “OLYMPE” (INRA, France; Deheuvels and Penot in Rapidel et al. 2009), “RECORD” (Bergez et al. in Rapidel et al. 2009), and other approaches may

2090

R.G. Muschler

improve decision-making to strengthen the role of farmers’ organizations for accessing markets and generating policy impact. – Economic modeling approaches for the transition of a cacao-service tree to a cacao-timber tree agroforestry system (Navarro and Bermudez in Rapidel et al. 2009). – Modeling of the impact of “payments for environmental services” in favor of rubber agroforestry systems over rubber monocultures (Villamor et al. 2013). (C) System models to simulate whole systems and economic performance and provide tools for decision-makers: – Epidemiological models, like the ones used for modeling coffee leaf rust (Avelino, in Rapidel et al. 2009), combine factors of the host, the pathogen, the environment, and the cropping systems, including management. However, they tend to be limited to statistically descriptive and predictive models; a call is made for developing more mechanistic models in the future. – Modeling habitat quality and landscape connectivity for birds in silvopastoral and riparian systems (Sanfiorenzo et al. in Rapidel et al. 2009). – Simulate the impact of biodiversity loss on ecosystem functioning (DeClerck et al. in Rapidel et al. 2009). Obviously, these are but a few examples in this rapidly evolving field. Doubtlessly, the next steps will generate a better integration of individual modeling modules to expand the complexity of the models and, hence, improve the accuracy of predictions. One example of such new integrative tools was presented by Jackson et al. (2013) to identify synergies and trade-offs among the impacts of land use change on different ecosystem services. Other examples, both from the World Agroforestry Centre, are the model developed within the “Land Degradation Surveillance Framework (LDSF)” to study carbon dynamics, vegetation changes, as well as soil functional and hydrological properties at the landscape level (http:// landscapeportal.org/blog/2/) and the Negotiation-support toolkit for learning landscapes (van Noordwijk et al. 2013) which provides a wealth of tools for fostering the development of landscapes that combine production with protection. Undoubtedly, these tools fill a key gap in linking field and farm level actions with ecosystem service provision at landscape scales.

Conclusions and Recommendations: Priorities for Research and Development Over the past four decades, a solid body of research has revealed the potential of agroforestry for increasing or maintaining system productivity while protecting natural resources and providing environmental services (cf. Fig. 24). Thus, agroforestry is well suited as a central tool on the path toward “sustainable intensification” (The Montpellier Panel 2013), a new name for the quest of increasing production and conservation at the same time. The recent call by FAO in its 2014 “International Symposium on Agroecology for Food Security and Nutrition” that the future

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2091

paradigm of land use should be based much more on biology and agroecology, rather than on chemistry and fossil fuels, endorses also a more widespread use of agroforestry and agroecological practices for food production. Now, it is up to us to apply the powerful tools at hand to design and custom-tailor agroforestry practices and systems to the needs of each particular ecozone or community. In order to fine-tune the systems to the respective environmental and socioeconomic conditions, and rise to the challenge of sustainably producing more food that is less contaminated and less contaminating, we should advance in the following directions: (i) expand the species characterizations, (ii) widen the scope of crops and trees used by including “neglected and underutilized species” (NUS), (iii) strengthen work on “using” beneficial soil organisms for soil and plant health, (iv) optimize the system design and management to maximize resource use efficiency and minimize pest incidence, (v) create climate-smart and pest-suppressive landscapes based on an improved understanding on how different land use types interact at the landscape level to link production with environmental services, and, finally, (vi) advance toward more holistic socioeconomic assessments including an improved valuation of environmental services.

Characterization of Crop and Tree Species: Expanding the Passport Information Based on the fundamental importance of using locally adapted, productive, and nutritious tree and crop species that can be associated in mutually beneficial ways, greater importance should be given to expand the species characterization beyond the standard botanic and agronomic attributes of crops and trees (cf. section “Information to Be Added to Crop and Tree Databases”). To facilitate the identification of the most appropriate crops and trees for designing climate-smart agroforestry systems, i.e., systems with minimal requirements for (unsustainable) external inputs (see section “Requirements for Sustainable Landuse”), the passport information for each species should be expanded by relevant attributes (preferably in quantitative terms translated to easily understandable categories for farmers and practitioners), including the following: – Shade tolerance (at least in four categories: 75 %) – Drought tolerance (months without rain, seasonal minimum water requirement) – Heat tolerance (preferred range, tolerated range) – Wind tolerance (with an indication of strength; at least in three categories: not, moderate, strong) – Water logging tolerance (days, weeks) – Disease- and pest-tolerance (at least in three categories: high, medium, low) – Resprouting capacity after inundation (to identify fast-start crops after environmental disasters) – Soil pH preference and tolerance (preferred range, tolerated range)

2092

R.G. Muschler

– “Anti-erosion effectivity”: the capacity of its root system to retain soil on slopes (at least three categories: weak, moderate, strong, and extreme, e.g., bamboo) – Productivity of edible parts (kg/plant/year) – Nutrient and micronutrient profile (including Fe, Zn, and vitamin A content) – Space requirement per plant (m2) – Shelf life and storability – Ease of preparation and transformation as human food (energy and time requirements) – Water use efficiency (l water/kg edible product) – Compatibility with other crops or trees (e.g., highly compatible with species A, B, C; not compatible with D, E) – Ease of reproduction under tropical conditions (incl. number of seeds and storability) – Potential invasiveness With this information in a digitalized searchable database, the user can, then, easily choose the most appropriate plant components and “custom-tailor” their mix and arrangement according to the biophysical environment and the other plants in a specific place. While this information is currently not readily available for many crops or trees, its inclusion in standard databases would facilitate “filling in the gaps” as experiences around the world get analyzed. A comprehensive format is currently under development at CATIE.

Increasing the Scope for the Selection of Crops, Trees, and Animals The scope of species should be broadened by strengthening work on neglected underutilized species (NUS), also called “orphan crops” or “Cinderella species.” The increased passport information will facilitate finding the best niche for a particular species within the agroforestry system, integrated with compatible trees, crops, or animals around it. An interesting example is the widely promoted “drumstick tree” (Moringa oleifera) with highly nutritious leaves and pods, which can be planted in many arrangements and uses in agroforestry systems, including as a windbreak or living fence, in alley cropping, or as a support for climbing plants such as passion fruit, yams, or beans. Moringa can be intercropped with a wide range of vegetables such as cluster bean (Cyamopsis tetragonoloba), hot peppers, cowpeas, pigeon peas, and onions (Ebert 2014). The same holds for a long list of other species and NUS (see section “Lessons Learnt from Tree, Crop, and Animal Domestication: Widening the Search”), the selection and characterization of minor animal species for specialty systems (cf. section “Selection of Animal Species”), including bees such as the stingless Melipona and other insects of local importance, but also for warm-blooded animals and fish in their native setting. However, caution must be taken to avoid negative effects. As has been learned from the (ecologically) painful experiences in Lake Victoria after the introduction of the Nile perch (see the

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2093

documentary “Darwin’s Nightmare”8) or in many lakes in Central America after the liberation of Tilapia, the displacement of local fish species and the disturbance of local stability should also be considered before introducing exotic species. The same holds, of course, for invasive trees and shrubs (see section “Plant Selection for Agroforestry”), as well as other organisms.

Management of Beneficial Soil Fauna and Microorganisms for Soil Health and Fertility For the next quantum leap in sustainable plant production, another group of organisms of central importance, yet largely ignored, should receive the attention it deserves: beneficial soil fauna and flora. The existing studies are but the tip of the iceberg (section “Selection and Use of Microbial Symbionts and Other Beneficial Soil Organisms”). Clearly, there is a great potential of using these organisms for maintaining or increasing soil fertility and even for rehabilitating degraded soils after decades of chemical-intensive monocultures, exposure to rain and sun, or following massive erosion as in Haiti and so many other places around the world with similarly adverse conditions. After decades of a predominantly chemical focus on soil fertility, more emphasis should now be given to study soil ecology and effective agroecological management interventions (Bardgett and van der Putten 2014). Particular attention should be given to systematically study how to manage soil health for better crop health and reduced pest incidence (Altieri et al. 2005; Altieri and Nicholls 2003) by learning about the effects of: – Soil fauna on plant growth, demonstrated already as highly positive for trees, grasses, and perennial crops such as tea. Considering that only about 10 of the more than 6,000 species of earthworms have been studied in detail, more attention should be given to studying also native earthworms and their interactions with the soil and its fauna and flora (Brown et al. 1999; Lavelle et al. 1999). The importance of other groups such as dung beetles (Nichols et al. 2008) underlines the need for studying soil fauna. – Soil microorganisms, including fungi, actinomycetes, and bacteria (Cardoso and Kuyper 2006; Stamets 2005; Margulis 1998). The active management of these microorganisms should be studied in more detail both in soils and on plant surfaces. In many organic production systems around the world, microbial ferments are applied to the foliage of crops to stimulate plant growth and to prevent the growth of diseases or to minimize their impact. The positive experiences from intensive organic agriculture from around the world (Martı´nezViera and Dibut-Álvarez 2012; Restrepo-Rivera y Hensel n.d.; Benzing 2001)

8

Hubert Sauper 2004. Darwin’s Nightmare. An Austrian-French-Belgian documentary. Nominated for the 2006 Academy Award for Documentary Feature. 102 min. https://www.youtube. com/watch?v=IV7Y9FHcdFk

2094

R.G. Muschler

should be adapted to optimize agroforestry systems and combat pests and diseases (Vázquez-Moreno 2014; Altieri and Nicholls 2003). Given that fungi are among the main decomposers and provide many essential products and services to their (mycorrhizal) symbiotic partners, greater attention should be given to their active use as inoculants for plant roots and for their conservation in agroforestry systems (Sieverding 1991). Unfortunately, the use of synthetic fertilizers, herbicides, fungicides, and other pesticides leads to a drastic impoverishment of the soil biota with large negative impact on the soil suppressiveness (Thuerig et al. 2009; Stamets 2005; Weller et al. 2002). The same holds for the symbiotic partners of N-fixing legumes. While the use of mycorrhizal fungi (such as Glomus spp.) or N-fixing bacteria is already a widespread practice in horticulture and some agriculture or forestry applications, this should also become a standard practice for improved agroforestry systems. A big part of the much needed “biological intensification” will have to rely on choosing and using the right inoculant for the crop and tree species of future agroforestry systems with decided attention on generating the appropriate soil conditions, particularly SOM, that favor their activity.

Optimizing the System Design and Management for Maximum Resource Use Efficiency As mentioned in section “Design and Modeling of Agroforestry Systems,” the appropriate design of agroforestry systems should aim for maximum resource use efficiency and combine high productivity with the provision of environmental services, including pollination, suppressiveness to pests and diseases, as well as adaptation and mitigation of climate change. The central tools for this are the right choice and management of crop, tree, and associated species as a function of the environmental and socioeconomic factors. Choosing the right amount of shade and the best spacing and arrangement of the trees are essential tools. While the interactions among the components within specific agroforestry systems are increasingly understood and documented (particularly for systems with coffee, cacao, tea, rubber, and animals), more work is needed to elucidate the roles and management of associated biodiversity, including birds, insects, and microorganisms which contribute greatly to providing essential ecosystem services. Furthermore, of particular importance for long-term resource use efficiency is the wider exploration of the great potential of biochar for reconstructing and maintaining long-term soil fertility, for carbon sequestration on the scale of decades to centuries, as well as for reducing the leaching of agrochemicals and the emission of green-house-gases (Cayuela et al. 2013; Lehmann and Joseph 2009).

Creating Climate-Smart and Pest-Suppressive Landscapes Since many ecosystem services, such as the protection of soils, water, and biodiversity, but also pollination and the suppression of mobile pests, are emergent properties

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2095

of a landscape, we need to greatly expand our understanding on how different land use types interact at the landscape level. This knowledge is essential to strike the right balance between production and environmental services (Vaast and Somarriba 2014). Undoubtedly, the emerging tools for mapping and valuing ecosystem services (Kareiva et al. 2011; Rapidel et al. 2011) and for facilitating community-driven processes (van Noordwijk et al. 2013) will help greatly with the transition toward truly sustainable landscapes that satisfy the needs of production and protection.

Biophysical Research Methods and Improved Experimental Design One of the central lessons of early agroforestry research is that tree roots reach much further than expected from standard agronomic experiments. In some cases, roots of adult Erythrina poeppigiana and other trees absorbed nutrients from adjacent “control” plots without trees (sometimes at distances of more than 20 m), causing spurious and distorted results (Somarriba et al. 2001). This realization had fundamental implications for the appropriate design and analysis of agroforestry research as summarized by Coe et al. (2003). The most obvious result is to drastically increase the dimensions of the experimental plots (sometimes to more than 50 m of net plot dimensions); another one is to work with artificial root barriers or to practice root pruning. Unfortunately, the need to establish agroforestry plots on sufficiently large tracts of land (depending on the tree species, planting density, and tree pruning, individual plots may require often 0.25 ha or even more) and the need to manage them under the experimental regimes for many years, sometimes decades, put severe financial and biophysical constraints to such work. Besides the great difficulties to obtain funding for such endeavors, it is difficult to assure access to land of appropriate dimensions and homogeneity. Today, there are very few examples of such long-term studies. One successful example is the long-term experiment to test the effects of different tree species and fertilization regimes on coffee, which was set up at CATIE in 2000 with long-term funding from Norway (Box 3). A sister experiment is running in Nicaragua. Box 3. The “Mesoamerican Scientific Partnership Platform” (PCP) at CATIE, Costa Rica: Technical Backstopping for Regional Development

In order to pool research capacity for studying agroforestry systems with coffee and cacao, six organizations (Bioversity Int’l, CABI, CATIE, CIRAD, INCAE, and PROMECAFE) established in 2007 the “Mesoamerican Scientific Partnership Platform” (PCP) at CATIE. Today, this consortium provides essential technical information on productivity, product quality, disease and pest management, nutrient and gas flows, environmental services, and economic performance of different systems. The clients include national and regional institutions, as well as development projects working on systems with perennial (continued)

2096

R.G. Muschler

crops. In the photo below, a team of experts visited a strategic experiment at CATIE for studying the interactions between tree species, shade, and different input levels on organic and conventional coffee production. Information on nutrient budgets and flows is essential for optimizing the systems and for mitigating greenhouse gas emissions in coffee production. Costa Rica’s pioneer project on “Nationally Appropriate Mitigation Actions” in coffee production (NAMA-Cafe´, started in 2014 with key support by GIZ, Germany) will receive central inputs from this and related strategic experiments (Fig. 29).

Fig. 29 Top: in 2000, CATIE established its “coffee systems experiment” in Turrialba, Costa Rica. This experiment, scheduled to run for at least 20 years, is designed to study the interactions among three tree species (Erythrina poeppigiana, Chloroleucon eurycyclum, Terminalia amazonia, full sun as control), five fertilization regimes (with synthetic and/or organic fertilizers at different levels), and two varieties of coffee. The dimension of this project reflects the need for controlling the root and shade interactions among neighboring plots: the experiment occupies 9 ha and each plot occupies about 1,500 m2. Early results indicate that a significant portion of synthetic fertilizers can be substituted by organic sources, reducing nutrient losses and the emission of nitrous oxides to the atmosphere. A “sister” experiment is being run in Nicaragua. Both projects are financed by NORAD. Bottom: a 30 m tall tower in a commercial coffee farm (Cafetalera Aquiares) allows to monitor C flux at high resolution, essential information for measuring C sequestration. Scientists of CATIE, CIRAD, and other partners pool their capacities in the “Mesoamerican Scientific Partnership Platform” (PCP) (Photo credits: R. Muschler)

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2097

For studying the complex relationships between trees, crops, and soils, Schroth and Sinclair (2003) collected fundamental concepts and research methods, which remain of great relevance to date. This multiauthor landmark publication covers standard and advanced methods for studying soil organic matter, soil nutrient availability and acidity, decomposition and nutrient supply from biomass, nutrient leaching, nutrient capture, nutrient exchange with the atmosphere, soil structure, soil water, root systems, biological nitrogen fixation, mycorrhizae, rhizosphere processes, soil macrofauna, and, finally, soil erosion. For research on coffee and cacao systems, research methods were reviewed by Somarriba et al. (2001). Methods for measuring and valuing ecosystem services were compiled by Rapidel et al. (2011). For the way ahead, it is imperative to not only optimize the systems ecologically but also to transform the ecological benefits at medium and long time scales into tangible economic payments and incentives.

Socioeconomic Aspects: Paying for Externalities and Services A wealth of studies exists on the economic performance of different agroforestry systems. Examples from around the world and tools for “financial and economic analyses of agroforestry systems” were compiled by Sullivan et al. (1992) and Current et al. (1995). The assessment of the trade-offs between crop losses due to tree-crop competition or harvest damages from tree felling, on the one hand, and economic benefits from harvesting high-value timber species, on the other hand, has shown a positive balance for the agroforestry systems, particularly for coffee (Somarriba 1992) and cacao (Ryan et al. 2009). Similar benefits are often generated when fruits, medicinal plants, and other crops are included in the economic evaluations. However, due to the complexity of measuring and valuing non-tangible products or services, most of the economic studies have ignored nonmarket benefits and environmental services of agroforestry (Mercer and Miller 1998). This frequent shortcoming of economic analysis, a fundamental limitation when considering the essential importance of conserving natural resources (see sections “Setting the Stage for Agroforestry: Lessons from Monocultures” and “Agroforestry: Evolution, Definition, Practices and Systems”), is only recently receiving more attention with the evolution of metrics for measuring such services (Rapidel et al. 2011). Clearly, for a more holistic assessment of the full benefits of agroforestry systems, it will be important to strengthen the quantitative assessment of all the services provided by agroforestry systems, including their capacity to reduce agronomic and financial risks, to improve the system resilience to extreme climate events, and to contribute to biodiversity conservation and climate change mitigation. The recognition of these benefits drives the current evolution of different schemes of payment to compensate land stewards for the services they provide to society at large. For example, based on experiences from Africa, Reid et al. (2004) concluded that carbon credits for maintaining trees in savanna grasslands could contribute around 15 % of additional income to pastoralists. For many pastoralists, this was a sufficiently high incentive to retain and protect the trees. Another successful example is the evolution of “payments for environmental services” (PES)

2098

R.G. Muschler

as implemented in Costa Rica since 1997. In this country, a tax is levered on gasoline which is used to pay for environmental services such as the protection of biodiversity, soils, and water through forest conservation and reforestation and, recently, also through the establishment for trees in agroforestry systems. For details about the performance of the PES scheme and the possibilities for sustaining such a program, please consult the “National Forestry Financing Fund” (www. fonafifo.go.cr) and the evaluations by Robalino et al. (2011) and Murillo et al. (2011). As demonstrated by Villamor et al. (2013), there is a significant potential for PES also in Asian rubber plantations to increase the attractiveness of biodiversity-rich systems. Together with additional benefits from certified products for specialty markets and increased support of farmers to transform and sell their products, the overall benefits can favor the adoption and maintenance of more species-rich systems. However, integrative work is urgently needed on the best ways to combine public and private payments not only for the products but also for the increasingly important environmental services of biodiverse agroecosystems within climate-smart territories. A third example of the effectiveness of PES comes from their highly beneficial application to silvopastoral systems in Central and South America (Box 4). Box 4. “Payments for Environmental Services” (PES) for Silvopastoral Systems in Central and South America

In a project funded by the “Global Environment Facility” (GEF), the Tropical Agricultural Research and Higher Education Center (CATIE) in Costa Rica evaluated, together with FAO, the World Bank, Nitlapan in Nicaragua, and the Centre for Research on Sustainable Farming Systems (CIPAV) in Colombia, the impacts of PES on the adoption of silvopastoral systems. From 2003 to 2006, cattle farmers from Colombia, Costa Rica, and Nicaragua received between US$ 2000 and US$ 2400 per farm (equivalent to 10–15 % of their net income) to implement silvopastoral systems. Overall, the program led to a 60 % reduction in degraded pastures in the three countries, while increasing the land under silvopastoral systems, such as improved pastures with highdensity trees, fodder banks, and live fences. The environmental benefits associated with the project included a 71 % increase in carbon sequestration (from 28 M t CO2-eq. in 2003 to 48 M t in 2006). At the same time, milk production increased by 10 % and farm income by 115 %, while herbicide use dropped by 60 %, and the use of fire to regenerate the pasture is now less frequent (FAO 2010). These positive changes give ample evidence that PES can be a viable avenue to foster positive change. While the mentioned benefits and their economic equivalents should reach all members of the families and communities providing the services, there is also a great challenge to foster gender equality. Analyzing gender issues in agroforestry in Africa, Kiptot and Franzel (2012) have recommended to (i) empower women by forming or strengthening women’s associations, (ii) help women improve the

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2099

productivity and marketing of the crops and animals under their responsibility, and (iii) foster women’s access to information. Obviously, these recommendations also apply to Asia and Latin America where gender inequality prevails. Undoubtedly, much remains to be done to foster gender equality in development initiatives. So, how can we put all of these biophysical and socioeconomic aspects together?

Integrative and Cross-Disciplinary Work: Stitching it all Together at the Landscape Level In order to address the challenges of food insecurity, poverty and inequality, climate change, ecosystem degradation, and biodiversity loss at the same time, it is critical to put all our tools for sustainable use of natural resources to work to create “multifunctional agriculture” subject to “integrated landscape management” (UN SDSN 2012; Milder et al. 2012). For the past 50 years or so, despite local and temporary relief generated by punctual efforts (e.g., green revolution and monocultures in agriculture), individual uncoordinated efforts in agronomy, livestock farming, forestry, or ecological engineering were not able to sustain environmental services, which are emergent properties of landscapes (Pope Francis 2015; Kimbrell 2002). Consequently, the development paradigm has to change and integrate the advances in all these fields. Agroforestry and agroecological practices have much to offer in the creation of such “climate-smart landscapes” (Gliessman 2015; Leakey 2012; Nair and Garrity 2012). Our focus must widen to encompass all actions, despite their immense complexity and our resulting lack of mechanistic understanding. When trees, crops, animals, and their respective microbial symbionts interact with the atmosphere and the pedosphere, complicated by nonlinear relationships over different scales of time and space, predictions based on linear dose-response curves for the application of individual nutrients are often inappropriate, no matter how hard we like to cling to them. Add extreme climatic events and increasing population pressure to upset the systems, plus the global spread of pests and diseases to other continents (cf. coffee leaf rust, coffee berry borer, or the existence-threatening diseases of bananas and cacao), and we can easily see how “business as usual” with its reliance on technological quick fixes in the form of pesticides or genetically modified crops not only has contributed greatly to the dimensions of our current problems (Kimbrell 2002) but will also likely impede effective solutions to overcome them. Clearly, a new paradigm is needed to substitute the ineffective reductionist approaches of linking individual factors often with (mostly inappropriate) linear relations for the sake of modeling. Increasingly, farmers and land managers are reaching out across traditional sectorial boundaries to forge partnerships with conservation and development organizations, researchers, local governments, businesses, and others to address these interconnected problems. According to the UN SDSN (2012), more than 200 such initiatives have already been documented in Latin America, Africa, and Asia. Our future efforts should facilitate and build on such multi-actor and multidimensional initiatives as illustrated by the following examples from UN SDSN (2012): in Lari-Kijabe in Kenya, smallholder farmer organizations are

2100

R.G. Muschler

partnering with local governments, banks, and conservation groups to expand agricultural markets and protect high conservation value forests and watersheds. In the Maasai Steppe land of Tanzania, commercial avocado producers, pastoralists, and conservation organizations are partnering to raise incomes and food security, while protecting wildlife. In Tigray, Ethiopia, restoration of highly degraded watersheds by community-government-NGO partnerships has enabled irrigation and water access, increased food production, and greatly reduced the need for food aid during droughts. From Latin America, where agroforestry is increasingly important (Somarriba et al. 2012), another highly successful example is CATIE’s “Mesoamerican Agro-environmental Program” (MAP), which has, over the past 10 years, facilitated positive changes in many communities throughout Central America. This multi-actor platform (http://map.catie.ac.cr/web_en/) has provided central tools for promoting effective practices for improving agroforestry systems with cacao, for promoting the agroecological production of vegetables in homegardens and integrated production systems, and for strengthening local governments in their quest for combining production and income generation with the effective protection of natural resources. The creation of climate-smart territories for food security and, where possible, food sovereignty requires the judicious integration of agricultural, livestock, agroforest, and forest ecosystems in a way that maximizes productivity on the most appropriate lands for production while guaranteeing the essential ecosystem services by protecting critical areas. As indicated in Fig. 30 (and recalling Fig. 2 from the beginning of this chapter), efforts should be strengthened to develop tools which help us design landscapes as a mosaic of land use patches whose species composition, management, dimensions, and spatial distribution foster the effective functional integration of production and conservation objectives. Obviously, the conservation must include cultivated and wild biodiversity (from micro- to macroscales), as well as soil fertility and water resources, the three key elements for sustainable production (Frison et al. 2011; Jackson et al. 2007). In 2012, a global coalition of more than 50 agriculture, environment, and development organizations came together to implement the “Landscapes for People, Food and Nature Initiative” (www.landscapes.ecoagriculture.org). The target is to combine sustainable development in food production, ecosystem health, and human well-being. The top priority is to strengthen the capacity of existing landscape initiatives and mobilize cross-site learning, coordinated investment, and documentation. To accelerate the scaling up of integrated landscape approaches, the initiative is assisting countries to put in place supportive policy frameworks, encouraging businesses to pursue sustainable sourcing through landscape partnerships, expanding financing for integrated landscape investments, and promoting science and knowledge systems for landscape solutions. Undoubtedly, this is an illustrative example of what is needed. Another example is the evolution and success of the Cuban programs on agroforestry, reforestation, and knowledge-intensive urban agriculture over the

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2101

Fig. 30 At the landscape level, it is essential to put the different agroforestry practices to work in a “climate-smart” mosaic. In this example from the Central Highlands of the Dominican Republic, we see the integration of different agroforestry practices that are (or should be) incorporated in order to protect the natural resources (Photo credit: R. Muschler)

past 25 years. Productivities exceeding 10 kg of organic vegetable production per m2 (INIFAT 2011) illustrate the potential from pooling effective agroecological practices, even under a demanding climate and on lateritic soils. This is remarkable and the accumulated know-how should be applied to many similar situations (Niggli et al. 2007). Undoubtedly, there is now a wealth of effective agroecological practices which can help reduce the yield gap of 5 to 34% between high-chemical input and organic production systems even further (Seufert et al. 2012). Finally, as illustrated by the examples in sections “Principal Agroforestry Practices” (Fig. 11) and “Roles and Potential of Agroforestry for Sustainable Land and Landscape Management,” decided community efforts to build artificial lakes and establish agroforestry systems around them can bring about the effective rehabilitation even of degraded landscapes subject to high population pressure like in Haiti. As these examples illustrate, when the tools of agroforestry and agroecological management are used to their full potential to contribute to sustaining multifunctional landscapes, the title of the recent review Agroforestry – The Future of Global Landuse (Nair and Garrity 2012) may be well justified. In fact, the creation of climate-smart landscapes with agroforestry at its core offers substantial benefits over the model of continued agricultural intensification and land sparing (Perfecto and Vandermeer 2010). While agroforestry is not a magic wand for solving all human and environmental challenges, it clearly provides powerful tools to address many of them.

2102

R.G. Muschler

References Abukutsa-Onyango MO (2014) Strategic repositioning of African indigenous vegetables and fruits with nutrition, economic and climate change resilience potential. In: Gurib-Fakim A (ed) Novel plant bioresources: applications in food, medicine and cosmetics. Wiley, London, pp 361–370, Chapter 25 Akinnifesi FK, Ajayi OC, Sileshi G, Chirwa PW, Chianu J (2010) Fertilizer trees for sustainable food security in the maize-based productions systems of East and Southern Africa. A review. Agron Sustain Dev. doi:10.1051/agro/2009058 Altieri MA (1999) Agroecologı´a. Bases Cientificas para la Agricultura Sustentable. Editorial NORDAN-Comunidad, Montevideo, 325 p Altieri M (2002) Agroecology: the science of natural resource management for poor farmers in marginal environments. Agric Ecosyst Environ 93:1–24 Altieri M, Nicholls C (2003) Soil fertility management and insect pests: harmonizing soil and plant health in agroecosystems. Soil Tillage Res 72:203–211 Altieri M, Nicholls C, Fritz MA (2005) Manage insects on your farm. A guide to ecological strategies. Sustainable Agriculture Network, Beltsville, 119 p Amaranthus M, Simpson L, Malajczuk N (2012) Inoculate with mycorrhizae – it’s as easy as A-B-seeds. ACRES (USA) 42(1) Amundson R, Berhe AA, Hopmans JW, Olson C, Sztein AE, Sparks DL (2015) Soil and human security in the 21st century. Science 348(6235) doi:10.1126/science.1261071:1–6 Antoniou M, Robinson C, Fagan J (2012) GMO myths and truths – an evidence-based examination of the claims made for the safety and efficacy of genetically modified crops. Earth Open Source, London, 331 p Aranda E, Barois I, Arellano P, Irisso´n S, Salazar T, Rodrı´guez J, Patron JC (1999) Vermicomposting in the tropics. In: Lavelle P, Brussaard L, Hendrix P (eds) Earthworm management in tropical agroecosystems. CAB International, Wallingford, pp 253–287, 300 p Avelino J, Romero-Gurdián A, Cruz-Cuellar HF, DeClerck FAJ (2012) Landscape context and scale differentially impact coffee leaf rust, coffee berry borer, and coffee root-knot nematodes. Ecol Appl 22:584–596 Barantal S, Schimann H, Fromin N, Ha¨ttenschwiler S (2014) C, N and P fertilization in an Amazonian rainforest supports stoichiometric dissimilarity as a driver of litter diversity effects on decomposition. Proc R Soc B 281:20141682, http://dx.doi.org/10.1098/rspb.2014.1682 Bardgett RD, van der Putten WH (2014) Belowground biodiversity and ecosystem functioning. Nature 515:505–511. doi:10.1038/nature13855, pmid: 25428498 Barradas VL, Fanjul L (1986) Microclimatic characterization of shaded and open-grown coffee (Coffea arabica L.) plantations in Mexico. Agric For Meteorol 38:101–112 Beck M (1997) The secret life of compost: a guide to static-pile composting – lawn, garden, feedlot or farm. Acres USA :170 Beer J (1987) Advantages, disadvantages and desirable characteristics of shade trees for coffee, cacao and tea. Agroforest Syst 5:3–13 Beer J (1988) Litter production and nutrient cycling in coffee (Coffea arabica) or cacao (Theobroma cacao) plantations with shade trees. Agroforest Syst 7:103–114 Beer JW (ed) (2000) Highlights of CATIE’s agroforestry research in Latin America during the 1990s. Special issue. Agrofor Syst 51:75–175 Beer J, Fassbender H, Heuveldop J (eds) (1987) Advances in agroforestry research: proceedings of a seminar. CATIE, Turrialba, 379 p. (also published in Spanish) Beer JW, Muschler RG, Somarriba E, Kass D (1998) Shade management in coffee and cacao plantations – a review. Agroforest Syst 38:139–164 Benavides JE (ed) (1994) Arboles y Arbustos Forrajeros en Ame´rica Central, vol 236, Serie Te´cnica, Informe Te´cnico. CATIE, Turrialba, 2 Volumes. 721 p

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2103

Benavides JE, Arias R (eds) (1995) Sistemas Tradicionales y Agroforestales de Produccion Caprina en Ame´rica Central y Repu´blica Dominicana, vol 269, Serie Te´cnica, Informe Te´cnico. CATIE, Turrialba, 266 p Benyus JM (2002) How will we feed ourselves? Farming to fit the land: growing food like a prairie. In: Benyus JM (ed) Biomimicry. Innovation inspired by nature. Perennial Publ, New York, pp 11–58, 308 p Benzing A (2001) Agricultura Orgánica. Fundamentos para la Regio´n Andina. Neckar Verlag, Villingen-Schwenningen, 682 pp Boonkird SA, Fernandes ECM, Nair PKR (1984) Forest villages: an agroforestry approach to rehabilitating forest land degraded by shifting cultivation in Thailand. Agroforest Syst 2:87–102 BOSTID (1989) Lost crops of the Incas. Little-known plants of the Andes with promise for worldwide cultivation. Board on science and technology for international development. National Research Council, National Academic Press, Washington, DC, 415 p BOSTID (1991) Microlivestock. Little-know small animals with a promising economic future. Board on science and technology for international development. National Research Council, National Academic Press, Washington, DC, 449 p BOSTID (1996a) Lost crops of Africa. Volume I (grains) and volume II (cultivated fruits). Board on science and technology for international development. National Research Council, National Academic Press, Washington, DC BOSTID (1996b) Foods of the future: volume I (promising tropical fruits) board on science and technology for international development. National Research Council, National Academic Press, Washington, DC Brandle J, Hintz D, Sturrock J (eds) (1988) Windbreak technology. Elsevier Science, Amsterdam, 598 pp Brittain C, Williams N, Kremen C, Klein A-M (2013) Synergistic effects of non-Apis bees and honeybees for pollination services. Proc R Soc B 280:20122767, http://dx.doi.org/10.1098/ rspb.2012.2767 Brown GG, Pashanasi B, Villenave C, Patro´n JC, Senapati BK, Giri S, Barois I, Lavelle P, Blanchart E, Blakemore RJ, Spain AV, Boyer J (1999) In: Lavelle P, Brussaard L, Hendrix P (eds) Earthworm management in tropical agroecosystems. CAB International, Wallingford, pp 87–137, 300 p Budelman A (1988) The decomposition of the leaf mulches of Leucaena leucocephala, Gliricidia sepium and Flemingia macrophylla under humid tropical conditions. Agroforest Syst 7:33–45, 47–62 Budelman A (1989) Nutrient composition of the leaf biomass of three selected woody leguminous species. Agroforest Syst 8:39–51 Budowski G (1987) Living fences in tropical America, a widespread agroforestry practice. In: Gholz HL (ed) Agroforestry: realities, possibilities and potentials. Martinus Nijhoff, Dordrecht, pp 169–178 Buresh RJ, Cooper PJM (1999) The science and practice of improved fallows. Agroforest Syst 47:13–58 Cadisch G, Giller KE (1997) Driven by nature; plant litter quality and decomposition. Oxford University Press, Oxford, UK, 409 p Callo-Concha D, Krishnamurthy L, Alegre J (2002) Secuestro de carbono por sistemas agroforestales Amazo´nicos. Rev Chapingo: Cienc Forestales y Medio Ambiente 8:101–106 Camero A, Ibrahim M, Kass M (2001) Improving rumen fermentation and milk production with legume-tree fodder in the tropics. Agroforest Syst 51:157–166 Cannell MGR (1983) Plant management in agroforestry: manipulation of trees, population densities and mixtures of trees and herbaceous crops. In: Huxley PA (ed) Plant research and agroforestry. ICRAF, Nairobi, pp 455–486 Caramori PH, Androcioli Filho A, Leal AC (1996) Coffee shade with Mimosa scabrella Benth. for frost protection in southern Brazil. Agroforest Syst 33:205–214

2104

R.G. Muschler

Cardoso IM, Kuyper TW (2006) Mycorrhizas and tropical soil fertility. Agric Ecosyst Environ 116:72–84 Castan˜eda P, Castan˜eda O (2000) El Cafe´ Ecologico. Algunas Recomendaciones para su Cultivo, Procesamiento y Comercializacion. Vecinos Mundiales, Guatemala, 230 p Castello´n JU, Muschler RG, Jimenez F (2000) Abonos orgánicos: efecto de sombra y altitud en almácigos de cafe´. Agroforesterı´a en las Ame´ricas (CATIE) 7:30–33 CATIE (1999) Agroforesterı´a en el CATIE, vol 27, Bibliografı´a anotada. CATIE, Turrialba, 423 p CATIE (2001) Agroforesterı´a en el CATIE, vol 28, Suplemento Bibliográfico. CATIE, Turrialba, 171 p Cayuela ML, Monedero M, Roig A, Hanley K, Enders A, Lehmann J (2013) Biochar and denitrification in soils: when, how much and why does biochar reduce N2O emissions? Scientific Reports 3, 1732 Chizmar-Fernández C, Coronado-González I, Mejı´a-Ordo´n˜ez T, Raymond-House P, RuizValladares I, Menjı´var-Cruz JE, Lara LR, Cere´n-Lo´pez JG, Quesada-Hernández A, LoboCabezas SL, Chang-Vargas G, Correa-Arroyo MD (2009) Plantas Comestibles de Centroame´rica. Instituto Nacional de Biodiversidad (INBIO), Santo Domingo de Heredia, Costa Rica, 360 p Christiansen JA (2004) Cafe´ Organico con Diversificacion. Ideas Litograficas, Tegucigalpa, 345 p Cifuentes-Jara M (2008) Aboveground biomass and ecosystem carbon pools in tropical secondary forests growing in six life zones of Costa Rica. PhD thesis. Oregon State University, 195 p Cleugh H, Prinsley R, Bird R, Brooks S, Carberry P, Crawford M, Jackson T, Meinke H, Mylius S, Nuberg I, Sudmeyer R, Wright A (2002) The Australian national windbreaks program: overview and summary of results. Aust J Exp Agric 42:649–664 Coe R, Huwe B, Schroth G (2003) Designing experiments and analyzing data. In: Schroth G, Sinclair FL (eds) Trees, crops and soil fertility. Concepts and research methods. CABI Publishing, Wallingtford, pp 39–76, 437 p Colburn T, Dumanoski D, Myers JP (1997) Our stolen future: are we threatening our fertility, intelligence, and survival? A scientific detective story. Plume Publishers, New York, 336 p Conway GR, Pretty JN (1991) Unwelcome harvest. Agriculture and pollution. Earthscan Publications, UK, 645 p Cordero J, Boshier DH (eds) (2003) Árboles de Centroame´rica. Un Manual para Extensionistas. Oxford Forestry Institute – CATIE, Costa Rica, 1080 p Corlett JE, Ong CK, Black CR (1989) Modification of microclimate in intercropping and alleycropping systems. In: Reifsnyder WS, Darnhofer TO (eds) Meteorology and agroforestry. ICRAF/WMO/UNEP/GTZ, Nairobi, pp 419–430 Crews TE, Peoples MB (2005) Can the synchrony of nitrogen supply and crop demand be improved in legume and fertilizer-based agroecosystems? A review. Nutr Cycl Agroecosyst 72:101–120 Cunningham SJ (2000) Great garden companions: a companion-planting system for a beautiful. Chemical-free vegetable garden. Rodale Press, Emmaus, 288 p Current D, Lutz E, Scherr S (eds) (1995) Costs, benefits, and farmer adoption of agroforestry: project experiences in central America and the Caribbean, vol 14, World Bank environment paper. World Bank, Washington, DC Dagang ABK, Nair PKR (2003) Silvopastoral research and adoption in Central America: recent findings and recommendations for future directions. Agroforest Syst 59:149–155 Daghela Bisseleua HB, Fotio D, Yede, Missoup AD, Vidal S (2013) Shade tree diversity, cocoa pest damage, yield compensating inputs and farmers’ net returns in West Africa. PLoS One 8(3):e56115. doi:10.1371/journal.pone.0056115 Daily G (ed) (1997) Nature’s services. Societal dependence on natural ecosystems. Island Press, Washington, DC, 392 p DeClerck FAJ, Negreros-Castillo P (2000) Plant species of traditional Mayan homegardens of Mexico as analogs for multistrata agroforests. Agroforest Syst 48:303–317

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2105

Derpsch R, Friedrich T, Kassam A, Li HW (2010) Current status of adoption of no-till farming in the world and some of its main benefits. Int J Agric Biol Eng 3:1–25 Diamond J (2011) Collapse. How societies choose to fail or succeed. Penguin, New York, 608 p Dixon RK (1995) Agroforestry systems: sources or sinks of greenhouse gases? Agroforest Syst 31:99–116 Dossa EL, Fernandes EC, Reid WS, Ezui K (2008) Above- and belowground biomass, nutrient and carbon stocks contrasting an open-grown and a shaded coffee plantation. Agroforest Syst 72:103–115 Douglas I (2009) Climate change, flooding and food security in south Asia. Food Secur 1:127–136 Dronen S (1988) Layout and design criteria for livestock windbreaks. Agric Ecosyst Environ 22 (23):231–240 Ebert A (2014) Potential of underutilized traditional vegetables and legume crops to contribute to food and nutritional security, income and more sustainable production systems. Sustainability 6:319–335. doi:10.3390/su6010319 Ebert A, Astorga C, Ebert I, Mora A, Uman˜a C (2007) Securing our future. CATIE’s germplasm collections. CATIE, Turrialba, Costa Rica. Boletı´n Te´cnico No 26. 204 pp Elevitch CR (ed) (2011) Specialty crops for pacific islands. Permanent Agriculture Resources, Hawaii, 558 p Elevitch CR, Wilkinson K (1999) A guide to orchard alley cropping for fertility, mulch and soil conservation. AgroForester, Hawaii, 10 pp. Available at http://www.agroforestry.net/images/ pdfs/oachbk.pdf Espina D, Ordetx G (1983) Flora Apicola Tropical. Editorial Tecnologica de Costa Rica, 406 p Evans TA, Dawes TZ, Ward PR, Lo N (2011) Ants and termites increase crop yield in a dry climate. Nat Commun 2:262. doi:10.1038/ncomms1257 Ewel JJ (1986) Designing agricultural ecosystems for the humid tropics. Annu Rev Ecol Syst 17:245–271 Ewel JJ (1999) Natural systems as models for the design of sustainable systems of land use. Agroforest Syst 45:1–21 Fallon S (1999) Nourishing traditions: the cookbook that challenges politically correct nutrition and the diet dictocrats. Newtrends Publishing, Indiana, 688 p FAO (2010) An international consultation on integrated crop-livestock systems for development, the way forward for sustainable production intensification, vol 13, Integrated crop management. FAO, Rome FAO (2011) Los bosques para una mejor nutricio´n y seguridad alimentaria. Roma, 12 p FAO (2013) Climate-smart agriculture sourcebook. Food and Agriculture Organization, Rome, 570 p FAO (2015) Genetic resources and biodiversity for food and agriculture. A treasure for the future. Infographic at http://www.fao.org/assets/infographics/FAO-Infographic-CGRFA30-en.pdf Fassbender HW (1993) Modelos Edafolo´gicos de Sistemas Agroforestales, 2nd edn. CATIE, Costa Rica. Serie de Materiales de Ensenanza No 29. 491 p Fassbender HW, Beer J, Heuveldop J, Imbach A, Enriquez G, Bonnemann A (1991) Ten years balances of organic matter and nutrients in agroforestry systems at CATIE, Costa Rica. For Ecol Manage 45:173–183 Fernandes ECM, Nair PKR (1986) An evaluation of the structure and function of tropical homegardens. Agric Syst 21:279–310 Figueroa R, Fischerworring B, Rosskamp R (1998) Guia para la Caficultura Ecolo´gica: Cafe´ Orgánico. Novella Publigraf, Lima, 176 p Francesconi W, Montagnini F, Ibrahim M (2011) Living fences as linear extensions of forest remnants: a strategy for restoration of connectivity in agricultural landscapes. In: Montagnini F, Finney C (eds) Restoring degraded landscapes with native species in Latin America. Nova Science, New York, 244 p Frison EA, Cherfas J, Hodgkin T (2011) Agricultural biodiversity is essential for a sustainable improvement in food and nutrition security. Sustainability 3:238–253

2106

R.G. Muschler

Funes-Monzote F, Freyre Roach EF (eds) (2009) Transge´nicos ¿Que´ se gana? ¿Que´ se pierde? Textos para un Debate en Cuba. OXFAM, RAP-AL. Publicaciones Acuario, Centro Fe´lix Varela, Habana, Cuba, 321 p Gantheret L (2010) Capitaines de l’esperance. Film documentary about the successful establishment of artificial lakes in the Central Highlands of Haiti. Argus Productions, Port-au-Prince, Haiti. 52 min Garen EJ, Saltonstall K, Ashton MS, Slusser JL, Mathias S, Hall JS (2010) The tree planting and protecting culture of cattle ranchers and small-scale agriculturalists in rural Panama: opportunities for reforestation and land restoration. For Ecol Manag. doi:10.1016/j. foreco.2010.10.011 Geilfus F (1994) El Arbol al Servicio del Agricultor. Manual de Agroforesterı´a para el Desarrollo Rural. ENDA CARIBE – CATIE, Turrialba Geilfus F (2002) 80 herramientas para el desarrollo participativo: diagno´stico, planificacio´n, monitoreo, evaluacio´n. IICA, San Jose´, 217 p Gliessman SR (2015) Agroecology. The ecology of sustainable food systems, 3rd edn. CRC Press, Boca Raton, 371 p Haggar JP, Warren GP, Beer JW, Kass D (1991) Phosphorous availability under alley cropping and mulched and unmulched sole cropping systems in Costa Rica. Plant and Soil 137:275–283 Haggar JP, Beer JW, Tanner EVJ, Rippin M (1993) Nitrogen dynamics of tropical agroforestry and annual cropping systems. Soil Biol Biochem 25:1363–1378 Hardner J, Rice R (2002) Replanteamiento del mercado ecolo´gico. Investigacion y Ciencia 76–83 Hartemink AE (2005) Nutrient stocks, nutrient cycling, and soil changes in cocoa ecosystems: a review. Adv Agron 86:227–253 Harvey CA, Tucker NIJ, Estrada A (2004) Live fences, isolated trees, and windbreaks: tools for conserving biodiversity in fragmented tropical landscapes. In: Schroth G, da Fonseca GAB, Harvey CA, Gascon C, Vasconcelos HL, Izac A-MN (eds) Agroforestry and biodiversity conservation in tropical landscapes. Island Press, Washington, pp 261–289, 523 p Harvey C, Medina A, Sánchez D, Vı´lchez S, Hernández B, Sáenz JC, Maes JM, Casanoves F, Sinclair F (2006) Patterns of animal diversity in different forms of tree cover in agricultural landscapes. Ecol Appl 16:19–86 Hassan J (2011) Introduction of partial carbon footprint on double-purpose systems in Azuero province, Panama. M.Sc. thesis. CATIE, Turrialba, Costa Rica Hauggaard-Nielsen H, Jensen ES (2005) Facilitative root interactions in intercrops. Plant Soil 274:237–250 Heller NE, Zavaleta ES (2008) Biodiversity management in the face of climate change: a review of 22 years of recommendations. Biol Conserv 124:14–32 Hergoualc’h K, Blanchart E, Skiba U, He´nault C, Harmand J-M (2012) Changes in carbon stock and greenhouse gas balance in a coffee (Coffea arabica) monoculture versus an agroforestry system with Inga densiflora in Costa Rica. Agric Ecosyst Environ 148:102–110 Hoehn P, Steffan-Dewenter I, Tscharntke T (2010) Relative contribution of agroforestry, rainforest and openland to local and regional bee diversity. Biodivers Conserv 19:2189–2200 Houghton JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, Xiaosu D (eds) (2001) Climate change 2001: the scientific basis, report of working group I of the intergovernmental panel on climate change. Cambridge University Press, UK, 944 p Huxley P (1999) Tropical agroforestry. Blackwell, Oxford, 371 p IASS (2015) Soil Atlas. Facts and figures about earth, land, and fields. Institute for Advanced Sustainability Studies (IASS), Potsdam, 68 pp. www.iass-potsdam.de/en/publications/soilatlas Ibrahim M, Chaco´n M, Cuartas C, Naranjo J, Ponce G, Vega P, Casasola F, Rojas J (2007) Almacenamiento de carbono en el suelo y la biomasa ae´rea en sistemas de uso de la tierra en paisajes ganaderos de Colombia, Costa Rica y Nicaragua. Agroforest Am (CATIE, Costa Rica) 45:27–36

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2107

Ibrahim M, Casasola F, Villanueva C, Murgueitio E, Ramı´rez E, Sáenz J, Sepu´lveda C (2010) Payment for environmental services as a tool to encourage the adoption of of silvo-pastoral systems and restoration of agricultural landscapes dominated by cattle in Latin America. In: Montagnini F, Finney C (eds) Restoring degraded landscapes with native species in Latin America. Nova Science, New York, 244 p INIFAT (2011) Manual Te´cnico para Organopo´nicos, Huertos Intensivos y Organoponı´a Semiprotegida. 7ma edicio´n. Instituto de Investigaciones Fundamentales en Agricultura Tropical (INIFAT), Habana, Cuba, 210 p Jackson LE, Pascual U, Hodgkin T (2007) Utilizing and conserving agrobiodiversity in agricultural landscapes. Agric Ecosyst Environ 121:196–210 Jackson B, Pagella T, Sinclair F, Orellana B, Henshaw A, Reynolds B, Mcintyre N, Wheater H, Eycott A (2013) Polyscape: a GIS mapping framework providing efficient and spatially explicit landscape-scale valuation of multiple ecosystem services. Landsc Urban Plan 112:74–88 Jaenicke H, Ho¨schle-Zeledon I (eds) (2006) Strategic framework for underutilized plant species research and development with special reference to Asia and the pacific, and to Sub-Saharan Africa. International Center for Underutilized Crops, Colombo, Sri Lanka and Global Facilitation Unit for Underutilized Species, Rome, 33 p Jalonen R, Nygren P, Sierra J (2009) Transfer of nitrogen from a tropical legume tree to an associated fodder grass via root exudation and common mycelial networks. Plant Cell Environ 32:1366–1376 Jama BA, Palm CA, Buresh NJ, Niang AI, Gachengo C, Nziguheba G, Amadalo B (2000) Tithonia diversifolia as a green manure for soil fertility improvement in western Kenya: a review. Agroforest Syst 49:201–221 Jeavons J (2014) How to grow more vegetables (and fruits, nuts, berries, grains, and other crops) than you ever thought possible on less land than you can imagine, 8th edn. Ten Speed Press, Willits, Ecology Action, 256 p Jensen M (1993a) Soil conditions, vegetation structure and biomass of a Javanese homegarden. Agroforest Syst 24:171–186 Jensen M (1993b) Productivity and nutrient cycling in a Javanese homegarden. Agroforest Syst 24:187–201 Jha S, Bacon CM, Philpott SM, La¨derach P, Rice RA (2014) Shade coffee: update on a disappearing refuge for biodiversity. Bioscience 64:416–428 Jime´nez JJ, Thomas RJ (eds) (2001) Nature’s plow: soil macroinvertebrate communities in the neotropical savannas of Colombia. CIAT, Cali, 389 p Jime´nez F, Muschler RG, Ko¨psell E (eds) (2001) Funciones y Aplicaciones de Sistemas Agroforestales. Modulo de Ensen˜anza No 6. Proyecto Agroforestal CATIE-GTZ, 187 p Jordan CF (1985) Nutrient cycling in tropical forest ecosystems. Principles and their application in management and conservation. Wiley, New York, 190 p Jordan CF, Gajaseni J, Watanabe H (eds) (1992) Taungya: forest plantations with agriculture in Southeast Asia. CAB International, Wallingford Jose S (2009) Agroforestry for ecosystem services and environmental benefits: an overview. Agroforest Syst 76:1–10 Jose S, Gordon AM (eds) (2008) Towards agroforestry design – an ecological approach. Springer, Dordrecht, 315 p Kahane R, Hodgkin T, Jaenicke H, Hoogendoorn C, Hermann M, Keatinge D, d’Arros Hughes J, Padulosi S, Looney N (2013) Agrobiodiversity for food security, health and income. Agron Sustain Dev 63(4):671–693 Kang BT, Duguma B (1985) Nitrogen movement in alley cropping systems. In: Kang BT, van den Heide J (eds) Nitrogen in farming systems in the humid and subhumid tropics. Institute of Soil Fertility, Haren, pp 269–284

2108

R.G. Muschler

Kang BT, Wilson GF (1987) The development of alley cropping as a promising agroforestry technology. In: Steppler HA, Nair PKR (eds) Agroforestry: a decade of development. ICRAF, Nairobi, pp 227–243 Kang BT, van der Kruijs ACBM, Cooper DC (1989) Alley cropping for food production. In: Kang BT, Reynolds L (eds) Alley farming in the humid and sub-humid tropics. International Development Research Center, Ottawa, pp 16–26 Kang BT, Reynolds L, Atta-Krah AN (1990) Alley farming. Adv Agron 43:315–359 Kareiva P, Tallis H, Ricketts TH, Daily GC, Polasky S (eds) (2011) Natural capital: theory and practice of mapping ecosystem services. Oxford Univ Press, Oxford, 365 p Kass D (1987) Alley cropping of annual food crops with woody legumes in Costa Rica. In: Beer JW, Fassbender HW, Heuveldop J (eds) Advances in agroforestry research: proceedings of a seminar. CATIE, Costa Rica, pp 197–208 Kass D, Somarriba E (1999) Traditional fallows in Latin America. Agroforest Syst 47:13–36 Keatinge J, Waliyar F, Jamnadas RH, Moustafa A, Andrade M, Drechsel P, Hughes J, Kadirvel P, Luther K (2010) Relearning old lessons for the future of food – by bread alone no longer: diversifying diets with fruit and vegetables. Crop Sci 50:51–62 Khoury CK, Bjorkman AD, Dempewolf H, Ramirez-Villegas J, Guarino L, Jarvis A, Rieseberg LH, Struik PC (2014) Increasing homogeneity in global food supplies and the implications for food security. www.pnas.org/cgi/doi/10.1073/pnas.1313490111 Kimbrell A (ed) (2002) Fatal harvest. The tragedy of industrial agriculture. Island Press, Foundation for Deep Ecology, Washington, 384 p Kindt R, Ordonez J, Smith E, Orwa C, Harja D, Kehlenbeck K, Luedeling E, Munjuga M, Mwanzia L, Sinclair F, Jamnadass R (2013) ICRAF species switchboard. Version 1.0. World Agroforestry Centre, Nairobi. http://www.worldagroforestry.org/products/switchboard/index. php Kiptot E, Franzel S (2012) Gender and agroforestry in Africa: who benefits? The African perspective. In: Nair PKR, Garrity D (eds) Agroforestry – the future of global land use. Springer, Dordrecht, pp 463–496, 541 p Klein A-M, Cunningham SA, Bos M, Steffan-Dewenter I (2008) Advances in pollination ecology from tropical plantation crops. Ecology 89:935–943 Kremen C (2005) Managing ecosystem services: what do we need to know about their ecology? Ecol Lett 8:468–479 Kumar BM, Nair PKR (eds) (2006) Tropical homegardens: a time-tested example of sustainable agroforestry, vol 3, Advances in Agroforestry. Springer, Dordrecht Kumar BM, Nair PKR (eds) (2011) Carbon sequestration potential of agroforestry systems. Opportunities and challenges. Springer, Dordrecht, 310 p Kuyah S, Rosenstock TS (2015) Optimal measurement strategies for aboveground tree biomass in agricultural landscapes. Agroforest Syst 89:125–133 Lal R (2015) Sequestering carbon and increasing productivity by conservation agriculture. J Soil Water Conserv 70:55A–62A Landauer K, Brazil M (eds) (1990) Tropical home gardens. United Nations University Press, Tokyo Lavelle P, Brussaard L, Hendrix P (eds) (1999) Earthworm management in tropical agroecosystems. CAB International, Wallingford, 300 pp Leakey RRB (1996) Definition of agroforestry revisited. Agroforest Today 8:5–7 Leakey R (1999) Potential for novel food products from agroforestry trees: a review. Food Chem 66:1–14 Leakey RRB (2012) Multifunctional agriculture and opportunities for agroforestry: implications of IAASTD. In: Nair PKR, Garrity D (eds) Agroforestry – the future of global land use. Springer, Dordrecht, pp 203–214, 541 p Leakey RRB, Weber JC, Page T, Cornelius JP, Akinnifesi FK, Roshetko JM, Tchoundjeu Z, Jamnadass R (2012) Tree domestication in agroforestry: progress in the second decade

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2109

(2003–2012). In: Nair PKR, Garrity D (eds) Agroforestry – the future of global land use. Springer, Dordrecht, pp 145–173, 541 p Lehmann J, Joseph S (2009) Biochar for environmental management - science and technology. Earthscan, London; Sterling, VA. 404 p Leterme P, Buldgen A, Estrada F, London AM (1996) Mineral content of tropical fruits and unconventional foods of the Andes and the rain forest of Colombia. Food Chem 95 (2006): 644–652 Leu A (2014) The myths of safe pesticides. Acres USA. 142 p Leupolz W (2000) Manual de Crianza y Explotacio´n de Ovejas de Pelo en los Tro´picos. ECONICA, Managua, 300 p Lin BB (2011) Resilience in agriculture through crop diversification: adaptive management for environmental change. Bioscience 61:183–193 Lok R (ed) (1998a) Huertos caseros tradicionales de America Central: caracterı´sticas, beneficios e importancia desde un enfoque multidisciplinario. CATIE, Turrialba, 232 p Lok, R (1998b) Introduccio´n a los Huertos caseros tradicionales tropicales. Mo´dulo de Ensen˜anza Agroforestal No. 3. CATIE. Serie Materiales de Ensen˜anza No. 41. Proyecto Agroforestal CATIE-GTZ, Turrialba, Costa Rica. 157 p Lorenzi H, Bacher L, Lacerda M, Sartori S (2006) Brazilian fruits and cultivated exotics. Instituto Plantarum de Estudos da Flora Ltda, Nova Odessa, 672 p Love BE, Bork EW, Spaner D (2009) Tree seedling establishment in living fences: a low-cost agroforestry management practice for the tropics. Agroforest Syst 77:1–8 Lowenfels J, Lewis W (2010) Teaming with microbes. The organic gardener’s guide to the soil food Web. Timber Press, Portland, 220 p Lyngbaek AE, Muschler RG, Sinclair FL (2001) Productivity and profitability of multistrata organic versus conventional coffee farms in Costa Rica. Agroforest Syst 53:205–213 MacDicken KG, Vergara NT (1990) Agroforestry: classification and management. Wiley, New York, 382 p Mafongoya PL, Hove (2008) Tree foliage polyphenols and nitrogen use in crop–livestock systems of Southern Africa: strategies for increasing efficiency. In: Jose S, Gordon AM (eds) Towards agroforestry design – an ecological approach. Springer, Dordrecht, pp 207–227, 315 pp Mafongoya PL, Barak P, Reed JD (2000) Carbon, nitrogen and phosphorus mineralization from tree leaves and manure. Biol Fertil Soils 30:298–305 Magdoff F, Van Es H (2009) Building soils for better crops. Sustainable soil management, 3rd edn. SARE-USDA, Washington, DC, 294 p Makumba W, Janssen B, Oenema O, Akinnifesi FK, Mweta D, Kwesiga F (2006) The long-term effects of a gliricidia–maize intercropping system in Southern Malawi on gliricidia and maize yields, and soil properties. Agric Ecosyst Environ 116:85–92 Margulis L (1998) Symbiotic planet. A new look at evolution. Basic Books, Perseus Books Group, New York, 147 p Marten GG, Abdoellah OS (1988) Crop diversity and nutrition in West Java. Ecol Food Nutr 21:17–34 Martı´nez-Viera R, Dibut-Álvarez B (2012) Biofertilizantes bacterianos. Editorial Cienı´´ıficoTe´cnica, La Habana, 279 p Martius C, Ho¨fer H, Garcia MVB, Ro¨mbke J, Fo¨rster B, Hanagarth W (2004) Microclimate in agroforestry systems in central Amazonia: does canopy closure matter to soil organisms? Agroforest Syst 60:291–304 Matocha J, Schroth G, Hills T, Hole D (2012) Integrating climate change adaptation and mitigation through agroforestry and ecosystem conservation. In: Nair PKR, Garrity D (eds) Agroforestry – the future of global land use. Springer, Dordrecht, pp 105–126, 541 p May T, Rodriguez S (2012) Plantas de intere´s apı´cola en el paisaje: observaciones de campo y la percepcio´n de apicultores en Repu´blica Dominicana. Rev Geogr Am Central 48:133–162 McCann KS (2000) The diversity-stability debate. Nature 405:228–233

2110

R.G. Muschler

Medina-Solı´s JA (1990) Manual de Apicultura Tropical. Imprenta Grafico Litho Offset SA. San Jose´, Costa Rica, 225 p Mendez VE, Beer J, Faustino J, Otarola A (2000) Plantacion de Árboles en Lı´nea, vol 1, 2nd edn, Mo´dulo de Ensen˜anza Agroforestal. CATIE, Turrialba Me´ndez VE, Lok R, Somarriba E (2001) Interdisciplinary analysis of homegardens in Nicaragua: micro-zonation, plant use and socioeconomic importance. Agroforest Syst 51:85–96 Mercer DE, Miller RP (1998) Socioeconomic research in agroforestry: progress, prospects, priorities. Agroforest Syst 38:177–193 Milder JC, Buck LE, DeClerck F, Scherr SJ (2012) Landscape approaches to achieving food production, natural resource conservation, and the millennium development goals. In: Ingram JC, DeClerck F, Rumbaitis del Rio C (eds) Integrating ecology and poverty reduction, vol I, Ecological dimensions. Springer, New York, pp 77–108, 425 p Mittal SP, Singh P (1989) Intercropping field crops between rows of Leucaena leucocephala under rainfed conditions in northern India. Agroforest Syst 8:165–172 Mollison B (1996) Permaculture. A designer’s manual. Tagari Publications, Tyalgum, 576 p Montagnini F, Finney C (eds) (2011) Restoring degraded landscapes with native species in Latin America. Nova Science, New York, 244 p Montagnini F, Nair PKR (2004) Carbon sequestration: and underexploited environmental benefit of agroforestry systems. Agroforest Syst 61:281–295 Montgomery DR (2012) Dirt. The erosion of civilizations. University of California Press, Berkeley, 285 p Morton JF (1987) Fruits of warm climates. JFM Publisher, Miami, 505 p Msuya JM, Mamiro P, Weinberger K (2009) Iron, zinc and β-carotene nutrient potential of non-cultivated indigenous vegetables in Tanzania. ISHS Acta Horticult 806:217–222 Murgueitio E, Calle Z, Uribe F, Calle A, Solorio B (2011) Native trees and shrubs for the productive rehabilitation of tropical cattle ranching lands. For Ecol Manage 261:1654–1663 Murillo R, Kilian B, Castro R (2011) Leveraging and sustainability of PES. In: Rapidel B, DeClerck F, Le Coq J-F, Beer J (eds) Ecosystem services from agriculture and agroforestry. Measurement and payment. Earthscan, London, pp 267–287, 414 p Muschler RG (1993) Chapter 13. Component interactions. In: Nair PKR (ed) Introduction to agroforestry. Kluwer, Amsterdam, pp 243–258, 499 p Muschler RG (1998) Tree-crop compatibility in agroforestry: production and quality of coffee grown under managed Tree Shade in Costa Rica. Agroforestry Program, University of Florida, Ph.D. Dissertation. 219 p Muschler RG (2001a) Árboles en Cafetales. Mo´dulo de Ensen˜anza Agroforestal. CATIE, Costa Rica. Proyecto Agroforestal CATIE/GTZ. 137 p Muschler RG (2001b) Shade improves coffee quality in a sub-optimal coffee zone of Costa Rica. Agroforest Syst 51:131–139 Muschler RG (2004) Shade management and its effect on coffee growth and quality. In: Wintgens J-N (ed) Coffee: growing, processing, sustainable production. A guidebook for growers, processors, traders and researchers. Wiley-VCH, Weinheim, pp 391–418 Muschler RG, Bonnemann A (1997) Potentials and limitations of agroforestry for changing land-use in the tropics: experiences from Central America. For Ecol Manage 91:61–73 Muschler RG, Nair PKR, Mele´ndez L (1993) Crown development and biomass production of pollarded Erythrina berteroana, E.fusca and Gliricidia sepium in the humid tropical lowlands of Costa Rica. Agroforest Syst 24:123–143 Muschler RG, Ye´pez C, Rodrı´guez A, Peters W, Pohlan HAJ (2006) Manejo y valoracio´n de la biodiversidad en cafetales. In: Pohlan J, Soto L, Barrera J (eds) El Cafetal del Futuro. Realidades y Visiones. ECOSUR, Chiapas, Me´xico. Shaker, Aachen, pp 333–360, 462 p Mutua J, Muriuki J, Gachie P, Bourne M, Capis J (2014) Conservation agriculture with trees: principles and practice. A simplified guide for extension staff and farmers. World Agroforestry Center, Nairobi, 93 p

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2111

NABCI Canada (2012) The state of Canada’s birds. North American bird conservation initiative Canada. Environment Canada, Ottawa, 36 p Nair PKR (1980) Agroforestry species – a crop sheets manual. ICRAF, Nairobi, 336 p Nair PKR (1989) Agroforestry Systems in the Tropics. Amsterdam: Kluwer. 680 p Nair PKR (1993) An introduction to agroforestry. Kluwer, Amsterdam, 499 p Nair PKR (2001) Do tropical homegardens elude science, or is it the other way around? Agroforest Syst 53:239–245 Nair PKR (2012) Climate change mitigation and adaptation: a low hanging fruit of agroforestry. In: Nair PKR, Garrity D (eds) Agroforestry – the future of global land use. Springer, Dordrecht, pp 31–67, 541 p Nair PKR, Garrity D (eds) (2012) Agroforestry – the future of global land use. Springer, Dordrecht, 541 p Nair PKR, Kumar BM, Nair VD (2009) Agroforestry as a strategy for carbon sequestration. J Plant Nutr Soil Sci 172:10/23 Nair MA, Sreedharan C (1986) Agroforestry farming systems in the homesteads of Kerala, southern India. Agroforest Syst 4:339–363 Nardi JB (2007) Life in the soil. A guide for naturalists and gardeners. Univesity of Chicago Press, Chicago, 293 p Newton AC, Johnson SN, Gregory PJ (2011) Implications of climate change for diseases, crop yields and food security. Euphytica 179:3–18 Nicholls Estrada CI, Rios Osorio LA, Altieri MA (eds) (2013) Agroecologı´a y resiliencia socioecolo´gica: adaptandose al cambio climático. Sociedad Cientı´fica Latinoamericana de Agroecologı´a (SOCLA), Medellin, 207 p Nicholls CI, Altieri MA (2013) Plant biodiversity enhances bees and other insect pollinators in agroecosystems. A review. Agron Sustain Dev 33:257–274 Nichols E, Spector S, Louzada J, Larsen T, Amezquita S, Favila ME (2008) Ecological functions and ecosystem services provided by Scarabaeinae dung beetles. Biol Conserv 141:1461–1474 Nicolas M (2010) Danser avec la Vie. Editions Nestor, Port-au-Prince, Haiti. 259 p Niggli U, Earley J, Ogorzalek K (2007) Organic agriculture and environmental stability of the food supply, Issues paper. FAO, Rome Nyambo A, Nyomora A, Ruffo CK, Tengnas B (2005) Fruits and nuts species with potential for Tanzania. World Agroforestry Centre (ICRAF) – Eastern and Central African Programme, Nairobi, 160 p Nygren P, Ramirez C (1995) Production and turnover of N2 fixing nodules in relation to foliage development in periodically pruned Erythrina poeppigiana (Leguminosae) trees. For Ecol Manage 73:59–73 OISAT (2015) Online Information Service for Non-Chemical Pest Management in the Tropics. Pesticide action network Ong CK, Huxley P (eds) (1996) Tree-crop interactions – a physiological approach. CAB Int’l, Wallingford Ong CK, Leakey RRB (1999) Why tree-crop interactions in agroforestry appear at odds with treegrass interactions in tropical savannahs. Agroforest Syst 45:109–129 Ong CK, Black CR, Marshall FM, Corlett JE (1996) Principles of resource capture and utilization of light and water. In: Ong CK, Huxley P (eds) Tree-crop interactions – a physiological approach. CAB Int’l, Wallingford, pp 73–158 Orwa C, Mutua A, Kindt R, Jamnadass R, Anthony S (2009) Agroforestree database: a tree reference and selection guide version 4.0. World Agroforestry Centre, Kenya. http://www. worldagroforestry.org/resources/databases/agroforestree Padulosi S, Hodgkin T, Williams JT, Haq N (2002) Underutilized crops: trends, challenges and opportunities in the 21st century. In: Managing plant genetic diversity. IPGRI, Rome, pp 323–338 PAF-Ngo¨be-Bugle´ (2003) Manual Keba Sula. Me´todos Te´cnicos y Organizativos para el Manejo Sostenible de los Recursos Naturales Renovables en la Comarca Ngo¨be-Bugle´. Proyecto Agroforestal Ngo¨be-Bugle´ PAN-ANAM – GTZ, San Fe´lix, Panamá

2112

R.G. Muschler

Perfecto I, Vandermeer J (2010) The agroecological matrix as alternative to the landsparing/ agriculture intensification model. Proc Natl Acad Sci U S A 107:5786–5791 Perfecto I, Rice RA, Greenberg R, van der Voort ME (1996) Shade coffee: a disappearing refuge for biodiversity – shade coffee plantations can contain as much biodiversity as forest habitats. Bioscience 46:598–608 (n.d.) Permaculture design. Background information on permaculture principles and how to set up a permaculture system: http://www.tropicalpermaculture.com/permaculture-design.html Pezo D, Ibrahim M (2001) Sistemas Silvopastoriles. Mo´dulo de Ensen˜anza Agroforestal No. 2. CATIE. Serie Materiales de Ensen˜anza No. 44. Proyecto Agroforestal CATIE-GTZ. Turrialba, Costa Rica. 275 p Phillips-Mora W, Arciniegas-Leal A, Mata-Quiro´s A, Motamayor-Arias JC (2013) Catalogue of cacao clones selected by CATIE for commercial plantings. Turrialba, CATIE. Technical manual No 105. 68 p Philpott SM et al (2008) Biodiversity loss in Latin American coffee landscapes: review of the evidence on ants, birds, and trees. Conserv Biol 22:1093–1105 Phiri S, Rao IM, Barrios E, Singh BR (2003) Plant growth, mycorrhizal association, nutrient uptake and phosphorus dynamics in a volcanic-ash soil in Colombia, as affected by the establishment of Tithonia diversifolia. J Sustain Agric 21:41–59 Pilatic H (2012) Pesticides and honey bees – state of the science. Pesticide action network North America. 30 p Pinard F, Boffa JM, Rwakagara E (2014) Scattered shade trees improve low-input smallholder Arabica coffee productivity in the Northern Lake Kivu region of Rwanda. Agroforest Syst 88:707–718 Pope Francis (2015) Encyclical letter ‘Laudato si’ on care for our common home. The Holy See, Rome. 82 pp http://w2.vatican.va/content/francesco/en/encyclicals/documents/papafrancesco_20150524_enciclica-laudato-si.pdf. Practical action. Practical hands-on information for agriculture and animal husbandry and other dimensions of sustainable living Pyke CR, Andelman SJ (2007) Land use and land cover tools for climate adaptation. Clim Chang 80:239–251 Rao MR, Nair PK, Ong CK (1998) Biophysical interactions in tropical agroforestry systems. Agroforest Syst 38:3–50 Rapidel B, Roupsard O, Navarro M (eds) (2009) Modeling agroforestry systems. Workshop proceedings. CATIE, Costa Rica, 330 p Rapidel B, DeClerck F, Le Coq J-F, Beer J (eds) (2011) Ecosystem services from agriculture and agroforestry. Measurement and payment. Earthscan, London, 414 p Reid RS, Thornton PK, McCrabb GJ, Kruska RL, Atieno F, Jones PG (2004) Is it possible to mitigate greenhouse gas emissions in pastoral ecosystems of the tropics? Environ Dev Sustain 6:91–109 Reifsnyder WS, Darnhofer TO (eds) (1989) Meteorology and agroforestry ICRAF/WMO/UNEP/ GTZ, Nairobi, Kenya. 546 p Restrepo-Rivera J, Hensel J (n.d.) El ABC de la Agricultura Orgánica – Fosfitos y Panes de Piedra. Feriva S.A., Santiago de Cali, Colombia. 396 p Reynolds MP, Hays D, Chapman S (2010) Breeding for adaptation to heat and drought stress. In: Reynolds MP (ed) Climate change and crop production. CAB International, Wallingford, pp 71–91 Rice RA, Greenberg R (2000) Cacao cultivation and the conservation of biological diversity. Ambio 29:167–173 Richardson DM, Binggeli P, Schroth G (2004) Invasive agroforestry trees: problems and solutions. In: Schroth G, da Fonseca GAB, Harvey CA, Gascon C, Vasconcelos HL, Izac A-MN (eds) Agroforestry and biodiversity conservation in tropical landscapes. Island Press, Washington, pp 371–396, 523 p Rı´os JN, Andrade H, Ibrahim M, Jime´nez F, Sancho F, Ramı´rez E, Reyes B, Woo A (2007) Escorrentı´a superficial e infiltracio´n en sistemas silvopastoriles en el tro´pico subhu´medo de Costa Rica y Nicaragua. Agroforesterı´a en las Ame´ricas (CATIE, Costa Rica) 45: 66–71

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2113

Riotte L (1998) Carrots love tomatoes: secrets of companion planting for successful gardening. Storey Publishing, North Adams, 224 p Robalino J, Pfaff A, Villalobos L (2011) Assessing the impact of institutional design of payments for environmental services. The Costa Rican experience. In: Rapidel B, DeClerck F, Le Coq J-F, Beer J (eds) Ecosystem services from agriculture and agroforestry. Measurement and payment. Earthscan, London, pp 305–318, 414 p Robinson R (2007) Return to resistance, 3rd edn. Sharebooks Publisher, CA. http://www. sharebooks.ca/?filename=ReturnToResistance.pdf Ruf FO (2011) The myth of complex cocoa agroforests: the case of Ghana. Hum Ecol 39:373–388 Ryan D, Bright GA, Somarriba E (2009) Damage and yield change in cocoa crops due to harvesting of timber shade trees in Talamanca, Costa Rica. Agroforest Syst 77:97–106 Salazar E, Muschler RG, Sanchez V, Jimenez F (2000) Calidad de Coffea arabica bajo sombra de Erythrina poeppigiana a diferentes elevaciones en Costa Rica. Agroforesterı´a en las Ame´ricas (Costa Rica) 7:40–42 Samsel A, Seneff F (2013) Glyphosate’s suppression of Cytochrome P450 enzymes and amino acid biosynthesis by the gut microbiome: pathways to modern diseases. Entropy 15:1416–1463. doi:10.3390/e15041416 Sanchez PA (1995) Science in agroforestry. Agroforest Syst 30:5–55 Sánchez de Leo´n Y, de Melo E, Soto G, Johnson-Maynard J, Lugo-Pe´rez J (2006) Earthworm populations, microbial biomass and coffee production in different experimental agroforestry management systems in Costa Rica. Carib J Sci 42:397–409 Sánchez-Salmero´n, D, Muschler RG, Prins C, Solano W, Astorga C (2015) Identifying underutilized edible plant species based on their potential to improve human nutrition and resilience to climate change: a case study from El Salvador (in Spanish). Submitted to agroecosistemas Scherr SJ, Sthapit S (2009) Mitigating climate change through food and land use, vol 179, World watch report. WorldWatch Institute, Washington, DC Schlo¨nvoigt A (1998) Sistemas Taungya. Mo´dulo de Ensen˜anza Agroforestal No. 4. CATIE. Serie Materiales de Ensen˜anza No. 42. Proyecto Agroforestal CATIE-GTZ, Turrialba, Costa Rica. 117 p Schlo¨nvoigt A, Beer J (2001) Initial growth of pioneer timber tree species in a Taungya system in the humid lowlands of Costa Rica. Agroforest Syst 51:97–108 Schroth G (1995) Tree root characteristics as criteria for species selection and systems design in agroforestry. Agroforest Syst 30:125–143 Schroth G, Harvey C (2007) Biodiversity conservation in cocoa production landscapes: an overview. Biodivers Conserv 16:2237–2244 Schroth G, Sinclair FL (eds) (2003) Trees, crops and soil fertility. Concepts and research methods. CABI Publishing, Wallingford, 437 p Schroth G, da Fonseca GAB, Harvey CA, Gascon C, Vasconcelos HL, Izac A-MN (eds) (2004) Agroforestry and biodiversity conservation in tropical landscapes. Island Press, Washington, 523 p Sekercioglu CH (2012) Bird functional diversity and ecosystem services in tropical forests, agroforests and agricultural areas. J Ornithol 153:153–161 Seufert V, Ramankutty N, Foley JA (2012) Comparing the yields of organic and conventional agriculture. Nature 485:229–234 Shibata R, Yano K (2003) Phosphorus acquisition from non-labile sources in peanut and pigeonpea with mycorrhizal interaction. Appl Soil Ecol 24:133–141 Sieverding E (1991) Vesicular–arbuscular mycorrhiza management in tropical agroecosystems. Gesellschaft fur Technische Zusammenarbeit (GTZ), Eschborn Siles P, Harmand J-M, Vaast P (2010) Effects of Inga densiflora on the microclimate of coffee (Coffea arabica L.) and overall biomass under optimal growing conditions in Costa Rica. Agroforest Syst 78:269–286 Sinclair F (1999) A general classification of agroforestry practice. Agroforest Syst 46:161–180

2114

R.G. Muschler

Smith J (2010) Agroforestry: reconciling production with protection of the environment – a synopsis of research literature. Organic Research Center, Elm Farm, UK, 24 p Smith NJH, Williams JT, Plucknett DL, Talbot JP (1992) Tropical forests and their crops. Comstock Publishing Associates, Ithaca, 568 p Smith N, Vásquez R, Wust WH (2007) Frutos del Rı´o Amazonas. Sabores para la Conservacio´n. Amazon Conservation Association, Lima, 274 p Somarriba E (1992) Timber harvest, damage to crop plants and yield reduction in two Costa Rican coffee plantations with Cordia alliodora shade trees. Agroforest Syst 18:69–82 Somarriba E, Beer J (1987) Dimensions, volumes and growth of Cordia alliodora in agroforestry systems. For Ecol Manage 18:113–126 Somarriba E, Beer J (2011) Productivity of Theobroma cacao agroforestry systems with legume and timber shade tree species. Agroforest Syst 81:109–121 Somarriba E, Beer J, Muschler RG (2001) Research methods for multistrata agroforestry systems with coffee and cacao: recommendations from two decades of research at CATIE. Agroforest Syst 53:195–203 Somarriba E, Beer J, Alegre-Orihuela J, Andrade HJ, Cerda R, DeClerck F, Detlefsen G, Escalante M, Giraldo LA, Ibrahim M, Krishnamurthy L, Mena-Mosquera VE, Mora-Delgado JR, Orozco L, Scheelje M, Campos JJ (2012) Mainstreaming agroforestry in Latin America. In: Nair PKR, Garrity D (eds) Agroforestry – the future of global land use. Springer, Dordrecht, pp 429–454, 541 p Somarriba E, Cerda R, Orozco L, Deheuvels O, Cifuentes M, Dávila H, Espin T, Mavisoy H, Ávila G, Alvarado E, Poveda V, Astorga C, Say E (2013) Carbon stocks in agroforestry systems with cocoa (Theobroma cacao L.) in Central America. Agric Ecosyst Environ 173:46–57 Somarriba E, Suárez-Islas A, Calero-Borge W, Villota A, Castillo C, Vı´lchez S, Deheuvels O, Cerda R (2014) Cocoa–timber agroforestry systems: Theobroma cacao–Cordia alliodora in Central America. Agroforest Syst 86:1–19. doi:10.07/s10457-014-9692-7 Soto G, Le Coq J-F (2011) Certification process in the coffee value chain. Achievement and limits to foster provision of environmental services. In: Rapidel B, DeClerck F, Le Coq J-F, Beer J (eds) Ecosystem services from agriculture and agroforestry. Measurement and payment. Earthscan, London, pp 319–345, 414 p Stamets P (2005) Mycelium running. How mushrooms can help save the world. Ten Speed Press, Berkeley, 344 p Staver C, Guharay F, Monterroso D, Muschler RG (2001) Designing pest-suppressive multistrata perennial crop systems: shade-grown coffee in Central America. Agroforest Syst 53:151–170 Steppler HA, Nair PKR (eds) (1987) Agroforestry: a decade of development. ICRAF, Nairobi, Kenya Stigter K (ed) (2010) Applied agrometeorology. Springer, Berlin, 1101 p Stigter CJ, Darnhofer C, Herrera H (1989) Crop protection from very strong winds: recommendations from a Costa Rican agroforestry case study. In: Reifsnyder WS, Darnhofer TO (eds) Meteorology and agroforestry. ICRAF/WMO/UNEP/GTZ, Nairobi, pp 521–529, 546 p Stigter K, Ofori E, Kyei-Baffour N, Walker S (2011) Microclimate management and manipulation aspects of applied agroforestry. Overstory #240 Stolton S, Geier B, McNeely JA (eds) (2000) The relationship between nature conservation, biodiversity and organic agriculture. IFOAM-IUCN-AIAB, Tholey, 224 p Sullivan GM, Huke SM, Fox JM (eds) (1992) Financial and economic analyses of agroforestry systems. Proceedings of a workshop held in Honolulu, Hawaii, USA, July 1991. Nitrogen Fixing Tree Association, Paia. 312 p Szott LT, Palm CA, Sanchez PA (1991) Agroforestry on acid soils of the humid tropics. Adv Agron 45:275–301 The Montpellier Panel (2013) Sustainable intensification: a new paradigm for African agriculture. The Montpellier Panel, London

Agroforestry: Essential for Sustainable and Climate-Smart Land Use?

2115

Thornton T (2009) Can I grow a complete diet? Designing a tropical subsistence garden. 4 pp. Available at http://www.agroforestry.net/images/pdfs/Can_I_Grow_a_Complete_Diet.pdf Thornton PK, Herrero M (2010) Potential for reduced methane and carbon dioxide emissions from livestock and pasture management in the tropics. Proc Natl Acad Sci U S A 107 (46):19667–19672. doi:10.1073/pnas.0912890107 Thuerig B, Fließbach A, Berger N, Fuchs JG, Kraus N, Mahlberg N, Nietlispach B, Tamm L (2009) Re-establishment of suppressiveness to soil- and air-borne diseases by re-inoculation of soil microbial communities. Soil Biol Biochem 41:2153–2161 Thurston HD (1997) Slash/mulch systems. Westview, Boulder Trumper K, Bertzky M, Dickson B, van der Heijden G, Jenkins M, Manning P (2009) The natural fix? The role of ecosystems in climate mitigation. A UNEP rapid response assessment. United Nations Environment Programme, UNEP- WCMC, Cambridge, UK. 65 p Tscharntke T, Clough Y, Bhagwat SA, Buchori D, Faust H, Hertel D, Holscher DH, Juhrbandt J, Kessler M, Perfecto I, Scherber C, Schroth G, Veldkamp E, Wanger TC (2011) Multifunctional shade-tree management in tropical agroforestry landscapes. J Appl Ecol 48:619–629 UN SDSN (2012) Solutions for Sustainable Agriculture and Food Systems. Technical report for the post-2015 Development agenda. United Nations, Sustainable Development Solutions Network. 108 p USDA (2015) National nutrient database for standard reference. URL: http://ndb.nal.usda.gov Vaast P, Somarriba E (2014) Trade-offs between crop intensification and ecosystem services: the role of agroforestry in cocoa cultivation. Agroforest Syst 88:947–956 van Bael SA, Philpott SM, Greenberg R, Bichier P, Barber NA, Mooney KA, Gruner DS (2008) Birds as predators in tropical agroforestry systems. Ecology 89:928–934 van Huis A, van Itterbeeck J, Klunder H, Mertens E, Halloran A, Muir G, Vantomme P (2013) Edible insects: future prospects for food and feed security. FAO Forestry Paper 171. FAO, Rome van Noordwijk M, Lusian B, Leimona B, Dewi S, Wulandari D (eds) (2013) Negotiation-support toolkit for learning landscapes. Bogor, Indonesia, World Agroforestry Centre (ICRAF) Southeast Asia Regional Program. 285 p van Nordwijk M (2014) Agroforestry as plant production system in a multifunctional landscape. Inaugural lecture at Wageningen University. 16 Oct 2014. http://worldagroforestry.org/ regions/southeast_asia/publications?do=view_pub_detail&pub_no=BL0051-15 van Nordwijk M, Lestari Tata H, Xu J, Dewi S, Minang PA (2012) Segregate or integrate for multifuntionality and sustained change through rubber-based agroforestry in Indonesia and China. In: Nair PKR, Garrity D (eds) Agroforestry – the future of global land use. Springer, Dordrecht, pp 69–104, 541 p van Rikxoort H, Schroth G, La¨derach P, Rodrı´guez-Sánchez B (2014) Carbon footprints and carbon stocks reveal climate-friendly coffee production. Agron Sustain Dev. doi:10.1007/ s13593-014-0223-8 Vázquez-Moreno LL (ed) (2014) Compendio de Buenas Prácticas Agroecolo´gicas en Manejo de Plagas. Editora Agroecolo´gica. La Habana, Cuba. 302 p Villachica H (1996) Frutales y Hortalizas Promisorios de la Amazonia. Tratado de Cooperacion Amazonica, Lima, 367 p Villamor GB, Djanibekov U, Le QB, Vlek PLG (2013) Modelling the socio-ecological system dynamics of rubber agroforests to design reward mechanisms for agrobiodiversity conservation. 20th International congress on modelling and simulation, Adelaide, Australia, 1–6 Dec 2013 www.mssanz.org.au/modsim2013 von Maydell HJ (1986) Trees and shrubs of the Sahel: their characteristics and uses. GTZ, Eschborn Walder F, Niemann H, Natarajan M, Lehmann MF, Boller T, Wiemken A (2012) Mycorrhizal networks: common goods of plants shared under unequal terms of trade. Plant Physiol 2012:789–797

2116

R.G. Muschler

Wambugo C, Franzel S, Cordero J, Stewart J (2006) Fodder shrubs for dairy farmers in east Africa. Making extension decisions and putting them into practice. World Agroforestry Center, Nairobi, Oxford Forestry Institute, Oxford, UK. 169 p Weller DM, Raaijmakers JM, Mcspadden Gardener BB, Thomashow LS (2002) Microbial populations responsible for specific soil suppressiveness to plant pathogens. Annu Rev Phytopathol 40:309–348 Willey RW (1975) The use of shade in coffee, cocoa and tea. Horticult Abstr 45:791–798 Witt C, Pasuquin JM, Pampolino MF, Buresh RJ, Dobermann A (2009) A manual for the development and participatory evaluation of site-specific nutrient management for maize in tropical, favorable environments. International Plant Nutrition Institute, Penang, 30 p Yamamoto W, Dewi I, Ibrahim M (2007) Effects of silvopastoral areas on milk production at dualpurpose cattle farms at the semi-humid old agricultural frontier in central Nicaragua. Agr Syst 94:368–375 Ye´pez C, Muschler RG, Benjamin T, Musálem M (2003) Seleccio´n de especies para sombra en cafetales diversificados en Chiapas, Me´xico. Agroforesterı´a en las Ame´ricas (Costa Rica) 9:55–61 Young A (1989) Agroforestry for soil conservation. CABI, ICRAF, Wallingford, 276 p Zomer RJ, Trabucco A, Coe R, Place F, van Nordwijk M, Xu J (2014) Trees on farms: an update and reanalysis of agroforestry’s global extent and socio-ecological characteristics. ICRAF Working Paper no 179. Bogor, Indonesia: World Agroforestry Centre (ICRAF) Southeast Asia Regional Program. doi:10.5716/WP14064.pdf

Community Forestry Carsten Schusser

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . What Is a Community Forest? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Community Forestry Institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Community Forestry Outcome Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Community Forestry Actor Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Actor-Centered Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interests of Actors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sequence of Surveys for a Comparative Analysis of Actor-Centered Power in Community Forestry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Preliminary Quantitative Network Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Follow-Up Qualitative Power Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Follow-Up Comparative Quantitative Network Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Annex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Questionnaire: Task and Experiences with Community Forestry . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2118 2119 2121 2123 2125 2128 2129 2130 2131 2135 2136 2138 2138 2138 2141

Abstract

Recent and ongoing international studies on community forestry in developing countries have begun to question the success of the international community forestry concept that was introduced by the end of the 1970s. Though it appears that community forestry does contribute to a positive ecological outcome, further results seem to reveal that other advantages promised by the model, i.e., devolution of power to the local resource users and improvement of their livelihoods, simply do not happen. C. Schusser (*) Forest Policy Expert, associated with Chair of Forest and Nature Conservation Policy, GeorgAugust-Universita¨t Go¨ttingen, Go¨ttingen, Germany e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_59

2117

2118

C. Schusser

But community forestry is a complex collective action by forest users that takes place within a broader network of multiple actors at local, national, and international levels. Apparently, the driving forces behind the programs are actors who are very powerful within the hierarchies. To understand this relationship, it is important to know the involved actors, their power and interests, as well as the outcome of community forestry as such. The following chapter therefore presents an approach which can help to unlock the complexity of community forestry. Keywords

Community forestry • Actor • Power • Interests • Outcomes

Introduction “Even the forestry world is driven by different factors and if you like it or not someone is always more powerful than others.” (Saying of a national community forestry Director, somewhere in Africa). Forests are important as a source for valuable products and fulfill environmental and social services. At the same time, several hundred million extremely poor people depend on forest resources for their daily survival. Therefore, it is necessary that policymakers understand the importance of community forestry , take forest governance seriously, and respond better to the needs of the people living nearby Community forests. “The poor conservation outcomes that followed decades of intrusive resource management strategies and planned development have forced policy makers and scholars to reconsider the role of communities in resource use and conservation. In a break from previous work on development which consider communities a hindrance to progressive social change, current writings champion the role of community in bringing about decentralisation, meaningful participation, and conservation” (Agrawal and Gibson 1999, p. 629). Decentralization approaches started in the end of the 1970s when policymakers and scientists realized that the central managed government systems had failed to stop the still ongoing deforestation. The term community forestry (CF) came into use at the same time, when the UN Food and Agriculture Organisation (FAO 1978) initiated activities and programs related to rural communities and their forest–related activities. Since then, community-based management concepts, in particular CF programs, are established in many countries around the globe. Agrawal (2007) mentioned 36 countries covering around 80 % of the forests from which he estimates that around 10 % are managed as community forests. In addition to the practice, the approach community forestry is highly researched and discussed. All approaches aim to improve the livelihood of local people which depend for their living on the natural resources. It follows the idea that if local people are made responsible for the management by handing over the management rights and some

Community Forestry

2119

benefits, the local people would start to protect the resource rather than to destroy it. The core policy objectives of the program are empowerment of direct forest users, improved livelihood of the direct forest users, and improved forest conditions.

What Is a Community Forest? De Jong (2012, p. 108) states that “Over the years, related terms have been suggested, such as communal forestry, community or communal forest management, community-based natural resource management and others.” The FAO accordingly interpreted the term community forestry already 1978 as “any situation that intimately involves local people in forestry activities. Important for all Community Forest approaches is the linkage among people, forests and the output of forests.” Agraval (2002, p. 41) analyzed several authors and their definitions and states that “The main positive lesson I deriver by comparing these authors are,. . ., members of small groups can design institutional arrangements that help sustainable management of resources.” McDermot (2009) follows partly. She goes further and states that “. . ., community forestry refers to the exercise by local people of power or influence over decisions regarding management of forests, including the rules of access and the disposition of products” (McDermot 2009, p. 158). With this, she highlights the role of the local resource user as being the key for a sustainable resource management. At the same time, she also mentions that the same resource user should decide about possible benefits. Baker and Kusel (2003, p. 8) identified community forests’ objective as being “to conserve or restore forest ecosystems while improving the well-being of the community that depend on them.” Charnley and Poe (2007, p. 301) state that “Community Forestry refers to forest management that has ecological sustainability and local community benefits as central goals, with some degree of responsibility and authority for forest management formally vested in the community.” Summarizing the different arguments, community forestry can be described as follows: Community Forestry is a concept which emphasises the involvement and the “wellbeing” of the resource users to conserve or restore forests. Therefor devolution of power to the forest user is the crucial necessity. Although it varies by places, community forestry shares common goals of improving ecological conditions in forests and encouraging ecologically sustainable forest use practises; increasing social and economic benefits from forests to local communities; increasing forest communities’ access to and control over nearby forests.

Poteete and Ostrom (2004, p. 218) state that “Communal management, for example, occurs when governments grant villagers formal control, but also when local residence exercise de facto control in the absence of formal rights.” Since this

2120

C. Schusser

is the crucial point, much research was and is conducted to investigate the problem of how to solve natural-resource-related problems when these involve local users. The general conclusion is that this requires power devolution to the local users, even at the community level. This can be, as Potetee and Ostrom (2004) stated, achieved formally (“government grants control”) or informally (“in the absence of formal rights”), whereby the absence of formal rights can also be seen as the absence of governmental control. Investigating more, it appears that at least community forest approaches deliver on their promises in that positive ecological outcomes are achieved (Brendler and Carey 1998; Chakraborty 2001; Dietz et al. 2003; Thomas 2006; Charnley and Poe 2007; Adhikari et al. 2007; Singh 2008; Wollenberg et al. 2008; Devkota 2010; Vodouhe et al. 2010; Maryudi 2011; Pandit and Bevilacqua 2011; Schusser 2013a, b). But what about the direct resource users? Maryudi (2011) analyzed community forests in Java, Indonesia, and concluded that local forest users were not benefiting significantly, neither in empowerment nor in livelihood improvements. Devkota (2010) has presented similar findings, and according to Edmunds and Wollenberg (2001, p. 192), it is likely that the poorest forest user has become worse off than before. Shackleton et al. (2002) conclude, “The way in which local people realize the benefits of devolution differs widely, and negative trade-offs, mostly felt by the poor, are common.” In addition, Wollenberg et al. (2008) conclude that neither the comanagement nor the local government model have met the high expectations of the community forest program. A number of researchers (Ribot 2004, 2009; Larson 2005; Blaikie 2006; Dahal and Capistrano 2006) have analyzed the common practice and have shown that decentralization policy is seldom followed by genuine power devolution to the local users. Edmunds and Wollenberg (2001) report similar findings, i.e., that local institutions are vulnerable to external powerful actors and that these powerful actors are more likely to dominate the processes. Agrawal and Gibson (1999, p. 629) suggested that it would be “more fruitful” to focus on “internal and external institutions that shape the decision-making process” and that it is important to know what the multiple interests of the actors are and how they make decisions regarding natural resource conservation. The same is suggested by Shackleton et al. (2002, p. 1): “More powerful actors in communities tend to manipulate devolution outcomes to suit themselves”. Similar findings are made by Schusser (2012b, p. 213) which states that “Outcomes of Community Forestry depend mostly on the interests of powerful actors.” Based on these findings, it is therefore important to know whether the influence of powerful actors on community forestry is dominant enough to drive the outcomes more often than not. If actors’ power is the decisive factor, it is not necessary to know the details of their interventions. Instead of analyzing complex influences, it would be sufficient to identify the power of the actors and their interests. From these two factors alone, it should be possible to predict the outcomes of community forestry. The direct power analysis will add to detailed findings about influences of different actors, a general framework which links the outcome of community

Community Forestry

2121

forestry directly to the power of actors. The advantage of such a rigid framework is its simplicity, which makes quantification and planning much easier. Additionally, the focus on power adds value, because it allows a realistic judgment about the involved actor’s situation, for example, and depends on the situation; the goal of community forestry could be modified to be devolution of power to a certain level, as needed. This might convince the respective powerful actors of relevance to support community forestry. Since powerful actors cannot be easily replaced, community forestry project design should cope with this circumstance. Through the development of a community forestry project design that incorporates the interests of powerful actors as well as those of local users, the chances for a success of the community forestry project can be increased. Following the discourse, the question could be raised if the CF program can ever fulfill all of its promises. It seems to be that CF has a positive outcome for the forest resource, but is this enough to be successful on a long run? So far, none of the countries which established community forests in the last 30 years were finally successful. Some are on a promising stage (like Namibia, Cameroon, Honduras, Indonesia, or Nepal), but the final step to become sustainable has not been taken yet (Schusser et al. 2015; Yufanyi Movuh and Schusser 2012; Maryudi 2011; Devkota 2010). The reader of this chapter will realize that sometimes the author became very scientific. This is done purposely because the author sees this as very crucial points and he wants the reader, while dealing with community forestry, doing it in a certain way because in this way, important lessons learned can be drawn and used for much broader comparison. This might speed up the process of discovering crucial findings important for the sustainability of community forestry. This is the reason why the author presents a chapter which reflects the current community forestry approaches as a total success story. With this view, no simple community forestry recipe can be presented. Since the conditions worldwide are so diverse, such a simple recipe will hardly exist and could never be described in a short book chapter. But what can be done is to identify the driving factors of community forestry which depend on the involved actors. It is therefore very important to know the involved actors, their actor-centered power, and interest relations. If such information is used smartly, it can become the keystone for the success of community forestry. This is the reason why this chapter will present important definitions and community forestry basics as well as a model to measure community forest outcomes, a concept to identify community forestry actors, as well as their power and interests and present a method how these results can be interpreted to understand the complexity of community forestry.

Community Forestry Institutions Worldwide, a variety of forest management institutions exists, ranging from statemanaged and -owned forests, informal and voluntary community forestry management without formal user rights, state–community comanagement with user rights

2122

C. Schusser

and without user rights, private managed with private ownership rights, and community ownership of forestland combined with private ownership of trees or historical community forestry with ideal user rights. The modern term “community forest” refers to forests where the people in the community are responsible for the sustainable management of the forest. Mostly every person in this community is formally an owner of the forest resource, without owning the land and without specifying which actual portion of the forest he owns (ideal share). Through this, he should have access to the forest and its products. Mostly the land rights itself belongs to other legal entities and only the user rights to the forest resources, mostly trees, deadwood, and nontimber forest products (NTFPs), are granted. This differentiates community forests from commune forests which are owned by political communes (i.e., state forests or regional government forests), where the residents have very limited user rights. In contrast, Europe’s community forests already have a long and ancient history. Extensive research has been conducted on the use of common resources, including forests which can be traced back in history, as can be the development of the common use of land and its legal status. All of these issues are still controversial; there is no consensus on these matters. What is sure is that today’s community forests emerged mostly from village cooperatives which oversaw common property, including forests. Throughout history, the structure and ownership of these cooperatives changed and developed in different ways. In the beginning of the eighteenth century, a new concept emerged which argued that common land could be better managed when transferred into private ownership. Therefore, several community divestiture orders were enforced which created the legal base upon which to split up the old village cooperatives. Soon after the divestiture of common land, it was realized that this would not lead to an improved output of the privately managed land. Shortly thereafter, most of the orders were replaced by laws regulating the management of common used and community forests, but until then, most of the old village forests had been attached to political communes or were privatized, and only few survive as community forests. The ideal concept has not changed much since then. This is the reason why the end of the eighteenth century can be seen as the beginning of the European community forestry concepts which are slightly different from the modern established community forests. As mentioned earlier, modern community forests got implemented since the end of 1970. The main establishment started in the southern countries of the world. But recently, even industrialized countries in the north discover or remember more and more the advantages of community forestry. In summary, the following phases characterize the establishment of a community forest which follows the above given definition. 1. Initiation Phase: • Step 1: Awareness Creation and Consultation • Step 2: Registration of Interest and Initialization of the Process • Step 3: Community Organization

Community Forestry

2123

2. Application and Declaration Phase • Step 4: Indicative Land Use and Resource Mapping • Step 5: Demarcation and Approval of Community Forestry Boundaries • Step 6: Socioeconomic Survey and Needs Assessments • Step 7: Provisional Forest Management Plan and Bylaws • Step 8: Developing Benefit- and Cost-Sharing Agreements • Step 9: Negotiating and Drafting a Community Forest Agreement • Step 10: Applying for a Community Forest Declaration 3. Implementation and Monitoring Phase • Step 11: Forest Inventory and Needs Assessment • Step 12: Integrated Forest Management Planning • Step 13: Implementing the Integrated Forest Management Plan • Step 14: Community-Based Monitoring • Step 15: Updating the Forest Management Plan • Step 16: Strengthening Community Forest Management and Organizational Capacity • Step 17: Continues from Step 13

Community Forestry Outcome Analysis The analysis of outcomes is oriented toward the core policy objectives of the concept of community forestry. These are the empowerment of the direct forest user (social outcome), the improved livelihood of the direct forest user (economic outcome), and improved forest conditions (ecological outcome) (Maryudi et al. 2012). The outcomes are operationalized as shown in Table 1. Table 1 presents an overview of the outcomes, their corresponding core objectives, the subcategories with their definition, and the key facts on how the outcomes can be evaluated. The subcategories indicate the level of the impact of community forestry according to their core objectives. The social outcome measures the empowerment by evaluating the means the direct forest user has to influence the management of the forest. It measures the degree to which he can make decisions about the management of the forest. Here, the access to forest-related information and becoming a part of the decision-making are important. In addition, the direct access to the forest and the use of its products empowers the end user. If the three criteria are fulfilled, the social outcome can be evaluated as high. By contrast, if there are limited information, decision rights, and/or access, the social outcome for the direct forest user is determined as intermediate (“middle”). If the direct forest user has no information, decision rights, or access, the social outcome is low. The economic outcome for the direct forest user is measured by the contribution of the forest to his livelihood. The options are all forest products, money from selling forest products, or exclusive access to such community development as school buildings, roads, or water pipes financed by community forestry. The degree to which the economic outcome contributes to livelihood improvement is

2124

C. Schusser

Table 1 Outcomes/core objectives of CF with definition and the key facts Outcome Social outcome

Definition (core objective) Empowerment of direct forest users

Low

No empowerment

Middle

Some empowerment

High

Full empowerment

Economical outcome

Contribution to the livelihood of direct forest users No contribution in livelihood Contribution up to subsistenceb level

Low Middle

High

Contribution above subsistence

Ecological outcome Low

Contribution to forest condition No contribution on forest stands and biodiversity

Middle

Contribution to sustained forest stands

High

Contribution to sustained stands and biodiversity

Key facts Access to forest related information Access to decision making Access to forest land and resources No access to information, decision making and/or forest land and resources Limited access to information, decision making and forest land and resources Maximum access to information, decision making and forest land and resources Forest products Monetary benefits Community developmenta No access to forest products, no monetary benefits and no community development Access to community development which was financed through community forestry and financial benefits and/or products providing subsistence Access to community development which was financed through community forestry and/or financial benefits and/or products supplied above subsistence level Forest growth Biodiversity Observation of decrease in stands and forest area No management activities Observation in increase of stands or forest area Forest Management plans Control of implementation In addition to sustained forest stands monitoring and increase of biodiversity

a

Illegal or legal Subsistence a economy without the possibility to save something

b

compared with the standard of living of the direct forest user. This means that if the economic contribution allows for a subsistence-level standard of living only, the economic outcome can be rated as middle. If the contribution is greater, the outcome becomes high. A small contribution compared to the standard of living will be rated as low, e.g., for Germany, the standard for comparison is the annual average income of households. The ecological outcome is twofold. The first part is sustained stands. This means the forest becomes green, grows, and may become larger. This forest’s sustained stands are rated as middle. The second part of ecology is biodiversity. If the forest contributes additionally to biodiversity, defined by Dirzo and Mendoza (2008) as species biodiversity, genetic biodiversity,

Community Forestry

2125

ecosystem biodiversity, or a combination of these, the ecological outcome can be rated as high. The outcome analysis is part of a sequence design method which is described at a later stage. In this sequence design method, expert interviews are conducted with actors of the community forestry network, and documents and observations are obtained and analyzed applying criteria which are summarized as key facts in the following table.

Community Forestry Actor Classification Since it is important to know who deals with community forestry, it is especially important to define the term “community forestry actor.” Many publications have looked at community forestry and identified different actors as important players. But none of it has defined the term “actors” explicitly. This inconsistency makes it impossible to identify actors and analyze them accordingly. Especially on a long run, such findings can’t be compared. This would be needed to draw common lessons learned about community forestry worldwide. Many publications used the terms “actor” or “stakeholder” to examine interrelations within community forestry, but none of the publications defines the terms theoretically. However, this actor-theoretical perspective is needed for community forestry actor comparison. As Schimank points out (2005, p. 29), “actors are source and bearer of actions.” He also observes that, in accordance with a methodological individualism, actors should be seen as individuals. According to him, this approach will not help much if research tries to cover societal issues, since individuals usually cannot accomplish much in terms of change. This is the starting point of Scharpf’s (2000, pp. 95–107) actor-centered intuitionalism approach, where he highlights the stronger position of composite actors, as opposed to individual ones. In much of the research, actors are seen through the lens of this theory. They are entities that have the possibility of influencing processes in order to achieve their own goals. Bo¨cher and To¨ller (2012) and Blum and Schubert (2011) go one step further and attribute the term “goal” to an actor’s distinct interest. Particularly Bo¨cher and To¨ller (2012) point out the importance of the actor’s interest as a determinant of how the actor acts. This is a crucial point, because an actor’s interests determine the involvement with the program and the way he behaves. For example, Grimble and Chan (1995, p. 123) indicate that stakeholder groups are defined “[. . .] on the basis of each group having a district set of interests that distinguishes.” Coleman (1990, p. 28) relates actors to resources and describes these resources as “things over which they have control and in which they have some interest.” Devkota (2010) and Maryudi (2011) see an actor as an entity that can influence CF outcomes based on its interests and power. According to Hermans and Thissen (2009), actors can “[. . .] influence the world around them, including other actors [. . .].” Schneider (2009, p. 192) defines an actor as an “acting entity which is involved in the formulation and implementation of a policy.” What these researchers all find is

2126

C. Schusser

a situation where an actor has a distinct interest and the possibility of action. To overcome this inconsistency, the following actor definition summarizes these facts and is given as follows: A Community Forestry actor is any entity that has a distinct interest and the possibility of influencing Community Forestry.

This definition allows for the different possibilities for what an actor can be, e.g., an individual person like a sawmill owner or a composite actor like a government institution. It strictly associates the term “actor” to a policy, e.g., community forestry, if it is possible for the actor to influence it. Several scholars conducting research about community forestry conducted comparative research on community-based natural resource management in 11 countries around the world. They mention government, traditional leaders, local government, NGOs, donors, universities, the media, the private sector, alliances, and people’s organizations as actors with an interest and the possibility of influencing the program. The self-descriptions of the actors are easy to acquire empirically, but due to the high diversity between different countries and the vagueness of the terms the actors use, the self-descriptions are not sufficient for a sound identification and comparison. The definition should be based on theory and should describe the identity of the actor well. In this way, the diversity of actors from several studies, seen above, is simplified based on theoretical considerations. Therefore, the introduction of a basic definition of “actor” in community forestry is a fundamental requirement for developing a theoretical actor classification. The role of the specific actor is formulated and legitimized within a society. Depending on the focus of the research, society in general can be divided into different function-based subsystems. These subsystems can be enlarged or adopted based on the research purpose and based on relevant theories. Luhmann (1986, p. 216) proposes political, economic, or social systems as subsystems. The application of these simple theoretical criteria leads to the actors defined in Table 2. Basic roles within the political system are politicians, public administration, boards, donors, and associations. Political theory describes their tasks and their legitimation. In addition, the traditional leaders are identified. They are not part of the formal political system but, at an informal level, still play their traditional roles in many countries and will be classified as politicians. Within the economic system, the classification discriminates between entrepreneurs, consultants, and the forest users and other user group representatives. They all conduct primarily economic activities related to the forest. The forest user group representative is the actor who manages the community forest. He acts formally on behalf of the user group. Finally, the social actors are the research institutions and the media. They define their key tasks as being independent from the political system. All actors exist on different geographic levels. (regional, national, and international levels).

Community Forestry

2127

Table 2 Theoretical actor classification, definition and examples Actor Political Politician

Public administration

Forest administration Traditional leader Board

International donor organizations Association

Support associations Economic Forest user group representative Other user group representatives Forest entrepreneur

Consultant

Societal Research Institutions Media

Religious organisations

Definition

Example

Actor who is elected by the people to fulfil a public mandate and who can legitimise binding decisions Public actor that makes decisions concerning specific problems on the basis of general legal standards, resolving those problems by implementing special measures (Krott 2005) Public administration focusing on forest tasks Actor who is legitimised to fulfil a public mandate and who can legitimise binding decisions for a community Actor formed by politicians, traditional leaders or administrations with public mandate International actor that offers funds for solving problems

Government and ministers, representatives of political parties, parliament, etc. Nature conservation authority, land use authorities, agriculture authorities, police, military, etc.

Actor that articulates the interests of the group he represents and attempts to implement them by lobbying politicians and public administration (Krott 2005) Actor that can be characterised as an association but also offers funds for solving problems Actor that articulates the interests of the local forest users and attempts to implement them Actor that articulates the interests of other community forestry user groups and attempts to implement them Actor using the forest for production or consumption of products and services

Actor providing information, funds and management for another actor, based on an contract

Department of forestry, forest office, directorate of forestry Village leaders, traditional healers, traditional authority, religious leaders, etc. Land-use boards, public-control boards, etc. KfW (German Development Bank), Sida (Swedish International Development Cooperation agency), etc. Forest user group association, carpenters association, foresters association, all etc. All kinds of NGOs which offer funds, health organisations, educational agencies etc. Forest management committee

Village development committee, conservancy management committee, management boards, etc. Sawmill operators, logging companies, professional hunters, illegal loggers, companies or individuals buying products or services etc. Consultants

Actor providing science-based knowledge

Universities, research centres, etc.

Actor distributing and generating information

International and national media, like newspapers, journals, radio and TV stations, etc. All kind Churches, mosques, religious or spiritual associations, etc.

Actor providing spiritual or religious backup

2128

C. Schusser

In general, the individual forest users’ possibility of carrying out collective action, in particular community-based forest management, is seen as an outcome of community forestry. Therefore, the forest user is not forgotten as an actor but is left out, since he is considered to be the actor who should benefit from community forestry (empowerment and livelihood improvements). Nevertheless, as soon as an individual forest user has a distinct interest and the possibility of influencing community forestry, he will become an actor and should be seen as a power player. The actors classified are displayed in the following table.

Actor-Centered Power The actor-centered power approach is defined by Krott et al. (2014) as a social relationship between actors in which one actor can alter the behavior of another actor without recognizing the latter’s will. Actor-centered power is linked to actors directly. They play the role of potentate or subordinate, depending on their power sources and the specific issue at hand. The most powerful actors can be identified by accumulating their roles as potentates. This can be done within the framework of a power network, discriminating a group of powerful actors from a group of weak ones. The model does not assume that the powerful actors are always most powerful because in specific relations they might be forced to the subordinate side. Actorcentered power specifies the following three elements of the general term “power”: • Coercion: altering the behavior of another actor by force • Incentives: altering the behavior of another actor by providing advantages (or disadvantages) • Dominant information (when building up power): alteration of another actor’s behavior due to his accepting information without verifying it Power is assumed only if behavior is altered by force, (dis)incentives, or unverified information. These particularities allow the separation of power from other social relations that alter the behavior of actors. Communication based on verified information is of the greatest importance. If two actors exchange information they both verify, they build a social relationship that is the opposite of a power-based relationship. This kind of communication constitutes political bargaining in which both can make informed decisions as long as all information is shared. In cases in which the outcome of bargaining is driven by dominant information or scarce sources, additional power processes could be identified. Open bargaining about sources means offering to other actors what they most urgently demand for themselves, at least in part. In addition, (dis)incentives are regarded as power because the will of the subordinate in respect of his prior resources is neglected by the potentate applying (dis)incentives. For example, the subordinate gets money for planting trees as long as he overrides his prior will to plant corn. The amount of the power source known as money decides the outcome and not the will of the subordinate.

Community Forestry

2129

The specified power elements are linked to observable factors (see Krott et al. 2014). These include the wielding of power as well as threats and sources. The sources of power offer the best opportunity for collecting information. They are specific and observable, like a weapon, economic resources, or written data.

Interests of Actors Krott (2005, p. 8) defines it as follows: “Interests are based on action orientation, adhered to by individuals or groups, and they designate the benefits the individual or group can receive from a certain object, such as a forest.” For analyzing community forestry, the object is the community forest itself. Therefore, the benefit an actor can receive from a specific community forest is important to know. In theory, the model assumes that the expected benefits directly influence the action of individual actors. The interests are linked to goals of community forestry, obligations, or values, but they are additionally shaped by informal aims. Interests cannot be observed directly, but the link to the actor’s behavior offers a chance to learn about the interests by observing the behavior of actors in the past. Quite often, an actor claims to have ecological concerns or to be convinced of the importance of sustainability, but by looking at his behavior in the past, it becomes evident that his actions can be explained wholly by the desire to achieve quick economic benefit or to augment his sources. How the actor behaves and what he does are indicators that show his interests. That is, if an actor has no interest in a positive biological outcome, he will be indifferent toward instruments measuring biodiversity or specific actions that benefit biodiversity. Therefore, interviews with powerful actors were conducted and field observations were made to assess these behaviors. To know how influential a certain actor is, his interests need to be related to the outcomes of community forestry. To test this, the PIDO (Powerful Interest Desired Outcome) indicator (Schusser et al. 2012a, b) can be used. It shows the powerful actors’ interests in specific outcomes for the final end users. The following scenarios are possible and are presented below: • • • •

PIDO (+1): the powerful actor has an interest in a high outcome PIDO (1): the powerful actor has an interest in a middle outcome PIDO (1): the powerful actor has an interest in a low outcome PIDO (0): the powerful actor has no interest in a specific outcome

A PIDO with the values (+1, 1, or 1) indicates that an actor prefers a specific outcome for the end user. Depending on the interests of the end user or the goals for community forestry, a specific PIDO might be evaluated as being positive or negative. Keeping the official program of community forestry in mind, a result would be assumed to be formally positive if all outcomes are middle to high. The PIDO is the final element needed to test dependencies between the interest of powerful actors and the real outcomes of community forestry. To use the PIDO makes only sense in cases where community forestry already exists over a longer period.

2130

C. Schusser

Sequence of Surveys for a Comparative Analysis of Actor-Centered Power in Community Forestry The following method is designed to identify the actors involved in a local community forest network and their actor-centered power. At the same time, additional information useful for the outcome and actor-interest analysis can be collected. The challenge is to find a sequence of quantitative and qualitative surveys which are suitable to identify the involved actors, to stratify these into a group of powerful actors and less powerful actors, and to observe their specific power behavior. To be practical, all this should be achieved with a small budget and limited time. Therefore, the sequence shown in Table 3 was developed. The preliminary quantitative network survey needs to be conducted to identify actors involved for a specific community forest as well as to stratify them into the two groups mentioned above. The follow-up qualitative power survey analyzes the power resources of the individual powerful actors according to three different power elements of the actor-centered power concept. The follow-up comparative quantitative network analysis builds on the data produced. A straightforward way would be to conduct empirical observations of all members of the network. Several cases show that the network of an individual community forestry comprises approximately 15 actors in average including the speaker of the committee of the community forest, the state forest agencies and other state agencies at different levels, donors, forest-based enterprises, and a number of associations lobbying for community forestry. Estimating on average 2 days of field work for each actor, 30 days would be needed to conduct field work for only one community forest. Keeping in mind that in many developing countries the weather conditions do not allow access to the field during the whole year, a much faster and efficient field survey technique is needed. This is achieved with the method presented. The quality of the single survey is quite similar to the second step in the sequence of surveys because the field observation applies the same combination of quantitative and qualitative questions, documents, and observations directly in

Table 3 Sequence of surveys for power analysis compared with single survey

Quality criteria Validity Reliability

Resource use

Sequence of surveys 1–3 1. Preliminary quantitative network survey High for simple hypothesis Sufficient for identifying the group of powerful actors Low

2. Follow-up qualitative power survey High for complex hypothesis Good due to combination of multiple sources Low

3. Follow-up comparative quantitative network analysis High for complex hypothesis Good due to triangulation of the results of the previous sequence steps Very low

Community Forestry

2131

the forest and the offices of the actors. These quality questions are discussed in the chapter about the second step in the follow-up qualitative power survey in detail.

Preliminary Quantitative Network Survey This huge amount of resources of a single survey approach can be diminished by focusing the observations on the findings of a preliminary network analysis. The method of network analysis follows the theoretical model of a power network closely. The theory assumes that actors are linked by complex power processes which become visible within a network only. The network analysis provides the observer with mostly quantitative tools for describing the power relations. Marsden (2001) draw the attention to the numerous errors which can occur in survey data about networks. The respondent answers within a “four-stage cognitive model: comprehending a question, retrieving relevant information from memory, integrating the information retrieved to develop a judgment about an answer and providing a response within the format given in the survey instrument” (Marsden 2001, p. 380). Trying to cover all these aspects properly would drive the sources needed for the complex survey instruments up. The solution suggested is to simplify the hypothesis. Instead, looking for a complex power network method is looking for a much simpler model only, namely, for the hypothesis that “Within the power network of a specific community forest there are only two groups of actors, powerful ones and less powerful ones.” This hypothesis contrasts two positions, namely, powerful or not powerful, rather than describing power processes exactly. To look for contrasting positions in order to get robust data is suggested by Marsden (1990, p. 456). If complexity is defined as the number of acknowledged variables, their diversity, and the multiple relations between them, it becomes obvious that this hypothesis is simple because it assumes that power is an unspecified attribute of a group of unspecified actors. The information provided from the simple hypothesis is much lower than from a complex network hypothesis. But the hypothesis indicates actors belonging to the powerful group which helps in focusing the follow-up steps of the analysis. The main argument is that for such a simple hypotheses, a preliminary networks analysis is able to achieve high validity. High validity does not require complex data about all individual power relations. Instead, it is sufficient already when the data indicates whether an actor belongs to the powerful group or not. Further, the validity is not hurt a lot when the survey misses one or two actors because the hypothesis did not deal with individual actors but with a group. The instrument used for the preliminary network analysis is a quantitative survey. The first question identifies the actors involved following a snowball technique. Starting with the chairperson of the specific Community Forest User Group Committee, he can be asked which actors he has to deal with within specific community forests. Afterward, this question is repeated to all actors mentioned, always referring to the specific community forest, until no new actor is mentioned.

2132

C. Schusser

Several case studies showed that after the group meets 10–15 actors, no new actor is mentioned anymore indicating that the core group is observed. Each actor is asked simultaneously with the first question other questions regarding the power of the other actors. The external estimation of power has the advantage that the bias of strategic answers about own power is avoided. Of course, also the external estimation has a bias caused by lack of knowledge and lack of willingness to tell about their knowledge. For the special case of looking for powerful actors, the lack of knowledge is regarded as low because the powerful actors influence other actors who feel them and know them within the context of the community forestry. General experiences of network analysis support the assumption because data about strong ties and about local networks are better. In contrast, this kind of survey is not very strong for the identification of weak actors, since most in the network pay little attention to them. Due to the prominent position of powerful actors, the first question to identify other actors is an indicator for power already. If actors are not mentioned at all, they can be considered as not powerful from the point of view of the specific actor asked. The social desirability bias caused by the selection of “social and political correct” answers instead of own opinion exists and might be higher in surveys conducted in countries with an uncertain justice system like in many developing countries. Even if an actor understands the question well, it might be that he avoids speaking about the power of other actors. Due to this bias, the reliability of the survey can be estimated as low to use the data for complex network analyses. But the reliability is sufficient to identify some of the powerful actors. The improvement by the follow-up qualitative survey is important. The survey measures the power of the actors in a quantitative manner, meaning that numeric data count how strong the power is. It creates standardized measures based on the theory of actor-centered power before data collection. As described earlier, the actor-centered power theory defines power as a social relationship in which actor A alters the behavior of actor B without recognizing B’s will. Altering the behavior can be achieved by coercion, incentives, or trust. In order to measure incentives, the actors can be asked, directly, from whom they had received any kind of incentives. This information can be transcribed in a Likert scale: the answer yes into a 1 and the answer no into a 0. In the same simple manner, they should be asked whom they trust in the network. Assuming that answering questions about trust is impossible to ask directly, indirect questions should be used. This can be done by asking “Who provides you with information?”, “How good was this information?”, and “Did you ever verify this information?” Finally, “coercion” as power should also not be addressed directly but rather with two questions: “Apart from the information and incentives provided, do you still need one or more actors to carry out your involvement in community forestry?” and “Do you need the permission of one of these actors?” If both answers are “yes” with regard to specific actors, it can be assumed they have strong coercive power whereby only one yes would refer to some coercive power and two no’s to no coercive power. All results should be coded in a Likert scale (see Annex for it). As many external estimates for the specific

Community Forestry

2133

power elements for each actor can be received as there were other actors in the network. The multiple external estimates are stable against the bias which would be inevitable if they were to ask an actor about his own power. Based on the data of all external estimates, the power for each actor for the three elements of coercion, incentives, and trust can be calculated separately. Having summarized estimates for each actor, the task remains of determining the group of most powerful actors. If all actors are weak but two are relatively stronger, these two should comprise the group of the most powerful. On the other hand, actors should not become part of the group of the most powerful, even if they are strong, if there are some other actors with a similar power level. The dominance degree is a suitably sensitive measurement to differentiate the relational habit of power in a network. The dominance degree can be calculated in the following way1: n Total number of identified actors Xi Sum of answers per actor and for one power element, 0 > Xi  ðn  1Þ

highes possible asnwer of the coresponding Likert scale ð1 or 3Þ, for n X Xi ¼ Total given answers per power element i ¼ 1, . . . , n , i¼1

hi

Ratio of power per actor and per power element (i), with 0 > hi  1, and n X hi ¼ 1 ¼ Total power per power element f or i ¼ 1, . . . , n and

r

Position of the sorted ratio of power per actor (hi); the sorting starts with the highest value until the lowest; equal values can be sorted continually anyway f or r ¼ 1, . . . , n Number of considered powerful actors Concentration ratio, shows the distribution of the power per actor (i.e., CR2 = 0.4 means that the first two actors hold 40 % of the total available power per power element in the network) Dominance Degree (Herfindahl-Dominance Degree or Deeffaa-Degree), with m = group of powerful actors and n  m group of less powerful actors

i¼1

m CRm

Dm

Xi hi ¼ X n i¼1

Xi

CRm ¼

m X hr

Dm ¼

j¼1

ðCRm Þ2 ð1  CRm Þ2 þ m nm

The point for the separation between the group of powerful actors and less powerful actors can be found at the maximum of the dominance degree values (highest Dm value). At this point, the Dm value for the last member of the group of powerful actors is still higher than the Dm value of the first member of the group of

1

Adopted from Duller and Kepler (2005, pp. 348–351).

2134

C. Schusser Dominance degree (Dm) value of the Power Element Trust for each actor of the Mbeyo Community Forest Network, Namibia 0,16 0,13

0,14 Dm-value

0,12 0,1

0,14

0,12

0,13

0,11

Dm- curve 0,12

0,11

0,09

0,10

0,09

0,09

0,08

0,08 0,06 0,04 0,02

0,00 Bo ar d En te rp Ba ric se e 1 d En te rp ric e 2

Ba se d

Fo re st

Fo re st

4 St at e

oc ia tio n

As s

2

1

3 St at e

St at e

St at e

Au Tr th ad or ity iti on 2 al Au th or ity 1

D on or

Tr ad iti on al

U se rG

ro up

R ep re se Fo nt re at st ive Ad m in is tra tio n

0

Graph 1 Dominance degree (Dm) value distribution for the power element of trust for all actors of the Mbeyo Community Forest Network, Namibia

less powerful actors. This is the point where the power mean value (Dm) for the assumed group of powerful actors plus the power mean value of the assumed group of less powerful actors is higher than in the following assumed actor- power constellation. As an illustrative example, Graph 1 shows the distribution of the dominance degree values for all actors, sorted from the strongest to the weakest, measured for the power element of trust. The peak is with the fifth actor, indicating that these five are members of the most powerful group. Based on the dominance degree, the group of most powerful actors is identified. Table 4 shows the group to which an actor belongs, for each power element (Trust, Incentives, and Coercion) for the quantitative and qualitative sequence as well as for the triangulated result. The result of the preliminary network survey (QT data in Table 4) is found using the rule which states that each actor who is part of the most powerful group with regard to at least one power element is considered to be part of the group of the most powerful actors. The actors in Table 4 are sorted according to the earlier introduced actor classification. Summing up, the preliminary network survey produces quantitative results indicating the members of the most powerful group. The resources needed to conduct this sequence are small. There are only some standardized questions which can be ticked quickly by the actors asked. Due to the size of the network, of approximately 15 actors in average, the survey for one community forest is done within 1 week. Of course, the empirical indicators are not sufficient for a power analysis, but they are a good starting point for a follow-up survey which would go deeper by focusing on the powerful actors only. A sample questionnaire to process as well as calculate the values for the preliminary quantitative analysis is provided in the Annex to this chapter.

Community Forestry

2135

Table 4 Quantitative, qualitative data and triangulated results for all power elements for the Mbeyo Community Forest Network, Namibia

Follow-Up Qualitative Power Survey The follow-up survey examines the power sources of the actors belonging to the group of most powerful actors individually, in a qualitative manner. The observations look for empirical evidence of specific power sources or processes within the framework of the three elements of power defined theoretically. For example, coercion can be exercised by using a power source or threatening to use it only. The power source could be the rifle of a forest guard, the physical strength of a truck, or igniting a fire. Qualitative, in-depth interviews shed light into such power features. They are accompanied by observations and secondary data like a forest management plan or law, written meeting minutes, and guidelines or letters of formal acts from the field. The interviewer identifies an empirical phenomenon and sees whether he can find a relation to the power element. If he can, the phenomenon supports the existence of the specific power element. For example, the possession of a rifle by a forest guard indicates that he can exert considerable coercion over a forest user with no gun. The hypothesis specified in the power features becomes complex. If no evidence could be found, the observation should be disregarded.

2136

C. Schusser

Conducting a qualitative field investigation which makes use of observations, interviews, and all kinds of documents requires good access to the field actors. An initial meeting between the researcher and actors for the purpose of introductions and the exchange of arguments which are largely symbolic is followed by other meetings which are more substantial. About 10 days were needed to carry out the field investigation of the five powerful actors which were identified within one case, on average (Schusser 2012a, b). In comparison with the quantitative preliminary survey, this means that the time spent with each interview partner is 400 % higher, but the overall time per case study is only 30 % higher (Devkota 2010; Maryudi 2011). The strict focus on the powerful actors increases the efficiency of the survey. This means the field observer can spend more time with the most relevant actors, looking for documents and making his own observations, which increases the reliability.

Follow-Up Comparative Quantitative Network Analysis The comparative quantitative network analysis builds on the data of the preliminary sequence triangulated with the results of the qualitative investigation. The triangulation follows the simple rule that if an actor is powerful, some evidence for it can be found during the qualitative follow-up survey. This means that if a proof or disproof of the results from the preliminary quantitative survey can be made with the qualitative survey, the triangulated result will be the finding of the second survey. Only if no information can be collected during the second survey will the result of the triangulation always be not powerful, regardless of the result from the first survey. For each power element quantified by the preliminary survey, qualitative support has to be found. If the quantitative data indicates a power element of an actor, the qualitative follow-up survey must identify power features. For example, if the survey estimated high coercive power, the qualitative investigation must find a “smoking gun” somewhere. The qualitative survey cannot quantify the power elements but rather, guided by theory, it looks for empirically based evidence of power features which may be a strong indicator as to whether they exist. Otherwise, the quantitative data is not recognized as being reliable and review them giving priority to the qualitative information (see Table 4). Giving stronger credit to the qualitative survey is justified by methodical arguments: (i) The quantitative survey is done in a methodical rudimentary way asking a few questions only in order to save resources. The data indicate the group of powerful actors but not more. For example, no complex network indicators are calculated, and a most simple scale with “0 (no),” “1 (some),” and “2 (strong)” for describing the quantitative results should be used. (ii) In contrast, the qualitative survey is done combining interviews, documents, and observations. The results relay not only on the judgment of the actor asked

Community Forestry

2137

but also on the documents which prove the answers and on observations, e.g., of his technical sources. (iii) The qualitative survey is linked much better to specific findings than the quantitative survey which measures a general power relation. If the quantitative survey indicates a powerful actor in general, it is not possible to describe his power process and sources. The weak link to detailed evidence justifies additionally overruling the quantitative data by qualitative ones. Even in the rare cases, whether the quantitative data are better remains unproven. (iv) Giving priority to qualitative data derives the question why to rely so much on the strong actors identified in the quantitative survey. First, it is not done fully. The qualitative survey may omit strong actors or add some if the data give evidence for power sources and processes. Second, it might oversee some powerful actors due to the weakness of the quantitative survey and the focus of the qualitative survey on the actors identified by the quantitative survey. Underestimating the powerful actors is not destroying the ability to identify powerful actors and to determine the outcome. A positive result can be seen as a proof. Of course, if this phenomenon turns out frequently, additional surveys in order to find the hidden powerful actor should be carried out. The preliminary actor-power network is reviewed focusing on the powerful actors based on the qualitative data. For example, in Table 4, and for all three power dimensions, the data for “powerful” (2) and “not powerful” (1) are examined to see whether they are supported by the qualitative results, and they are corrected in case of abbreviation. The final data goes into the follow-up comparative quantitative network analysis. The first two steps in the sequence build up a quantitative data set which comprises all cases (powerful actors per community forest) from all countries. All actors of the power networks of the community forests studied for all countries are classified according to their power elements as being “powerful” or “not powerful.” This set of data can be used for the quantitative comparative analysis of more complex hypotheses about power. The main progress of the comparative analysis is that it can classify the actors into categories which are meaningful. In line with the guiding question who decides about community forestry, the method can describe power processes and resources. But one restriction caused by the empirical method applied is the focus on powerful actors. The identification of weak actors and their specific power processes is not covered by this design. As discussed, this restriction can be justified with the hypothesis that in explaining the outcomes, the powerful actors make the difference. For example, state agencies use power, which can be set against the elements of coercion, incentives, and trust. For example, the quantitative data can prove whether state forest agencies, in case they are powerful, rely more on coercion or on trust in managing community forests, which is highly relevant for the discourse on governance.

2138

C. Schusser

Conclusion Community forestry assumes that the local resource user is the key for the success of a program. Since the program emphasizes that as a crucial point, this approach is not questioned. The method presented addresses the question of who drives CF at this present stage. Current studies clearly indicate that, so far, the forest user is not the one who determines this. However, certain actors have taken the chance to improve their positions. This is clearly visible by looking to the forest administration. Community forestry can help the forest administration to increase the governmental control over the forest resources through the involvement of the forest user (devolution of power). Nevertheless, the above already presented statement from Shackleton et al. (2002, p. 1) brings it to the point: “More powerful actors in communities tend to manipulate devolution outcomes to suit themselves.” The method presented in this chapter provides a tool to easily analyze the existing power structure of actors involved with community forestry. It facilitates the analysis of the actor’s interests as well as archived outcomes in case the program is implemented already for a certain time. The results of such an analysis can help to adjust existing programs or to plan new initiatives accordingly. If powerful actors and their interests are known, an appropriate approach can be developed. Most important is that community forestry depends on the people who live in community forests. They only do so if they see certain benefit for themselves. The forest administration is interested in a certain surface of forest resources which are essential to justify their existence. On the other side, local people are interested in the resources for use. If the local user has the feeling that the resources somehow belong to him, he might have a greater interest to protect the forests on a long run. Since the local user is not the only one who decides about the outcomes of community forestry, a sustainable way can only be achieved if the right mix of decision-making power, satisfaction of interests, and ownership of the forest resources is found.

Annex Questionnaire: Task and Experiences with Community Forestry Researcher: Interviewee: Community Forestry: Date: Please complete the table: Mention all actors you deal with related to the specific Community Forest (CF): Questions 1–5 1. Who of these actors provides you with information related to the specific CF and how good was this information according to your own judgment? (0 no or

Community Forestry

2. 3. 4.

5.

2139

unacceptable information, 1 acceptable- good information, 2 very good information) Have you ever verified this information? (0 always, 1 never, 2 sometimes) Who of these actors provides incentives (0 no incentives, 1 material or moral incentives or disincentives) Apart from the information or provided incentives is one of these actors still needed to carry out your activities related to the specific CF? (0 not needed, 1 needed) Do you need to get permission from one of your mentioned actors to carry out your activities related to the specific CF? (0 not needed, 1 needed)

Name of actor

1: Tq Information 0 no or unacceptable 1 good 2 very good

2: Tv Info verified 0 always 1 never 2 sometime

3: I Incentives 0 none 1 yes

4: Ci Needed actor 0 not needed, 1 needed

5: Cp Permission 0 not needed, 1 needed

Quantitative Actor- Power Network Analysis and calculation instructions for the preliminary quantitative power analysis Calculate based on the entries in the above table the values for the different power elements as followed:   Power element Trust T ¼ T q  T v , possible cases: • No Trust Power: T0 ¼



        T q ¼ 0 ^ ð T v ¼ 0Þ _ T q ¼ 0 ^ ð T v ¼ 1Þ _ T q ¼ 0 ^ ð T v ¼ 2Þ _       T q ¼ 1 ^ ðT v ¼ 0Þ _ T q ¼ 2 ^ ðT v ¼ 0Þ , code with 0

• Some Trust Power: T1 ¼



      T q ¼ 1 ^ ðT v ¼ 1Þ _ T q ¼ 1 ^ ðT v ¼ 2Þ _ T q ¼ 2 ^ ðT v ¼ 2Þ , code with 1

• Full Trust Power: T2 ¼

   T q ¼ 2 ^ ð T v ¼ 1Þ ,

code with 2

Power element Incentives ðI ¼ IÞ, possible cases: • No Incentives Power I 0 ¼ f0g,

code with 0

2140

C. Schusser

• Incentives Power: I 1 ¼ f1g,

code with 1

  Power element Coercion C ¼ Ci þ Cp , possible cases: • No Coercive Power:    C0 ¼ ðCi ¼ 0Þ ^ Cp ¼ 0 ,

code with 0

• Coercive Power: C1 ¼



      ðCi ¼ 0Þ ^ Cp ¼ 1 _ ðCi ¼ 1Þ ^ Cp ¼ 0 ,

code with 1

• Strong Coercive Power:    C2 ¼ ðCi ¼ 1Þ ^ Cp ¼ 1 ,

code with 2

Enter the calculated power element values accordingly into the following Table:

Name of Community Forest: T: trust: 0,1, 2,(0 no T -2 full), I: Incentive: 0 or 1 (0=none, 1= yes), C: coercion: 0 , 1 , 2(0= none, 1=some, 2 strong) Power Actor's name Factor Actor's name Actor 1

Actor 2

Actor 3

Actor 4

Actor 5

Total (xi)

Actor 1 T I C T I C T I C T I C T I C T I C

Actor 2

Actor 3

Actor 4

Actor 5

Actor x

no entry

no entry

no entry

no entry

no entry Sum Sum Sum

Sum Sum Sum

Sum Sum Sum

Sum Sum Sum

Sum Sum Sum

Sum Sum Sum

Community Forestry

2141

Sort the summarized power elements for all actors with the highest value to the lowest value per power element and proceed as followed

The peak on the graph dominance degree graph (Dm) indicates the point of separation between the group of powerful actors (Actor 4 and Actor 5) and the group of less powerful actors (Actor 1, Actor 3 and Actor 2).

References Adhikari B, Williams F, Lovett J (2007) Local benefits from community forests in the middle hills of Nepal. For Policy Econ 9:464–478 Agrawal A (2002) Common resourcces and institutional sustainability. (Ostrom edn.) Natl Academy Pr, Washington, DC. Agrawal A (2007) Forests, governance, and sustainebility: common property theory and its contributions. Int J Commons 1:89–110 Agrawal A, Gibson C (1999) Enchantment and disenchantment: the role of community in natural resource conservation. World Dev 27:629–649

2142

C. Schusser

Baker M, Kusel J (2003) Community forestry in the United States: learning from past, crafting the future. Islend Press, Washington, DC Blaikie P (2006) Is small really beautiful? Community- based natural resource management in Malawi and Botswana. World Dev 34(11):1942–1957 Blum S, Schubert K (2011) Politikfeldanalyse 2., aktualisierte Auflage. Springer VS, Wiesbaden Bo¨cher M, To¨ller AE (2012) Umweltpolitik in Deutschland, Eine politikfeldanalytische Einf€uhrung. Springer VS Brendler T, Carey H (1998) Community forestry, defined. J For 96(3):21–23 Chakraborty R (2001) Stability and outcomes of common property institutions in forestry: evidence from the Terai region of Nepal. Ecological Economics 36:341–353 Charnley S, Poe M (2007) Communit forestry in theory and practise: where are we now? Anu Rev Anthropol 36:301–336 Coleman JS (1990) Foundations of social theory. The Belknap Press, Cambridge, MA Dahal G, Capistrano D (2006) Forest governance and institutional structer: an ignored dimension of community-based forest management in the Philippines. Int For Rev 8(4):377–394 de Jong W (2012) Discurse of community forestry. In: Arts B, van Bommel S, Ros-Tonen M, Verchoor G (eds) Forest- people interfaces. Wageningen Academics Publishers, Wageningen, pp 107–118 Devkota R (2010) Interests and powers as drivers of community forestry: a case study of Nepal. University Press Goettingen, Go¨ttingen Dietz T, Ostrom E, Stern P (2003) The strugle to govern the commons. Science 302:1–8 Dirzo R, Mendoza E (2008) Encyclopedia of ecology. Academic, Stanford Duller C, Kepler J (2005) Die o¨sterreichische private Krankenversicherung- Ein Monopol?, Austrian J Stat 34:348–351 Edmunds D, Wollenberg E (2001) Historical perspectives on forest policy change in Asia: an introduction. Environ Hist 6:190–212 Grimble R, Chan MK (1995) Stakeholder analysis for natural resource management in developing countries: some practical guidelines for making management more participatory and effective. Nat Res Forum 19(2):113–124 Hermans LM, Thissen WAH (2009) Actor analysis methods and their use for public policy analysis. Eur J Oper Res 196:808–818 Krott M (2005) Forest policy analysis. Springer, Dordrecht Krott M, Bader A, Devkota R, Schusser C, Maryudi A, Giessen L, Aurenhammer H (2014) Actorcentred power: The driving force in decentralised community based forest governance. For Policy Econ 49:34–42 Larson A (2005) Democratic decentralisation in the forestry sector: eassons learned from Africa, Asia and latin America. In: Colfer C, Capistrano D (eds) The politics of decentralisationforests, power and people. Earthscan, London ¨ kologische Kommunikation. Kann die moderne Gesellschaft sich auf Luhmann N (1986) O o¨kologische Gefa¨hrdungen einstellen? Westdeutscher Verlag, Opladen Marsden P (1990) Network data and measurement. Annu Rev Sociol 16:435–463 Marsden P (2001) Survey methods for network data. In: Socott J, Carrington PJ (eds) The sage handbook of social network analysis, London, SAGE: 370–388 Marsden PV (2012) Survey methods for network data, in Scott J, Carrington PJ (eds) The SAGE Handbook of Social Network Analysis. London, SAGE: 370–388 Maryudi A (2011) The contesting aspirations in the forests: actors, interests and power in community forestry in Java, Indonisia. University Press Goettingen, Go¨ttingen Maryudi A, Krott M (2012) Poverty alleviation efforts through a community forestry program in Java, Indonisia. J Sustain Dev 5(2):43–53 Maryudi A, Devkota R, Schusser C, Yufanyi C, Salla M, Aurenhammer H et al (2012) Back to basics: considerations in evaluating the outcomes of community forestry. For Policy Econ 14:1–5

Community Forestry

2143

McDermot MH (2009) Equity first or later? How US community-based forestry distributes benefits. Int For Rev 11(2):207–220 Pandit R, Bevilacqua E (2011) Forest users and environmental impacts of community forestry in the hills of Nepal. For Policy Econ 13:345–352 Poteete A, Ostrom E (2004) In pursuit of comparable concepts and data about collective action. Agr Syst 82:215–232 Ribot J (2004) Waiting for democrazy: the politics of choise in natrural resource decentralisation. World Resource Institute, Washington, DC Ribot J (2009) Authority over forests: empowerment and subordination in Senegal’s democratic decentralisation. Dev Chang 40(1):105–129 Scharpf FW (2000) Interaktionsformen: Akteurszentrierter Institutionalismus in der Politikforschung. Westdeutscher Verlag, Opladen Schimank U (2005) Die Entscheidungsgesellschaft, VS Verlag f€ ur Sozialwissenschaften, Wiesbaden Schneider V (2009) Akteurskonstellationen und Netzwerke in der Politikentwicklung. In: Schubert K, Bandeloe NC (eds) Lehrbuch der Politikfeldanalyse 2.0. Oldenbourg Verlag, M€unchen, pp 191–218 Schusser C (2012a) Community forestry: a Namibian case study. In: Broekhoven G, Svanije H, von Scheliha S (eds) Moving forward with forest governance. Trobenbos International, Wageningen, pp 213–221 Schusser C (2012b) Who determines biodiversity? An analysis of actors, power and interests in com-munity forestry in Namibia. For Policy Econ. Special issue: Biodiversity & Climate Policy. doi:10.1016/j.forpol.2012.06.005 Schusser C (2013) Comparative analysis of community forestry: theoretical and methodological requirements. VVB Laufersweiler, Giessen Schusser C, Krott M, Devkota R, Maryudi A, Salla M, Yufanyi Movuh MC (2012) Sequence design of quantitative and qualitative surveys for increasing efficiency in forest policy re-search. AFJZ 183(3/4):75–83 Schusser C, Krott M, Logmani J, Sadath N, Yufanyi Movuh MC, Salla M (2013a) Community forestry in Germany, a case study seen through the lens of the international model. J Sustain Dev 6(9):88–100. doi:10.5539/jsd.v6n9p88 Schusser C, Krott M, Logmani J (2013b) The applicability of the German community forestry model to developing countries. Forstarchiv 84:24–29 Schusser C, Krott M, Yufanyi Movuh MC, Logmani J, Devkota RR, Maryudi A, Salla M (2015) Powerful actors as drivers of community forestry – results of an international study, forest policy and economics. Special issue: Community Forestry. doi:10.1016/j.forpol.2015.05.011 Shackleton SC, Wollenberg E, Edmunds D (2002) Devolution and community-based natural resource management: creating space for local people to participate and benefit? Nat Resour Perspect 76:1–6 Singh P (2008) Exploring biodiversity and climate change benefits of community- based forest management. Glob Environ Chang 18:187–195 Thomas C (2006) Conservation success, livelihoods failure? Policy Matters 14:169–179 Vodouhe FG, Coulibaly O, Adegbidi A, Sinsin B (2010) Community perception of biodiversity conservation within protected areas in Benin. For Policy Econ 12:505–512 Wollenberg E, Iwan R, Limberg GM (2008) Locating social choise in forest co-managment and local governance: the politics of public decision making and interests. In: Sikor T (ed) Public and private in natural resource governance: a false dicotomy? Earthscan Research Editions, London Yufanyi Movuh MC, Schusser C (2012) Power, the hidden factor in development cooperation. An example of community forestry in Cameroon. Open J For 2(4):240–251. doi:10.4236/ojf.2012

Wildlife Management in the Tropics: An Overview Johannes Bauer

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Definitions and Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . From Hunting to “Game Management” to “Wildlife Ecology” and “Wildlife Management” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Four Elements of Wildlife Management (WM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wildlife and/vs Biodiversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wildlife as a Nontimber Forest Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wildlife and Global Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Use of Wildlife: The Awkward Space of Firearms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Worldwide Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . African Wildlife: Between Myth, Colonial Legacy, and Modernity . . . . . . . . . . . . . . . . . . . . . . . . . The Asian Wildlife Dilemma: Between Economic Boom and Rural Poverty . . . . . . . . . . . . . . . Nepal’s and India’s Tourism Industry in a Vanishing Landscape, the Terai . . . . . . . . . . . . . Wildlife in the Neotropics: Between Overhunting, Habitat Destruction, and Enlightened New Policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Oceanic Tropical Islands: Extinction Traps and Wildlife Sanctuaries . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Status of Wildlife . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Status of Wildlife Harvest and Hunting/Fishing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2147 2148 2148 2148 2149 2150 2150 2151 2151 2153 2159 2162 2164 2168 2170 2171 2172 2172

Abstract

This chapter gives an overview of wildlife management (WM) as it is currently conducted in different tropical regions of the world. These are divided into four WM realms as defined by natural and political geography. It examines, with the benefits of hindsight, what Dasman, one of the seminal American writers on WM, stipulated for the tropics in the early 1960s and how his model of western J. Bauer (*) Australian Carbon Co-operative Ltd., Bathurst, NSW, Australia e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_172

2145

2146

J. Bauer

intervention has been applied around the tropical world as it changed from colonies to independent states. It exposes how western myths have, often negatively, affected that management and how collapses of many traditional and indigenous wildlife management systems, the proliferation of firearms, conflicts, wildlife trade but also the spread of the environmental and protected area movements and tourism have further affected that world. It concludes that wildlife continues to play a crucially important role, in particular for poor and disadvantaged people, for many of which it has however become inaccessible through legislation and global society trends. It also shows, however, that models have started to emerge, often not from the west and community based, which hold promises for the future. What should a forester working in the tropics know about wildlife management? In this chapter for the 2nd edition of the Tropical Forestry Handbook I have chosen the wide view because I believe that our increased specialization and expertise has come at a cost. The understanding of why these things are being done, who does them, what they will do, and, most importantly, what we will achieve by that is what really matters, and that understanding is not so readily available. In order to provide that I have divided this chapter into five sections. In the first one I will show the differences we deal with when we talk about various regions of the vast tropical belt as it stretches around our globe. In this section I have tried to dwell on the specific, aware of the many similarities those regions share. In the second section I have given an overview of the crisis resulting from the growing impacts modern society has on the wildlife in the tropics. In the third section I will present some of the responses to this wildlife crisis by a growing number of parties and stakeholders. This section describes a growing arsenal of tools to better manage wildlife. It shows that global and national communities have developed not only science but, more importantly, frameworks, conventions, programs, networks, and databases, for an informed and unified response. In the fourth section I will identify the programs and approaches where we have made real progress but also will be critical where I think the international responses can – need to – be improved. In that section I will also examine Dasmann’s (1964) premise with the benefits of 50 years of hindsight. In the last chapter I will “reimagine” wildlife management for the tropical world where the new meets the not-so-conventional and where I suggest we have to change our approach. If I have managed to make the reader realize that wildlife management in the tropics is not so much about the application of western science but about the development of sustained visions and activities by a growing number of empowered and collaborating actors, I have succeeded. Keywords

African wildlife • Asian wildlife • Oceanic tropical islands • Wildlife and biodiversity management (W&BD Management) • Global change and W&BD Management in Africa • Global change and W&BD Management in Asia • Global change and W&BD Management neotropics • Global change and W&BD Management oceanic tropical islands • Non timber forest product • Wuyishan Biosphere Reserve

Wildlife Management in the Tropics: An Overview

2147

Introduction In trying to define the relationship between wildlife and people, Dasmann (1964) distinguishes between its commercial, recreational, aesthetic, ethical, and scientific values and introduces the concept of wildlife as a natural resource. This natural good can be, like soils, “mined” and destroyed but also cared for and permanently maintained. He suggests that “the world can now be divided into two areas: the first: where the greatest damage to land and natural resources has been done in the past and where conservation movements are now firmly established”. He puts ‘Anglo-America, Western Europe, the Soviet Union, Japan, Australia, New Zealand’, and “a few other areas” into this category. The other regions, he goes on, are where: “the levels of public education are so low, poverty. . . so widespread, or the pressure of population . . . so great, that destruction of natural resources is still going on at a rapid rate”. He concludes that “conservation practices, although known to some, are not generally applied” and “Many of the countries that are in this category cannot do much about conservation problems themselves, but must rely on outside assistance from the more fortunately situated lands”. What are these lands? He suggests that “Much of Africa, Asia, Latin America is in this area”, where “rapid population growth” (South America), “governmental indifference” (SE Asia), “political turmoil” (Africa), and “widespread destruction of natural areas and native wildlife” (Oceania) are widespread. In short, he means the tropics! In contrast to this rather somber assessment of the tropical world, Dasmann suggests that the status of wildlife conservation in Europe, Australia, and New Zealand is satisfactory. The situation in North America, his homeland, he calls ‘generally satisfactory’ however affected by “rapid population growth” also. If we read this assessment of wildlife management in the tropics now, some 50 years later, we are embarrassingly aware how easy it is, then and now to adopt such a patronizing, colonial view of wildlife and conservation for the “Third World” where in Dasmann’s eyes a combination of high population growth, governmental indifference, and generally a lack of “enlightened attitudes by government” destroyed wildlife and natural ecosystems at an alarming rate. We also realize, however, how that view of the world has been implemented across the tropics as the “outside assistance” from “more fortunately situated lands” Dasmann suggested. And now we must ask ourselves the question whether that “approach,” or should we say “western intervention,” had been the one which was successful, and both morally defensible and required and, if not, how to correct that. In this book chapter I will assess the outcomes of the “wildlife intervention” by the western world in the tropics and attempt to chart a future of wildlife management, which learns from the many often disastrous mistakes made but also gives credit to the ones which worked. I will also, reflective of Dasmann’s premise, “that the tropics must rely on outside assistance from the more fortunately situated lands,” try to find answers as to whether wildlife and poverty in the tropics need

2148

J. Bauer

that unqualified assistance he suggested, or whether there are answers emerging from within and not based on western assistance. And last but not least I will also look at the role of science in that, asking some uncomfortable questions.

Definitions and Terminology Over the years a great many terms and definitions in ecology and wildlife management, often expressed as acronyms, have been established, and those have been further confused by “new” terms such as “Non-(Wood) Timber Forest Products” (NWFP/NTFP) and biodiversity. I have added to that confusion by introducing a number of terms which generally appear in the text as acronyms. Although I do share an aversion to those, as many of my readers do suspect, they do save space; they also might focus our minds.

From Hunting to “Game Management” to “Wildlife Ecology” and “Wildlife Management” The science of game management as used by the American Aldo Leopold in 1933 has been the American equivalent of humanity’s oldest land use and what was, for example, in Germany, “The Science of Hunting” (Jagdwissenschaft), synonymous with land use terms such as agriculture or forestry and/or fishing. In much of the conservation discourse, hunting and to a lesser extent fishing have been replaced by the term “Wildlife Management,” which is more general, some think more “scientific,” and less conflict laden nowadays. I have based the logic of this book chapter on the term WM as the overarching land use activity, which needs to be based on sound science (through research), the development of sensible and harmonized (internationally, between states, between land uses) policies, and national and international legislation. These guide WM systems as implementation tools of the above. Most importantly in the tropics and when applied to wildlife-dependent indigenous people, the overarching framework needs to be developed around what is (or has been) already in place as traditional/ indigenous systems. These it needs to protect and harmonize with new forms of wildlife use and land management (e.g., tourism, protected areas, etc.).

The Four Elements of Wildlife Management (WM) It was Caughley (1977) who suggested that in WM we have three major processes we want to manage. We want to maintain populations as they are, we want to increase them because they are too small, or we want to reduce them because they are overabundant. I have added to this additional processes as follows (Table 1):

Wildlife Management in the Tropics: An Overview

2149

Table 1 The four types of management aims in Wildlife management WC WH WPC

WR

Wildlife Conservation Wildlife Harvest Wildlife Pest Control Wildlife Rehabilitation

We want to retain wildlife in currently healthy wildlife populations We want to harvest a wildlife surplus we would like to sustain +/ We want to reduce wildlife which by high abundance threatens our other aims, this might include the extermination of invasive/alien species We want to increase the low numbers of wildlife (+), most dramatically we want to reintroduce wildlife, which has either disappeared locally (translocations), or gone extinct (Captive Breeding Programs (CBP)

Wildlife and/vs Biodiversity We suggest that the current fad of biodiversity inventories will pass rather soon as a central focus of conservation. Ricklefs and Renner (1994)

Ricklefs and Renner (1994) were only partially right in their prediction, so it seems. Twenty years later, the “fad” might have somewhat faded but it still around wherever one looks – it has become mainstream. While my own assessment of the contents of biodiversity studies (Bauer, unpublished) as opposed to wildlife studies would suggest that it is often little more than a revamp of old themes under new terminology, there is also more to it. Obviously they have underestimated the political tenacity of the term and the domination of biodiversity research through politically and economically motivated research grants (in Australia the great majority one might get). They also have, perhaps, underestimated the need for a generalized inventory and planning tool. What were their reservations? I suggest it is instructive to discuss this briefly. They challenged the term on two grounds: first, the intrinsic value of species lists (the major target of many “biodiversity inventories”) for conservation (which they find questionable), secondly, the philosophical change of such an approach to systematics, denigrating a great scientific discipline and its repositories, museums, to service providers for political ends. They are not convinced that biodiversity inventories (ultimately species lists and often very incomplete) have intrinsic conservation value, they doubt if they influence decision-making, they fear they compromise scientific rigor and integrity (obviously lots of lay people doing them), and they are alarmed that it might sap scarce funds from serious systematic work. These are the scientific and ethical arguments, and they are very valid ones. Twenty years later, however, it seems obvious that the term and species inventories have somewhat developed a life of their own, and if one is just, the term has become a powerful driver for the collection, analysis, synthesis, and extrapolation of that large variety of life we find in biological diversity. It seems also noteworthy that science and in particular computing power and GIS have developed quite

2150

J. Bauer

astonishing new contents around it. I also cannot help but notice, however, that the way biodiversity is applied reflects very much what Ricklefs and Renner (1994) feared and that the term has all but economized the rigor of the discipline of biology and wildlife ecology. As for practicalities, “biodiversity” remains an “unfocused conservation driver” and needs substance to be applied meaningfully. It has also, as Ricklefs and Renner (1994) feared, continued to be compromised by its suitability for the superficial, the political, and the economical. Instead of becoming a serious tool to promote systematics, the science of diversity, biodiversity has all too often become the rallying cry of a very superficial and economic worldview. In this chapter I introduce the term ‘Wildlife & Biodiversity Management’ (W&BDM) in order to reground and rejoin what has become another of our scientific distractions which confuse our minds and our purpose.

Wildlife as a Nontimber Forest Product This is another term for wildlife again and obviously targeting the wildlife in forests (other than wood or timber). If we read the papers about that and in particular the FAO expert meeting discussion (FAO 1995) we can easily see that this has, unlike wildlife and biodiversity, not a biological but a socioeconomic focus and that it is an attempt of international development that tries to describe wildlife in analogy to “agriculture” as a collection of crops and products. While this is entirely legitimate and even has practical value for community planning, if one considers all the ‘biological and ideological ramifications’ which come from the terms wildlife and biodiversity, I have happily replaced this term (as I have done with biodiversity) with wildlife.

Wildlife and Global Change Change is present wherever we look and is accelerating whatever we seem to do, yet the detection of change might not be as straightforward as it would seem, and our ability to detect change seems to be also dependent on the scale we look at. Large-scale changes (biodiversity, for example) seem to be easier to detect than small-scale changes (a species of wildlife, for example) yet more difficult to interpret and are best with great uncertainty. Small-scale changes seem to be everywhere, however most of the time impossible to interpret simply because they mostly reflect dynamics of ecosystems, communities, and populations at local and temporally possibly irrelevant (for management) scales. Only few meaningful data sets to do so are available. One example to detect change of an entire unit of biodiversity over a significant time frame and at a continental scale has been the comparison of bird data collected by hundreds of thousands of amateur ornithologists around Australia between 1977–1981 and 1998–2000. The preliminary analysis of these data as reported in Australia’s

Wildlife Management in the Tropics: An Overview

2151

‘State of the Environment (SoE) of 2001’ suggested, for example, that a total of 65 species had contracted in range to 13 species showing a substantial and systematic decline over this time frame. Interpretation of changes is, however, a difficult task. When five specialists were asked to interpret these changes, there was consensus that four of five species, the brush turkey, the Australian bustard, the wedge-tailed eagle, and the fuscous honeyeater, had genuinely contracted in their range, all of them mostly through loss of habitat, one through habitat loss and loss of its major prey, the exotic rabbit. About the rest – hundreds of species – nobody was so very sure. And this was an analysis based on a data set size and time frame a biologist generally could only dream about.

The Use of Wildlife: The Awkward Space of Firearms The use of wildlife is generally known not as “wildlife management” but as hunting, gathering, and fishing. These three often gender- and age-specific activities in wildlife-dependent societies are based on a wide range of cultural specializations, techniques, and knowledge. They involve tools which range from the cultural and traditional to the highly technological and often lethal (firearms). (Ab)use of such modern weapons, and requisition of those, have adversely affected attitudes and legislation toward hunting. There is also the interface with wildlife, arms, and drugs trade, often connected, which has affected the legitimate users of arms for hunting. Generally, the dialectic and often bitter discourse around that has compromised support for hunting as a traditional and legitimate land use by western aid including charities.

Worldwide Perspective For the purpose of this book chapter I have distinguished between four different wildlife management zones around the tropics (Fig. 1). Three of them are based on biogeography (Neotropical, Asian, and African realms) with one special group “Oceanic Islands” defined by the isolated and generally greatly modified (Tropical Islands and Tropical Australia) nature of island environments. Wildlife management in these four regions is carried out within the political environments of 53 different nations which might (or might not) be signatories to international conventions and are under growing pressures from growing populations and associated resource use (see http://stateofthetropics.org/). Current systems to manage wildlife depend on the countries’ native traditional/indigenous systems and communities, colonial history, and the role of the political system states. For each country agricultural, forestry, water development, protected area, and tourism policies are, generally, of higher importance than wildlife or indigenous legislation. There is also, in each of these places, a multitude of donors and charities carrying out projects and supporting government and communities, each with their political agenda and ideologies. Political history has a major bearing on the current situation.

2152

J. Bauer

Fig. 1 A suggested classification of Tropical Wildlife Management Zones based on biogeography (Neotropics, Tropical Africa, Tropical Mainland Asia (3) and major Oceanic Islands (island groups) around the tropical oceans (Map adapted from State of the Tropics n.d.)

The African region had a long and chequered colonial history. The Neotropics are where, with the exception of some small northern colonies, the Spanish and Portuguese became masters for centuries, while the Asian Region was predominantly English (India) with some Spanish (Philippines) and Dutch (Indonesia) elements. Significantly, either powerful or remote countries resisted colonization (China, Thailand, Japan, Bhutan, Nepal). This different history, and colonial masters who viewed wildlife differently, has also shaped the fate of modern wildlife populations. The modern plight of the tiger in India has to be understood in the context of an imposed British bounty system which decimated populations beyond recovery. So have the Anglo-Saxon experiments with acclimatization of exotic species for hunting and pest control and the commercial exploitation, often ending in the destruction of native wildlife resources. After colonial states gained independence, wildlife, often at greatly lowered densities and affected by land use changes under colonial rule, rarely had a chance to recover. New pressures and political instability, along with disempowerment of indigenous people and loss of traditional land rights of minorities and rural communities, ensured that most of the wildlife did not recover. Significantly, this post-WWII phase was also characterized by increasing exploration for oil and minerals, megahydrodevelopment, industrial agriculture with its proliferation of chemicals, GM crops, and logging, all of them leading not only to vast environmental destruction but also to the loss of land rights and growing impacts and pressures on communities and remaining wildlife. This phase was also characterized by a growing and global wildlife trade for the emerging western pet market, medicinal research, poorly regulated (international) hunting, a dramatic growth of demands on wildlife for the emerging Asian economies, and the development of tropical mass tourism, much of it targeting with its

Wildlife Management in the Tropics: An Overview

2153

development beautiful and natural regions. International fishing fleets, transport, and hundreds of millions of tourists started to invade marine and terrestrial tropical environments, all with their own specific impacts and multiplying the dispersal of alien species. Not surprisingly, this time was also defined by many thousands of species threatened with extinction. While the growing responses (as described in chapter “▶ The Development of Wildlife Governance, Science, and Management Capacity in the Tropics”) from the 1960s onward sought to stop this trend in wildlife decline, the growth of human populations, poverty, and environmental degradation mostly offset these programs. And that was before the growing impacts of a changing climate became clear and science showed many frightening future scenarios, additional and exacerbating to all the old ailments. In this chapter I will briefly describe the current status of wildlife and of wildlife utilization/management in these different regions. Not more than a glimpse, I have tried to emphasize the differences as they unfold around the tropical world.

African Wildlife: Between Myth, Colonial Legacy, and Modernity The survival of our wildlife is a matter of grave concern to all of us in Africa. These wild creatures amid the wild places they inhabit are not only important as a source of wonder and inspiration but are an integral part of our natural resources and of our future livelihood and well-being. From the Arusha Declaration on Wildlife Protection, Julius Nyerere, 1961

Africa, a vast continent, the cradle of humankind, and to many still a nearmythical place where lions and elephants roam in endless numbers, has over the past century been greatly transformed. Now, with the colonial masters all but gone, the modern African states (re)established, often at great costs to humans and wildlife, a mixed picture emerges. There are places such as South Africa, which is overcoming apartheid to become a modern state and thrive economically. There are countries such as some Central African nations which had and continue to have wars killing millions of people. There are also countries such as Namibia or Botswana, where postcolonial legacy, social progress, and the conservation of their environments including wildlife have progressed into modern states. And there are places such as the Gambia where forest cover has increased almost 10 % over the past 15 years. The African Tropics are characterized by a vast continental mass which had its uniquely rich fauna of large mammals (megafauna) often in great abundance. As a continent African ungulates were greatly affected by the introduction of the rinderpest (Italian cattle in Ethiopia), which reached the Cape Horn within few years at the beginning of the nineteenth century devastating wildlife to such an extent that many lions turned to maneaters for lack of food (see Sinclair 1977). Before African wildlife could recover, colonial powers decimated the ungulate fauna until the conservation movement was born with the Serengeti, and country after country started to set aside vast land areas for wildlife

2154

J. Bauer

conservation – often with great costs to human communities. Densely forested regions in Central Africa were for many years and up until the 1970s relatively unaffected by major development, but they have also been opened up and changes accelerate as Africa joins the rest of the world in its legitimate search for a better life. The magnificent African wildlife continues to survive in this modern world, and sometimes against many odds, as it stumbles from one crisis to the next. Species such as the black and white rhinoceros, once numbering in their millions, then almost extinct, recovering again in the 1990s, are now greatly endangered again. This recent demise is a result of the high demand for rhino horn (US $90,000 kg) in Asian countries where the middle class has become wealthy. There are countries such as Tanzania which have dedicated almost half of their land (44 %) to wildlife conservation – and created a tourism industry around that (No 1 Destination in the NYT Tourism Hotspot Ranking of last year). This industry employs 27,000 people, attracts almost a million tourists per year, and generates 25 % of its foreign exchange, mostly around its wildlife migrations in the Serengeti and Ngoro Ngoro Crater at Mt. Kilimanjaro. Other countries such as Rwanda lead the tourism world around endangered species, here the mountain gorilla, with significant benefits for gorillas but also for the poor rural people. Perhaps even more significantly, these real societal benefits have played a critical role for the mountain gorilla to survive war and genocide and resume its role as major tourism attraction. Parts of the abstracts of two papers in 1995 and 2010 show how the fates of mountain gorillas and poor people have become entwined. They also show their survival in times of great adversity. Box 1: Gorilla Tourism in Rwanda: A Remarkable and Lasting Success Story

Until April 1994 gorilla tourism was the basis of the Rwandan tourism industry part of which was returned to finance gorilla conservation programmes (Shackley 1995). This study showed that the gorillas had survived the war unexpectedly well and increased political stability has permitted research and protection teams to return. The paper discusses competing gorilla tourism in Uganda and the uncertain future of this industry. 15 years later: Spencely et al. (2010) study shows that Tourism is still the leading export sector in Rwanda and continues to grow. It also shows that it provides income and opportunities for the poor.

The emerging modern African wildlife world remains, however, full of conflict. There are places such as South Africa where a powerful western tourism prevents the reduction of the elephant population in Kruger NP against the wishes of the local communities and the park administration, which is to retain the park’s diversity and plant productivity as a basis for many herbivores. In the same country there are also farmers who have replaced their marginal income from livestock by

Wildlife Management in the Tropics: An Overview

2155

thriving “wildstock” farms, where they breed, sell, and trade endangered wildlife for profits. And, of course, there is the vast Congo region, where the “bushmeat trade” thrives and where diseases such as HIV emerged from primates. Ownership of and benefits from wildlife in Africa have been greatly confused during colonial times, and Tanzania may be used as a rather representative African example on how colonial legacy, the western conservation movement, and the new-found value for wildlife through tourism have created an uneasy and always contested relationship between the power of the state, International Conservation NGOs, the tourism industry, and local communities. Communities are now, after empowerment for some years, in retreat from the power of the state as it seeks its rent from wildlife and often conspires with foreign industries against local communities. Box 2: “The Government’s Animals”. Tanzania: Between National Park Legacy and Contested Community Rights

The use of wildlife in Africa, traditional and modern (tourism including big game hunting) is a fine balance between the power of the state, the leaders and actors it encourages and the communities of people, who had to live under the “yoke” of conservation. I have chosen excerpts from Minwary (2009) who showed the fickle nature of this “participation” and benefit sharing, as communities try to survive, the state seeks its “resource rent” and the industry joins hands with the state and leaders. This reality is almost an allegory of sorts as it describes the nature of what happens in many other parts of Africa and indeed the world where communities want to regain rights over wildlife management but find, that they have only limited power to do so.

And then of course there was the lure of hunting in Africa, stronger than anywhere else and inextricably linked with colonialism. All through Anglo-Saxon colonial literature, from Rider Haggard to modern books, the particular fascination of Africa for the colonial hunter reverberated over centuries. Trophy hunting of the big five (lion, elephant, rhino, buffalo, leopard) was a pursuit of the very rich during the nineteenth and most of the twentieth centuries. In the 1970s and 1980s the general decline of game (more often than not through political instability than trophy hunting) and the establishment of many national parks and antihunting sentiments from the west strongly affected big game hunting. Over the last decade, however, trophy hunting has again expanded, and a serious attempt has been made to include it as a facet of nature conservation (Bauer and Herr 2004; Baker 1997; Baskin 1994; IIED 1994; Lewis and Alpert 1997; Meier 1989). However, while Lewis and Alpert (1997) demonstrate the substantial benefits hunting can bring (e.g., to the Zambian economy), Baker (1997), in an analysis of hunting in the southern parts of Africa, concluded that a lack of appropriate monitoring and exceeding hunting quotas made sustainability doubtful. Additionally, corruption prevents communities from truly benefiting from (hunting) tourism in Tanzania and Botswana. This has currently been reiterated in Tanzania, where community

2156

J. Bauer

participation despite much rhetoric and hype remains elusive (Minwary 2009). In contrast Namibia stands as an example where hunting is carried out based on private landownership, a system also adopted in Southern Africa, where farmbased wildlife breeding, trade, and tourism (including hunting) have developed. Numerous game farms now rely on their hunting income in a significant way. This harvest encompasses approximately 22 species of wild ungulates and provides a very substantial contribution to farm income, with trophy hunting being a superior land use on marginal land (Meier 1989). This and many other studies also show that this “new” land use, if well regulated, can have significant conservation benefits. Box 3: Wildlife Ranching and Hunting in Southern Africa

In 1989, Meier, analysed the profitability of three landuse schemes in Namibia, conventional livestock pastoralism, (1), game farming (2) and trophy hunting (3). He concluded, that anything to do with game was more profitable for the farmers, if it involved marginal grazing land. He further found that trophy hunting was the most favourable economic option on such land and that income derived from it, even compared favourably with livestock on good land. Since this and similar other studies were carried out, farmers and the farmer markets in Namibia and South Africa have reacted and there has been a market adjustment towards it which is so far unique in the world. Farmers have started to rear, trade and sell wildlife instead of cattle and sheep and much of it is done in auctions, where they buy species that have disappeared on their large farms to restock (as fishermen do in many rivers and lakes). These species then propagate and can be either resold during the next auction or sold to a hunting tourist, mostly from Europe, but increasingly so from the US, who wants to: (i) (ii) (iii) (iv)

Have that unique hunting experience for that species Complement his/her collection of hunting trophies Mount it over his/her fire place Complement his/her farm stay holiday in the Savanna with something exciting (v) Have a trophy of that animal that is larger than Geoff’s at home

Over the years this industry has matured and in 2004 according to Damm (2002) 17,569 heads of game were auctioned. While this is a rather impressive number, in particular as it involved many rare and endangered species (e.g., 21 elephants, 39 Lions, 4 scimitar-horned oryx, 1 Black Rhino, 137 White Rhino) it has dropped from 21,101 heads in previously (16.7 %), a fall which Damm attributes to that fact, that the “market” has run its way and “most land, which could be converted back from agricultural to wildlife habitat, has experienced this transformation already” (Damm 2006). (continued)

Wildlife Management in the Tropics: An Overview

2157

Past farmers who wanted to restock their farms have turned into producers, wildlife markets are starting to become saturated and stagnate. Damm (2002) predicts that with “tighter new legislation” with regards to breeding, trade and landownership tax being currently considered this negative trend will continue. This is of course how markets work. The saturation will not work the same for all species and that is evident with some dropping dramatically (e.g., the single black rhino auctioned in 2005 fetched just under 100,000 Rand, down from half million Rand in 2001 and 2002), while for others such as springbuck, kudu, eland and impala prices were “astonishingly” (Damm also, according to the mentality of markets, there have been some farmers who were “inventive”. Exploring new market opportunities e.g., for Bengal Tiger and Water Buffalo, while giving the new income source for farmers a bad name in conservation, something it cannot afford. Hence the need for regulation, but one which is quite achievable.

Controversy also rules in Africa as to what to do about the illegal trade in ivory and rhino horn, a trade that has re-emerged as Asia grew rich and that needs to be fought with great determination, as a recent high-level meeting between heads of state in London decided, in the meantime with the aid of military drone technology. There is, however, also a valid argument to be made that rhino conservation would greatly benefit from making this wildlife product rather a legitimate farmer venture than driving it into illegality by becoming “owned” by western conservation elites and charities. Box 4: “Rhinos Belong to the Future. . . Five Rhino Species Forever”

The current re-emergence of rhino and elephant poaching, after many past, seemingly successful campaigns over more than 30 years, have, once again, demonstrated the precarious existence of the tropical mega fauna (Save the Rhino 2015). This resurgence is closely linked to the emergence of the Asian markets which, continue to cherish (rhino horn, ivory, tiger bones) what the western markets have, either never valued or successfully abolished. The current massive poaching resurge has led to so far unheard of international condemnation, most recently during a conference in London where leaders of 52 countries pledged to dramatically step up efforts and political commitment. Ultimately it will be decided whether Asian leaders in particular China and Vietnam are taking their international commitments more seriously. That the Chinese Public in China is ready is evidenced by the shark fin trade where, almost overnight, a consumer campaign (targeting a highly environmentally conscious Chinese public), endorsed by the government, and drastically reduced it. (See also: http://savetherhinotrust.org). But then again there is (continued)

2158

J. Bauer

another way of viewing this and it can be found on http://www.rhino-econom ics.com/. It looks at the trade less emotional and while the author calls it an “economic” argument one might also call it “evidence-based”. Significantly it reduces if not eliminates a conflict, creates opportunities and income for local communities and saves/grows rhino populations. Saving Rhinos: Success versus Failure In the year 1800 about one million rhinos lived on earth. Today less than 28,000 survive in the wild, due to the combined effects of habitat loss and uncontrolled hunting. Throughout history, humans have hunted rhinos for their meat and other body parts, which are used for ornaments and traditional medicines. Despite this bleak situation, there has been at least one notable success story. The southern white rhino was close to extinct by 1900, but today it is the most abundant species. In 1900 there were less than 50 in the world-today there are more than 20,000! Why has the southern white rhino fared better than the other species and what can we learn from this? Economics provides the answer! White rhino conservation efforts were driven by South Africa, which has developed a vibrant market economy for wildlife within the last 50 years. This economy rests on three pillars: • Recognizing and actively developing legal markets for things that people value about rhinos, such as tourist viewing and trophy hunting • Allowing private landowners to legally own rhinos, thereby giving them strong direct incentives to manage them responsibly • Enabling all landowners (private, communal or public) to retain the money they earn from selling live rhinos and rhino products, thus making rhinos a lucrative long-term investment In the last two decades, the market values of live white rhinos have soared–from around $1,000 a rhino in the early 1980s to more than $30,000 in recent years. These rising values created strong incentives to protect and breed more rhinos. The market approach has also been applied to South Africa’s black rhino population and in neighbouring Namibia. Today South Africa and Namibia protect 75 % of the world’s wild black rhino population of about 5,000 animals. In 1970 there were an estimated 65,000 black rhinos in Africa, mostly in other African countries, but almost all lost their lives to poachers. Rhino populations in other countries are protected by laws against poaching and illegal trade, but there are limited incentives to enforce these properly. Government ownership and trade restrictions simply do not create strong enough incentives to invest in rhino protection and breeding, especially not to the people that matter: the people on the ground, (continued)

Wildlife Management in the Tropics: An Overview

2159

who ultimately decide the rhinos’ fate. We need to learn from the southern African experience! Unfortunately the southern African success remains under threat because of the world’s refusal to recognize a legitimate demand for rhino horn. For more than 35 years, the world has attempted to end the rhino horn trade by banning it–and has failed. Rhino horn demand and illegal trade persists, and the ban appears to have simply driven black market prices to extraordinary levels, with disastrous results. The rhino horn trade ban no longer makes either economic or conservation sense. The natural mortality rate of rhinos in Africa alone yields as much horn as has been poached to supply the market in recent years. Furthermore, rhino horn is a renewable resource that can be easily harvested without killing rhinos. And African conservation agencies and landowners already hold several years’ supply of rhino horn (at the current rate of black market supply). These stockpiles are worth millions of dollars, money that could be usefully spent on rhino conservation, but the ban will not allow them to be sold to raise this money. The rhino horn trade ban is quite possibly the greatest remaining threat to the rhino! Public ignorance and misunderstanding allow this policy to persist. It is time to dispel some myths and think more creatively about the most sensible way to ensure the future of all rhinos.

The Asian Wildlife Dilemma: Between Economic Boom and Rural Poverty Tropical and subtropical Asia, from the Himalayan peaks through the monsoon belts of India and Burma, the rainforests of SE Asia through the vast coastal and island worlds of the countless archipelagos to the great tropical island landmasses of Borneo and New Guinea, each with high mountain ranges, even glaciers, are centers of natural and cultural diversity and history which are unique in the world. They are areas where four biogeographic regions meet, where the collision of India with its Gondwana heritage – the vast continent of Asia – not only created the world’s highest and most extensive mountain range but changed the world’s climate. In this region human migrations, agriculture, and great civilizations have longer and more intensively interacted with a rich tropical world than anywhere else. This deep connection has led to the domestication of the largest terrestrial creature existing on Earth, the Asian elephant, and has brought forth rainforest cultures, which had a uniquely rich way of life around thousands of species of wildlife and going back many thousands of years. Perhaps most significantly, this rich use of wildlife as food and medicine, common to many “primitive” societies, was maintained and greatly developed in the great civilizations, in particular in China and India. Although much has changed and disappeared over

2160

J. Bauer

time, one of the enduring legacies of this interconnectedness of humans and wildlife in Asia is its use for human needs and cultural enjoyment, for example, in a food culture and cuisine which, in the case of China and India alone, encompasses thousands of species of wild plants and animals or in the traditional medicine chests (Chinese traditional medicine, Ayurvedic medicine in India, PNG’s indigenous medicine, etc.) where wildlife continues to bring benefits to hundreds of millions of people and remains a great symbol of status (elephant, tiger, rhino). Asian wildlife is caught between this history, the economic boom, and still widespread poverty. Treading its thin line between tradition and the West, this tension defines the Asian wildlife dilemma. While having the power and wealth to continue its use, Asia now has to learn restraint if it does not want to lose it all and it has risen to the challenge. Asia as a whole has made great strides as it implements protected areas, as this suits the powerful role of its governments. It has, however, been less successful to change the habits of its people or their continuing dependence on its wildlife. While its governments struggle to accept this responsibility, and opportunity, its wildlife continues to disappear, and its cultural diversity, including its cuisine, will be greatly diminished. “The west” has not recognized this “Asian dilemma” as it condemns and tries to control a trade around wildlife, which is so deeply entwined with the peoples’ cultures. It still fails to see that conservation and regulation have mostly led to the emergence of a vast illegal market, while the sustainable use was either made impossible or prevented to improve. With this view the challenge in Asia is not so much any longer to establish more protected areas. The real task of wildlife conservation in Asia is to make societies understand the need to manage wildlife sustainably for its many uses and benefits it can bring. Last but not least, it is the human population which has to have a bearing in our approaches to conservation. Asia contains more than half of the world’s population, and excluding them from protected areas will be no long-term solution. There are many examples of this Asian wildlife dilemma. There is, for example, the modern wild swallow nest trade. Emanating from China and carried out for many centuries it has all but destroyed a great resource for the communities at the South Andaman Coast of Thailand. These people remain excluded to harvest them as they had done for centuries (because national parks are there now), while the “Swallow Mafia,” with unhindered access, has now almost destroyed the resource. Throughout India, Bhutan, and Nepal tourism industries thrive on wildlife and specifically around the tiger without accepting responsibility (Furze et al. 1996; O’Riordan 2002; Eglert 2002). There is a legacy of inappropriate western and hegemonic wildlife legislation in Bhutan and China as it prevents its farmers to kill wild boar, an overabundant agricultural pest (Boyd et al. 2003; Bauer 2002), accompanied by a sad loss of traditional ecological knowledge in the Wuyi mountains of China as conservation goes too far. A similar trend is evident (loss of coral reefs, fish, primates, swallows, and indigenous people) in Thailand’s south while the government ignores the long-term needs and responsibilities of its greatest industry: tourism.

Wildlife Management in the Tropics: An Overview

2161

Box 5: Ecosystem and Wildlife Change in Protected Landscapes in Wuyishan Biosphere Reserve, China

This case study shows, that many protected areas with human communities and activities around and within it continue to change (Bauer et al. 2002; Boyd et al. 2002, 2003). Wuyishan Nature (Biosphere) Reserve in Fujian Province China may serve as an example how, under enlightened community forestry development by the Chinese government, along with far reaching protection of wildlife, one of the uniquely rich natural heritage areas in wildlife has been able to recover from destruction of forest and wildlife until the 1980s. It also has provided however an example which shows how various trends, erosion of TEK, Bamboo community forestry and tourism have introduced a new dynamic setting whose outcomes are difficult to predict and need to be monitored and managed. A main change is the economic and cultural disengagement of the community from wildlife through changes in reliance on and loss of cultural knowledge about wildlife (Fig. 2). – Many people had direct knowledge of species (had seen or seen signs of), and therefore information about species should be considered to have a high level of validity. – There does appear to be an overall decrease in the time spent in the forest by local people. This can be related to a decrease in hunting following protection of the species and also changing lifestyles away from direct reliance on the natural environment for subsistence. – The majority of the study target species are considered to be common or very common, although 54 % of species may be decreasing in abundance. Overall, at least eight species may be considered to be threatened: Crabeating Mongoose, Fox, Large Indian Civet, Dhole, River Otter, Leopard Cat, Leopard and Tiger. These results suggest that negative pressures are ongoing in the reserve. – The results for changing abundance and abundance of species are not conclusive, but may suggest that species that are more often found in disturbed areas (according to interview statements) are more likely to be increasing in abundance or, to a lesser extent, stable and are more likely to be common than rare or locally extinct. Many local people were collecting wild plants in 1998 for food, medicine and sale. This indicates the continuing contribution of the natural environment to the well-being of the Wuyishan Biosphere Reserve community. While the data do not suggest negative impacts on plant species, those which are targeted for collection should be monitored for abundance and condition. – Most of the people interviewed indicated that they bought meat. This is most likely a change from earlier subsistence hunting following species (continued)

2162

J. Bauer

protection. However, collection of small fauna for food and sale was common in 1998. – The available evidence collected during this study suggests that Bamboo monocultures lead to lower levels of species diversity and favour species such as Wild Boar which become agricultural pests. – The collection and sale of small fauna suggest that significant pressure is being placed on frog, snake, fish and possibly rodent populations. While collection for subsistence has probably occurred at sustainable levels in the past, the increasing tourism market presents a threat to species survival. This study concluded that the economy of WBR is at an early stage of change, and that the present adult generations retain knowledge of species, with young generations retaining less knowledge. The results also suggest that unless more monitoring effort and better understanding of humanwildlife interactions are sought by the WNR administration it will not be possible to maintain the present level of mammal diversity and abundance and that increasing population shifts will result as a consequence of economic activities (bamboo community forestry, tourism), pollution and a presently unknown level of wildlife harvest. Education Programs in schools would appear to be of great importance to maintain the knowledge and interest of local people in wildlife.

These changes in wildlife habitat and erosion of traditional ecological knowledge (TEK) in Wuyishan Biosphere Reserve as bamboo monocultures develop might quite likely be reversed as the predominantly (and very fast-growing) Chinese tourism industry with its demand for wildlife experiences (and wildlife food also) grows and it is difficult to predict what the eventual outcome for this landscape is. Nor do gradual and unobserved shifts in “protected” landscapes only occur in human cultural environments and as a result of agricultural, forestry, tourism, or hunting/fishing activities. They are also happening at large landscape scales in the wake of megahydrodevelopment as it changes the face of many terrestrial habitats.

Nepal’s and India’s Tourism Industry in a Vanishing Landscape, the Terai Species also, like river systems, are caught between different uses and aspirations. The last 3,000 or so remaining wild tigers in the world are living a precarious existence between priced Chinese medicine tiger bone item and equally cherished live target for wildlife tourists (Fig. 3).

Wildlife Management in the Tropics: An Overview

2163

Fig. 2 The forest world of Wuyishan Biosphere Reserve (WBR) in Fujian Province, China: Conversion of natural forest into bamboo plantation (1), Traditional House style (2), Local man describing how he saw the last South China Tiger in the 1960s (3), Stump-tailed macaque, a regionally restricted primate species with one of its last strongholds in WBR

The Tiger: A Species of Contradictory Values and Approaches

There is probably no more suitable species than the tiger to describe Asians Conservation Dilemma (Global Tiger Initiative Secretariat 2012; Tepper 2013; Mills and Jackson 1994). Revered by the west, priced in China’s tiger bone market, major tourism drawing card in India’s and Nepal’s National Parks yet deadly neighbour for rural people the tiger has not fared very well over the past 30 years although what must be hundreds of millions of dollar were spent on its conservation. While its wild populations have kept declining despite of all these efforts, its captive population in China alone has not done so badly. As Tepper (2013) reports. A new report by the U.K.-based Environmental Investigation Agency (EIA) suggests that China is knowingly violating its own ban on the trade of tiger (continued)

2164

J. Bauer

bones, as stipulated in a 1993 State Council measure. . . the report, “Hidden in Plain Sight: China’s Clandestine Tiger Trade,” alleges that the government is allowing the use of captive-bred tiger bones for tonic wines thought to have medicinal properties. The EIA believes that several tiger farms in China are using what they claim is a secret government notification issued in 2005 as proof their tiger wine operations are legal. . . The head of group’s tiger campaign, Debbie Banks, expressed outrage on the EIA’s website: “The stark contradiction between China’s international posture supporting efforts to save the wild tiger and its inward-facing domestic policies which stimulate demand and ultimately drive the poaching of wild tigers represents one of the biggest cons ever perpetrated in the history of tiger conservation.”. This report also suggests that “Experts estimate the numbers of tigers in the wild to be between 3,200 and 3,500, although it’s believed that the captive tiger population in China may be as high as 5,000 animals among up to 200 farms and zoos, according to EIA. These farms are often touted as tourist attractions and sell tiger wine on the premises, which underlines the out-in-theopen nature of these operations”. This newspaper report not only shows the ‘two faced approach’ of China to tiger trade bones. It also shows the discrepancy between China and the urban “west”, China’s pragmatism towards most things, including tigers and its almost casual own approach as it legalises tiger bone trade, something many advocates in Africa also suggest would reduce the trade in Rhino horn. It remains to be seen which way will be more successful. The old one certainly did not work. While it is to be hoped a last ditch effort, the Global Tiger Initiative (GTI) will halt the clicking extinction clock for wild tiger, China’s captive breeding program can be viewed as the most organised effort, an insurance of sorts, to at least preserve the species,-if in captivity.

Wildlife in the Neotropics: Between Overhunting, Habitat Destruction, and Enlightened New Policies Latin America, South America, or the Neotropics encompasses a vast new world with unique wildlife, which has, after many continuing struggles, at times between indigenous empires against the Spanish or Portuguese invaders, and between US-type capitalism and Soviet communism, emerged as something of a truly New World. Although old regimes linger and indigenous people remain disadvantaged, there is progress in social conditions and in environmental management which is uniquely Latin American in its boldness and innovation. This is no more evident than in Brazil, a G20 nation now which has started to tackle its Amazon frontier with its vast environmental challenges with a sense of purpose, technological innovation, and the support of a passionate, growing number of environmental/ social leaders, not a few of them indigenous people. While this development is

Wildlife Management in the Tropics: An Overview

2165

Fig. 3 There are few if any other landscapes in the world which combine magnificent scenery with magnificent wildlife and culture to such an extent as the northern parts of the Indian subcontinent bordering the Himalaya’s (Bhutan, India, Nepal). The megafauna of this region (One-horned rhino, tiger, Asian elephant, Gangetic dolphin, gharial, gaur etc.) is now more or less confined to increasingly isolated protected areas. Small wildlife populations, after having been greatly affected and degraded by mega-hydro-development (Kosi Tappu, Sukhlaphanta, Bardia, Chitwan) now have to cope with growing tourism numbers, an increasing part coming from Asia, in particular India. Both short and long-term impacts are difficult to evaluate. On the positive side many communities around the parks, which in parts depend on their resources (fish in boundary rivers, thatch grass, firewood) adding to the pressure, now derive significant income from tourism (Bauer and Maskey 1990; Bauer et al. 1995, 1997, 1999)

possible through a growing educated middle class it is also based on a wealth of natural resources. For this reason the pressures to convert vast areas of rainforest into “more accessible” wealth (logging, cattle farming, soya beans, etc.) continue unabated, and successes remain frail. The 70 % reduction of clearing rates in the Amazon since 2004, the outcome of government efforts, and REDD expectations can easily turn to dust as the greed for land and minerals around the world grows unabated. As in Africa, wildlife’s greatest protection so far was the vastness of the land, the impenetrability of its forests and wetlands, and the lack of access through roads. This is, however, rapidly changing as roads and airstrips open up the endless wetlands of the Pantanal and Amazon and as growing numbers of settlers, fortune hunters, cattle ranchers, and mineral prospectors, often in the wake of logging and

2166

J. Bauer

mineral extraction, invade the rainforest and the equally vast tropical dryland forests (Cerrado, Caatinga). This situation is replicated in places such as Peru, Columbia, Ecuador, Venezuela, Chile, or Bolivia. A somewhat dated review of its use (Robinson and Redford 1991) showed that wildlife in South America, although greatly depleted in the past and still harvested mostly without much regulation, remains an important resource. This review distinguished five types of uses people derive from wildlife: subsistence hunting (1), market hunting and collecting (2), wildlife farming and ranching (3), sport hunting (4), and commercial uses which included trade and tourism (5). Wildlife used is many species of mammals, birds, reptiles, amphibians, fish, and thousands of species of plants. Despite the largescale tropical deforestation in this region, there is a great dependency on wildlife products, particularly among indigenous people and peasants (Bodmer 1995; Bodmer et al. 1997; Robinson and Redford 1991; Vickers 1991). At the same time there are many possibilities to develop this industry for value adding consumptive and nonconsumptive tourism (Dallmeier 1991; Groom et al. 1991; Purdy and Tomlinson 1991). Many of those, however, remain unrealized. Considers the implementation of more regulation and sustainable hunting in South America generally possible, however, only if the most common forms of market or commercial hunting can be eliminated. This is a problem in his eyes not so much of morality but of sustainability, closely linked to a regard of wildlife as “public property” or “commons” to be exploited for individual financial gain and as market demands increase and new technologies become available (e.g., also Hardin 1968). Latin American people who hunt for wildlife (primates, game birds, tapirs, spectacled caiman, green iguana, capybara, deer sp., guanaco, vicunja, waterbirds, etc.) have generally few other alternatives (Ojeda and Mares 1982). Generally speaking wildlife and nature-based tourism as well as wildlife trade have continued to grow while subsistence uses and markets were closely linked to the development of legislation, the trends of populations, and the fates of the communities, often indigenous, themselves. There are also uses, such as Capybara hunting, which have made it to the urban centers, with its own set of problems – and opportunities. The Capybara as a Modern and Competitive Resource?

Capybara, a much priced meat for Venezuela’s indigenous people, also made its way, for keeps so it seems, into the cuisine of the Spanish invaders. Catholic missionary monks soon realised that the Capybary not only were tasty, but allowed them to retain their meat diet on Friday (which back home in Europe was provided by the Beaver which they classified along with fish), it was also much bigger than beaver, reaching weights of up to 66 kg meat and valuable leather led to over-harvest and a greatly diminished resource. Venezuela responded in 1953 when it made Capybara subject to legal regulation which was little effective until 1968, when a 5-year national moratorium was declared, the species studied and a management plan developed. (continued)

Wildlife Management in the Tropics: An Overview

2167

Once populations had recovered, 35–40% of the censused animals could be harvested every year on licensed farms with populations exceeding 400 animals. Research had showed that in Venezuela on irrigated savannah optimal density (1.5–3 animals/ha) yielding some 27 kg meat/ha/annum. The yield of unmanaged wild population compared with around 8 kg/ha. Most importantly, further studies have also shown that capybaras do not, as previously thought, compete with domestic stock, but graze on short vegetation providing additional economic benefits to farmers. As it seems this trend in Venezuela has not continued. Will Grant, from BBC news (12.4.2009, http://news. bbc.co.uk/2/hi/americas/7987587.stm) in an article titled “Venezuela’s Giant Rodent Cuisine” reported that while the capybaras’ attraction as Christian “lent food” remains very alive (“many Venezuelans regard the semi-aquatic creature as more fish than meat – a useful description during Lent when it is eaten as a replacement for red meat in this largely Roman Catholic country”), legislation has started to lag behind, The high demand in the run-up to Easter, combined with widespread poaching and illegal hunting, means the “chiguire”, as it is called in Venezuela, is now under threat in some parts of the country. This trend, according to Deborah Bigio of FUDENA, an environmental NGO cited in that article, shows little sign of slowing down in the wake of tighter hunting legislation (special hunting permit in the month before Easter) as it remains poorly enforced in a poorly educated community of hunters. Capybara are listed by IUCN at lower risk (management dependant) yet seems to continue to decline due to lack of existing legislation or enforcement.

Nature- and wildlife-based tourism is proving a double-edged entry point into a more modern way of life, as desired by many people. Countries like Costa Rica have managed to develop this income source for rural communities through the proliferation of a national and international NGO (INGO) culture. As we can see this commercial and market-driven approach closely combined with tourism may hold some promise. Free Market Wildlife Conservation in Costa Rica?

Costa Rica may serve as an example where conditions have been created by the government that have encouraged foreign “conservation investment” by a wide range of INGO’s, supported by education and research programs from universities around the world and in particular from the US. In this policy environment, conservation has become an experimental ground for the worlds INGO’s to try out new concepts and ideas at a safe place where they could be reasonably sure that a government and society would support (continued)

2168

J. Bauer

it. Although the eventual outcome of that “foreign intervention” in matters W&BDM remains unclear (many land right issues remain or have been exacerbated), it emerges that such project activity, is generally supported by growing tourist numbers and that much of that tourism growth is based on small and community ecotourism enterprises. One might conclude that this diverse environment of free foreign access, mutual competition (raising standards building knowledge, capacity and networks) between a competing but also collaborating INGO community can do much to facilitate between government and tourism and communities. It has also already proven that it provides substantial income alternatives and complementary landuse options including REDD+ etc. for the disadvantaged rural and indigenous communities.

Oceanic Tropical Islands: Extinction Traps and Wildlife Sanctuaries Islands are unique places in the world where mostly birds became often large and helpless because of size. Humans arriving thought they had entered paradise, for a while at least. After the dodos, moas, Hawaiian geese, giant lemurs, and so on were driven to extinction, most tropical islands around the world were changed forever and in historic times. They offered a compressed and much clearer, more recent history of the relationship of humans with large mammals than on the continental world. This was before the European colonizers and their animal companions (goats, sheep, deer, rats, mustelids, foxes, cats) arrived. Then it was the turn of what had survived, smaller, less valuable, but still abundant like the Caribbean seal, which finally, and after the last sighting in 1952, was officially’ declared extinct in 2008. When everybody woke up, it seemed too late. Unique tropical island worlds, if diminished already by earlier invaders, had been changed to places where the native and endemic was restricted to some hilltops like in Hawaii or small offshore islands (like in new Zealand), while the remainder was covered with human crops, their foreign animals, and plants, which thrived in their new world, often replacing endemic species. Looking at the history of bird extinctions over the past 500 years one can easily see that the recent extinction wave started on small and larger islands in the 1600s. Continents only joined in after the 1800s. When comparing timing and geography of extinction events with the colonial expansion of Europe a pattern emerges, which suggests that many extinctions were related in space and time to the European fleets colonizing the world, taking over continents, and introducing European land use and which, more often than not, were little suited to the local conditions (DiCastri 1989; Flannery 1994, 2001; Diamond 1999; FernandezArmesto 2000). Between 1630 and 1999 around 117 species of birds went extinct in the world (based on assessments by Birdlife International and recorded by WCMC) during

Wildlife Management in the Tropics: An Overview

2169

two major extinction phases, the first between 1600 and 1700 centered on the Caribbean, the second, more pronounced, longer, and more consistent, between 1700 and 1950 centered in the vast Pacific region. Both waves coincided with the history of European expansion, and both focused on tropical islands, where almost all extinctions took place. A modern third phase, despite our modern vigilance, continues. One of the best-studied examples is the brown tree snake from Queensland, Australia, which, after having been introduced to Guam after WWII, almost eradicated the endemic bird fauna, some ten species, within 20 years (Meffe and Carrol 1994). Yet, despite many similar stories, mostly unrecorded, there is another side to islands. Some 5,000 km to the south, bird species extinct on the mainland in NZ managed to retain a last precarious hold and more importantly develop viable, if density-dependent, populations. These were assisted by wildlife management programs, which were daring and innovative yet proved successful. Similar things happen on Mauritius (pink pigeon, kestrel, parrot) and Galapagos, where collaborating national and international organizations have managed to arrest and even turn the extinction tide. The Island continent of Australia with its tropical upper third, its highly endemic island fauna, and its history of the introduction of large alien tropical mammal species (banteng, water buffalo, rusa deer, sambar deer, wild boar) but also amphibians (cane toad) may serve as continental case study on fauna change. They demonstrate how even a large island continent seems destined to become a new and composite ecosystem (native and alien) where the endemic and the native can only survive if supported by modern, well-funded, and, above everything else, consistent efforts in wildlife management. Australian SoE reports (1996–2011), e. g., give for the first time a relatively comprehensive threat assessment of Australian fauna and flora (State of the Environment 2011). If we eliminate trend estimates for flora and invertebrates (as currently too patchy and unreliable) it is clear that there is a reduction trend among vertebrates which does not simply reflect taxonomic and knowledge uncertainty. It seems that during this brief reporting period native amphibians and reptiles have joined birds and mammals in their extinction trends.

The Great Unknown: Continental Trends of Australia’s Bird Fauna

Australia is an object lesson of the poor predictability of the combined, longterm changes of highly endemic fauna as is generally found on islands. Currently, due to the large numbers of bird observers contributing to distribution lists and participating at large scale surveys and monitoring systems our understanding of trends in bird populations is by far the most advanced one. In Australia a forecast based on large datasets is worrying as a whole (Garnett and Crowley 2000). According to the 2000 Action Plan for Australian birds, there are 25 bird taxa (reporting to the subspecies level) extinct, 32 critically endangered, 41 endangered, 82 vulnerable and 81 near (continued)

2170

J. Bauer

threatened. The remaining 1,114 taxa are considered of least concern, including 28 introduced taxa and 95 vagrants. If this assessment is being compared with 1992, the time of the last comprehensive bird count, a downward trend is evident which is only partly offset by successes in the rehabilitation of species. During the time frame of the count seven taxa could be downgraded as a result of effective conservation management (2 from CE-EN, 4 from EN-VU and 1 from EN-NT). Conservation efforts at present are therefore not able to keep up with current downward trends. So far this trend has been most pronounced in the south west and south east where a combination of sheep grazing and grain growing has devastated entire landscapes and greatly impoverished regional mammal and reptile faunas (Goldney and Bauer 1998; Bauer and Goldney 2000). The Australian tropics are now the new Australian development frontier where the re-development of pastoralism, large scale mining and new tropical crops (with their water demands) combine with the impacts of long established (Water buffalo, Dromedaries, Feral horses and donkeys) and new arrivals (Cane toads in NT) species of exotic origin. This will be further exacerbated by the rising of sea levels which will change the very nature of many coastal wetland systems.

Conclusions When we talk about wildlife in the tropics Dasmann’s premise lurks in the back of our minds. Terms such as “the developing world,” disadvantage, and poverty come to mind. There remains a general “feeling,” akin to Dasmann (1964), that communities of animals are perhaps more important but also less “managed,” if managed at all, and certainly less “scientifically.” There is also the growing number of wildlife populations in the tropics which the western mind does not want to “manage” but confine to the national parks it has created there, accessible to the “tourist’s gaze” but not so much for the local to hunt, let alone eat. And there is that “megafauna,” mostly gone from our temperate and developed world as “highly inconvenient” for our agriculture and forestry but one we want to keep in the tropics, if only for the gaze of a discerning tourism industry. This chapter, drawn with a coarse brush, has shown that across the tropical world wildlife and its use have undergone dramatic changes. It continues to play a crucially important role, in particular for poor and disadvantaged people, for many of which, however, it has become inaccessible. Things are changing though. There are now examples in Africa where a colonial and postcolonial legacy of western-style livestock husbandry and wildlife protection has given way to one where indigenous communities have recovered what used to be their old wildlife heritage while having gained the confidence to embark in new or banished ones such as wildlife tourism and trophy hunting and fishing. Tanzania shows, however, that this is a fickle path as the state wants its share.

Wildlife Management in the Tropics: An Overview

2171

But there also are places like Sarawak, where a strong indigenous element (the Dajak people) was able to retain the use of wildlife as an important element in modern life and diet, despite the impacts of ongoing adverse logging. And there are places such as India, Nepal, Bhutan, Thailand, Vietnam, or many parts of China where the use of wildlife, including tiger bones, is all but illegal yet continues unabated and with perhaps different models where the influence of western attitudes and approaches from the 1960s lingers and remains in direct and often devastating contrast to rural realities where sustainable use by communities is made impossible because the illegal one thrives. This is also the reason why capybara ranching in Venezuela has not happened as what seemed to be a likely future scenario some 15 years ago. If we would try to summarize the negative and positive trends in wildlife populations and their use in the tropics we might identify the following general trends:

Status of Wildlife (a) Loss of tropical vegetation, in particular rainforest, continues unabated (despite half a century of efforts), and logging has depleted or destroyed tropical forests and their wildlife in many regions. While efforts also as part of REDD have increased, in particular in rainforest regions, they have yet to make a real impact. (b) Many tropical wetland systems have been affected/changed by megahydrodevelopment for energy, irrigation, urban infrastructure, and transport, greatly affecting their population of animals and plants with larger species (river dolphins, large fish, crocodilians) often particularly affected. This process (e.g., Pantanal in South America) continues to accelerate, and there are vast areas of coastal wetlands which will be increasingly affected (and changed) by rising sea levels. (c) There has been an emergence of new threats (soybeans, oil palm) which are closely linked to unabated growth of populations and consumerism. There is only little and very limited understanding on how land grabs and industrial agriculture (chemicals, GM, low labor demand) will further drive this change. (d) Megafauna around the world has reached a critical point where a combination of illegal trade (because of their often high value), loss of habitat, and conflict with rural communities have continued the dramatic losses in the past. Many populations have become too small and too isolated to be sustainable and will require more efforts and new approaches to be maintained. This even applies to Africa’s two species of elephant which have lost more than half of their population numbers over the past 20 years. (e) A similar situation has emerged for the world’s >300 species of primates of which around two-thirds are now endangered, often in small isolated populations. They continue to be under immense pressure from hunting and habitat loss. Although many international efforts are underway, generally this situation continues to worsen, in particular in SE Asia, where many primate populations have become unviable. This situation is even more pronounced for

2172

J. Bauer

remaining megafauna (e.g., wild cattle) and starts to affect a growing number of species of wildlife. (f) There has been an acceleration of the rate of spread of alien organisms with few successes to contain them in tropical countries, although the losses in agriculture (see Africa report) are undeniable. Although there have been many efforts to better understand and monitor the global megatrend of alienation of fauna and flora, action on ground is very limited, inconsistent, and largely ineffective even in places such as Australia.

The Status of Wildlife Harvest and Hunting/Fishing There has been a collapse of many traditional hunting/fishing systems partly because of loss of prey, of breakdown of rural traditions, or of legislation, in particular the establishment of protected areas without benefits for the local societies. This process has coincided and often been responsible for the growth of illegal trade and practices, which finds it easier to operate in this new legislative environment, leading to the loss of community practices and control. These practices are greatly facilitated with modern technology. The trade in wildlife emanating from urban centers of affluence has dramatically increased over the past 20 years and across the tropics. Many of the traders operate now globally, and the trade is closely linked with drugs, arms, and often conflict. Despite increasing efforts by the international community but also national governments, this trend continues. The corporatization of valuable hunting/fishing/wildlife resources has proceeded in many regions and often with support by the state. This might be the depredations of the European fishing fleet along Africa’s fish-rich east coast, foreign fishing vessels along the coasts of SE Asia, foreign logging companies (which apart from forest destruction greatly facilitate the exploitation of wildlife), or in a wider sense the globalization of all forms of (generally poorly regulated) wildlife-related tourism including hunting and fishing tourism. Tourism targeting nature and wildlife has emerged as a great player, seemingly justifying countries’ investments in wildlife and protected areas. As the tourism industry as a whole remains excluded (partly by choice, also by regulation) in matters of wildlife management nor shows generally any great commitment and interest to participate, its potential as a major positive force for wildlife and biodiversity remains mostly unrealized.

References Baker JE (1997) Trophy hunting as a sustainable use of wildlife resources in southern and eastern Africa. J Sustain Tour 4:306–321 Baskin Y (1994) Wildlife conservation – there’s a new wildlife policy in Kenya – use it or lose it. Science 265:733–734

Wildlife Management in the Tropics: An Overview

2173

Bauer JJ (2002) Development of a National Strategy for the Management of the Wild Boar-Farmer Conflict in Bhutan. Report, Ministry of Agriculture, Thimphu, Bhutan Bauer JJ, Goldney D (2000) Extinction processes in a transitional agricultural landscape. In: Hobbs RJ, Yates CJ (eds) Temperate eucalypt woodlands in Australia, biology, conservation, management and restoration. Surrey Beatty and Sons, Chipping Norton NSW, Australia Bauer J, Herr J (2004) Hunting and fishing tourism. In: Higginbottom K (ed) Wildlife tourism. Common Ground Publishing, Altona, Victoria, Australia Bauer JJ, Maskey TM and Rast G (1990) The impact of Karnali hydrodevelopment on the conservation potential of Royal Bardia Wildlife Reserve (RBWR) and other affected areas. Project document, WWF- International, Gland Switzerland, and WWF- Inst. for Floodplains ecology, Rastatt Bauer JJ, Maskey T, Rast G (1995) River systems, hydrodevelopment and the species crisis in the Terai. In: Bhandari T, Shresta TB, McEachern J (eds) Safeguarding Wetlands in Nepal. IUCNThe World Conservation Union. Heritage and Biodiversity Conservation Programme, Gland Bauer JJ, Maskey T, Rast G (1997) The environmental costs of river regulation in Nepal – present evidence and scenarios for the future. In: Proceedings of the international conference on Wetlands & Development, Selangor, 8–14 Oct 1995 Bauer JJ, Maskey T, Rast G, DeLacy T, Glazebrook H, Furze B (1999) The impact of mega hydrodevelopment on biodiversity conservation and community development in Nepal’s Terai- a Riverbasin perspective a case study from Nepal’s River Basins, UNEP/AWB. Johnstone Centre of Ecosystem Management, Kuala Lumpur/Nairobi/Kenya/Charles Sturt University/Albury. 85 pp Bauer JJ, Gadd L, Haohan W (2002) An analysis of the Wuyishan biosphere mammal fauna through a grad-sec-sampling technique. Cooperative Research Centre for Sustainable Tourism in collaboration with Chinese National Committee on MAN and BIOSPHERE, Bureau of Forestry, Environmental Protection Administration and Chinese Academy of Sciences, Charles Sturt University, Albury, Australia Bodmer RE (1995) Managing Amazonian wildlife: biological correlates of game choice by detribalized hunters. Ecol Appl 5:872–877. doi:10.2307/2269338 Bodmer RE, Eisenberg JF, Redford KH (1997) Hunting and the likelihood of extinction of Amazonian mammals. Conserv Biol 11(2):460–466 Boyd M, Bauer JJ, Ren Z, Haohan W, Gadd L, DeLacy T (2002) Traditional ecological knowledge (TEK) of wildlife: implications for conservation and development in Wuyishan nature reserve. Fujian Province the international program of the CRC for Sustainable Tourism, Griffith University, Green Globe Asia Pacific – Goldcoast, Australia Boyd M, Ren Z, DeLacy T, Bauer JJ (2003) An analysis of traditional knowledge on wildlife in Wuyishan Biosphere Reserve, Fujian Province, China, STCRC monograph series. STCRC, Griffith University, Goldcoast, Australia Caughley G (1977) Analysis of vertebrate populations. Wiley, London Dallmeier F (1991) Whistling-ducks as a manageable and sustainable resource in Venezuela: balancing economic costs and benefits. In: Robinson JG, Redford KH (eds) Neotropical wildlife use and conservation. University of Chicago Press, Chicago Damm GR (2002) The conservation game. Saving Africa’s biodiversity. Safari Club International African Chapter, Rivonia Damm GR (2006) Development of game prices in South Africa. Afr Indaba e-Newsl 4(3):22 Dasmann RF (1964) Wildlife biology. Wiley, New York Diamond (1999) Guns, Germs and Steel, the Fate of Human Societies.W.W.Norton & Company, Inc. New York DiCastri (1989) History of biological invasions with special emphasis on the old world. In: Biological Invasions: A Global Perspective by JA Drake et al.(eds.) Wiley, New York Eglert I (2002) Brazil: selling biodiversity with local livelihoods. In: O’Riordan T, Stoll-Kleemann S (eds) Biodiversity, sustainability and human communities. Cambridge University Press, Cambridge UK

2174

J. Bauer

FAO (1995) International expert consultation on non-wood forest products. Yogyakarta, 17–27 Jan 1995 Fernandez-Armesto (2000) Civilizations. Pan Macmillan. London,Basingstoke and Oxford UK Tim Flannery (1994) The Future Eaters: an ecological History of the australasian lands and people G Braziller, New York Tim Flannery (2002) The Eternal Frontier: an ecological history of North America and its peoples, Grove Press, New York Furze B, De Lacy T, Birckhead J (1996) Culture, conservation and biodiversity: the social dimensions of linking local level development and conservation through protected areas. Wiley, New York Garnett ST, Crowley GM (2000) The action plan for Australian Birds 2000. Environment Australia and Birds Australia, Canberra. http://www.environment.gov.au/biodiversity/threat ened/publications/action/birds2000/index.html. Accessed 11 June 2015 Global Tiger Initiative Secretariat (2012) Global Tiger Recovery Program Implementation Report 2012, Washington Goldney D, Bauer JJ (1998) Conservation in an agricultural landscape- fact or fiction. In: Pratley J, Candrel G (eds) Agriculture and the environmental imperative. CSIRO Publishers, Melbourne Groom MJ, Podolsky RD, Munn CA (1991) Tourism as a sustained use of wildlife: a case study of Madre de Dios, Southern Peru. In: Robinson JG, Redford KH (eds) Neotropical wildlife use and conservation. University of Chicago Press, Chicago Hardin G (1968) The tragedy of the commons. Science 162:1243–1248 IIED (1994) Whose eden? An overview of community approaches to wildlife management. International Institute for Environment and Development, London UK Lewis DM, Alpert P (1997) Trophy hunting and wildlife conservation in Zambia. Conserv Biol. doi:10.1046/j.1523-1739.1997.94389.x Meffe GK, Carrol CR (1994) Principles of conservation biology. Sinauer Associates, Michigan University Meier G (1989) Organisation und Wirtschaftlichkeit verschiedener Verfahren der Wildtiernutzung im s€udlichen Afrika. PhD thesis, Institut f€ ur Landwirtschaftliche Betriebslehre der Universita¨tHohenheim, Neuhofen Mills J, Jackson P (1994). In: Species in Danger Julie Gray (ed) Killed for a cure- a review of the worldwide trade in tigerbone. Species in danger series. TRAFFIC international Cambridge UK Minwary MY (2009) Politics of participatory wildlife management in Enduimet WMA, Tanzania. MSc thesis in development studies, Noragric, Norwegian University of Life Sciences (UMB) O’Riordan T (2002) Protecting beyond the protected. In: O’Riordan T, Stoll-Kleemann S (eds) Biodiversity, sustainability and human communities. Cambridge University Press, Cambridge Ojeda RA, Mares MA (1982) Conservation of South American mammals: Argentina as a paradigm. In: Mares MA, Genoways HH (eds) Mammalian biology in South America, vol 6, Special publication. Pymatuning Laboratory of Ecology, Linesville Purdy PC, Tomlinson RE (1991) The eastern white-winged dove: factors influencing use and continuity of the resource. In: Robinson JG, Redford KH (eds) Neotropical wildlife use and conservation. University of Chicago Press, Chicago Ricklefs ER, Renner SS (1994) Species richness within families of flowering plants. Evolution 48:1619–1636 Robinson JG, Redford KH (eds) (1991) Neotropical wildlife use and conservation. University of Chicago Press, Chicago Save the Rhino (2015) The rhino poaching crisis: a market analysis. http://savetherhinotrust.org/ programmes/84-the-rhino-poaching-crisis-a-market-analysis. Accessed 11 June 2015 Shackley M (1995) The future of Gorilla tourism in Rwanda. J Sustain Tour 3(2):1 Sinclair ARE (1977) The African buffalo. University of Chicago Press, Chicago Spencely A, Habyalimana S, Tusabe R (2010) Benefits to the poor from gorilla tourism in Rwanda. Dev South Afr. doi:10.1080/0376835X.2010.522828

Wildlife Management in the Tropics: An Overview

2175

State of the Environment (2011) Committee. Australia state of the environment 2011. Independent report to the Australian Government Minister for Sustainability, Environment, Water, Population and Communities. DSEWPaC, Canberra State of the Tropics (n.d.) Primary forests. http://stateofthetropics.org/wp-content/uploads/Pri mary-Forests_English2.pdf. Accessed 11 June 2015 Tepper R (2013) Tiger bone wine trade reveals China’s two-faced approach to conservancy (NSFW). http://www.huffingtonpost.com/2013/02/28/tiger-bone-wine-china_n_2782772.html. Accessed 11 June 2015 Vickers W (1991) Ten years in an Amazon Indian territory. In: Robinson JG, Redford KH (eds) Neotropical wildlife use and conservation. University of Chicago Press, Chicago

How Environmental and Societal Changes Affect Wildlife in the Tropics Johannes Bauer

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Population Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Poverty and Affluence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Impacts of Deforestation and Forest Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pastoralism and Tropical Desertification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wildlife and the Pillage of Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inland Water Development, Degradation and Wildlife . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Spread of Alien Plants and Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Urbanisation and the Growth of the Urban Sprawl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mining and Oil Exploration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Oil Exploration and Rainforests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Caught in Between: Wildlife Between Political Instability and Warfare . . . . . . . . . . . . . . . . . . . . Environmental Chemicals and Wildlife: The Big Unknown? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Neonicotinoids as an Emerging Threat to Beneficial Insects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Changes and Trends in Human Wildlife Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Consumptive and Non-consumptive Wildlife Tourism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wildlife Trade and Wildlife Crime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Human-Wildlife Conflict . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Climate Change in the Tropics as It Affects Wildlife . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Projections for Mexican Faunas Under Global Climate Change Scenarios . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2178 2178 2178 2179 2179 2180 2181 2181 2183 2183 2184 2185 2185 2185 2186 2187 2187 2189 2189 2191 2192 2192

Keywords

Affluence and wildlife conservation • Climate change and wildlife & biodiversity conservation • Impacts of Deforestation and forest degradation • Impacts of Human population growth on wildlife • Human-wildlife conflict 9 • Degradation J. Bauer (*) Australian Carbon Co-operative Ltd., Bathurst, NSW, Australia e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_173

2177

2178

J. Bauer

of freshwater systems • Invasive species • Environmental Impacts of Neonicotinoids • Mining and Oil exploration Impacts • Wildlife and biodiversity impacts of Pastoralism • Pillage of agriculture 3 • Poverty and wildlife • Wildlife and Biodiversity Impacts of the Tourism industry • Impacts of Urbanisation on wildlife • Impacts of Warfare on wildlife • Wildlife trade

Introduction The tropics are affected by a growing range of developments, which replicate what has been happening in developed nations over centuries within decades, even years at times, and at much larger scales. There are also newly emerging threats such as GM, mining and explorations at gigantic scales, mega-hydrodevelopment, and a host of new chemicals, some of them, like neonicotinoids, with systemic, longterm, and unknown impacts. An unprecedented assault takes place on ecosystems and their species through intensifying land uses to feed a growing human population, which consumes more and more per capita. “Super crops” such as oil palm and soy bean, increasingly genetically modified, to provide fuel, food, and fiber, replace natural forests and ecosystems at great and growing scales resulting in shrinking wildlife habitats and populations. And now, the threat of a changing climate has been added to all that.

Population Growth Over the past 21 years (since TFH 1st ed.) the human population has grown by more than 1000 million people; since Dasmann (1964) it has more than doubled. Most of this growth happened in the lesser developed world, much of that in the tropics. As land use intensity (including the use of wildlife) and land use change proceeds (deforestation, draining of wetlands, agricultural expansion, and urbanization) it becomes clear that this human population growth has been one of the major direct and indirect factors for a decrease in wildlife abundance and distribution. Nor have the hopes of population stabilization been realized. Newly revised estimates of human population growth suggest much higher than expected stabilization rates, and across-the-board declining efforts in population control suggest that high human population growth in the tropics will increase to add to the pressures on wildlife and biodiversity.

Poverty and Affluence Along with population growth adverse demographic shifts show no abatement or are increasing. There are two major social trends in the growth of the human population in the tropics, which affect wildlife greatly:

How Environmental and Societal Changes Affect Wildlife in the Tropics

2179

– The growth of poverty which is in our modern world now most pronounced in poorly developed regions (such as the tropics) which also happen to be wildlife rich – The growth in affluence which feeds in the west the tourism industry and the pet market, in the east the markets for Traditional Asian Medicine and food. Both have become major factors for the survival of many species. Each of them have had its own specific impacts, many of which are deeply engrained within traditions and culture (for example the Asian Traditional Food and Medicine systems). A large part of TCM dating back more than 5000 years, is based on plant and animal products. The use of wildlife for these medicines has long been recognised to provide a large drain on often already declining populations of wildlife (e.g., Mainka et al. 1995). They have historically affected the abundance of populations of tigers, bears, rhino’s and swiftlets and countless other species. They also have, since the Asian economic boom, greatly expanded. Both have added additional pressure to many already declining populations of wildlife.

The Impacts of Deforestation and Forest Degradation With tropical forests containing a large part of the world’s biodiversity and wildlife forest loss has great impacts on wildlife populations. Sarawak is an example how rare intact forest has become in many places throughout Asia (Fig. 1). This forest loss, mostly recorded where logging destroys valuable lowland/high rainfall forest, needs to be expanded. Much deforestation occurs in the tropical dry and mountain forest, often gradually and through the intermediate stages of degradation and it has received much less attention from the public, science and policy makers. “Scrub” here in Australia, “Jungle” in India, South American Cerrado or Caatinga, the many and vast areas of woodlands (e.g., Miombo woodlands) in Africa have long been affected by pastoralism, subsistence agriculture, charcoal production etc. with equally devastating effect including widespread desertification.

Pastoralism and Tropical Desertification With the “saturation” of many grassy landscapes (steppes, savannah, woodlands) with domestic ungulates, much of the current growth of the world’s livestock in particular cattle now happens at the expense of tropical rainforests. This has happened most dramatically in the Amazon where vast forest landscapes have been converted into cattle pasture along with soya fields, corn fields or even Eucalypt plantations. Of all these new landuses on formerly rainforest land, cattle has taken the largest toll with an Amazonian cattle population which had grown from 5 million to 70–80 million heads by 2003 (Veiga et al. 2003) At that time 15 % of the Amazon forest had been replaced by agricultural land and around 80 % of the

2180

J. Bauer

Border Non-forest Cloud/water Intact forest (not roaded) Degraded forest (roaded once) Severely degraded forest (roaded 2-7 times) Plantation/regrowth Mangrove Indonesia

0 40 80

160

240

320 Kilometers

N

A recent study estimated that by 2009 only 57% of Sarawak was covered in forest and that, conservatively, at least two-thirds of this was degraded or severely degraded by logging. An analysis by Global Witness based on satellite imagery from 2013 (Bryan et al. 2013) suggests that today only 5% of Sarawak is covered by intact forest. (The heavy grey line was added to show Sarawak’s border). Sarawak’s Wildlifedependent forest communities (Caldecott 1988; Horowitz 1998) are greatly affected as wildlife in logged forest has declined significantly).

Fig. 1 Extreme differences in forest degradation in Borneo (Bryan et al. 2013)

deforested areas is now covered by pastures (approximately 900,000 km2). This loss of natural land in Brazil is not restricted to the Amazon but progresses in its Caatinga and Cerrado regions where destruction of forest and woodland ecosystems happens at higher levels now as in the Amazon, albeit poorly reported. There are also growing concerns that changes in the Amazon reduce rainfall in these drier regions (“rivers in the sky”) with devastating impacts not only on southern Brazil’s domestic water supply, but wildlife. This process is not restricted to the Amazon. It is also happening in many other parts of the tropics and subtropics where pastoralism has led to clearing and subsequent landscape degradation and wildlife loss, which is similar to parts of Africa’s Sahel zone (Goldney and Bauer 1998; Bauer and Goldney 2000). Studies by Goldney et al. (1995), Bauer et al. 2002a, b; and Date et al. (2000) showed that during that change in Australia’s oldest pastoral landscape, one third of the mammal fauna has been lost during the first 50 years with another third being now on the verge of regional extinction. Similar direct and indirect extinction processes through pastoralism have been reported from many other regions around the world including Inner Mongolia (Thwaites et al. 1996, 2000) and happen in the Cerrado-Caatinga regions of Brazil, Mexican drylands, India, Africa and Madagascar.

Wildlife and the Pillage of Agriculture Mazoyer and Roudart’s (2006) “History of World Agriculture” starts off with the emergence of agriculture from the long human past of hunting and gathering and how it has developed its techniques in different regions of the world. These conditions have now led to what they perceive as the “pillage of agriculture in developing nations,” caused by a wide range of social conditions, international trade practices (neoliberal free trade), aid strategies, accepted development models, even research strategies, which have favoured the have’s and all but destroyed the have-nots (they estimate two billion farmers). Recently a new dimension to that has been added as an increasing number of companies and nations even are buying up

How Environmental and Societal Changes Affect Wildlife in the Tropics

2181

land in the tropics, from Ethiopia to Madagascar or Brazil and Northern Australia buy land at large scales. Such “land grabs” for land conversion into high tech (mechanical, chemical, GM) agriculture are an expanding practice. It will lead to further landuse conversion and reduce the suitability of traditionally diverse agricultural land for wildlife.

Inland Water Development, Degradation and Wildlife Inland (mostly freshwater) water systems, although comprising less than 1 % of the world’s water resources, contain a disproportionately high fraction of the world’s wildlife and play a crucial role for the health of terrestrial (and many marine) ecosystems and human society (agriculture, inland traffic, hydropower, fish resources etc. etc.). Pressures acting on them are many, the main ones being hydro-development, extraction, land reclamation (drainage), overfishing and overhunting (waterfowl), pollution and invasion by exotic species (e.g., Dugan 1990; Wescoat Jr and White 2003; Bauer 1993; Bauer et. al. 1995). Now there are indications that overall, freshwater species are declining faster than terrestrial or marine species (Groombridge and Jenkins 2000). There is also more and more evidence that such systems lose their many environmental functions, leading to a general “drying up” of entire landscapes, often as part of desertification (Bauer and Goldney 2000). Now, some 80 % of the world’s rivers are dammed, many lakes are drained or greatly polluted, in others dams have been “installed,” interrupting diverse and highly productive river ecologies. There has been the evolution of vast irrigation landscapes, canals have connected formerly separated systems and a great number of introductions of exotic organisms have occurred. River flows have been profoundly altered, and the demands of ever bigger cities have removed more and more water, while natural replenishment, and water quality generally declined. There are also now aquifers, huge sub-terrestrial water reservoirs which have been all but drained for agriculture and which cannot refill any longer because we remove water before it can recharge them. And above all, more and more of the water has become saline, because of irrigation and the pumping of water. The impacts of such “management schemes” have been many, poorly understood or not recognised or recorded and were profound beyond measure. Nor are there any intentions to learn from past mistakes. If we look at Finer and Jenkins (2012) study in one of the richest wildlife hotspots and ecologically fragile regions of the world, the western Amazon region, it is clear that many developments have hardly started (Fig. 2).

The Spread of Alien Plants and Animals Invasions of tropical ecosystems by alien and “invasive” plants and animals have been widespread and poorly recorded, let alone contained. The promotion of acclimatisation of exotic organisms by western governments and science, the

2182

J. Bauer ...”However, the current planning for hydropower lacks adequate regional and basin-scale assessment of potential ecological impacts. This lack of strategic planning is particularly problematic given the intimate link between the Andes and Amazonian flood plain, together one of the most species rich zones on Earth. We examined the potential ecological impacts, in terms of river connectivity and forest loss, of the planned proliferation of hydroelectric dams across all Andean tributaries of the Amazon River… The ecological impact analysis classified 71 (47%) of the planned dams as high impact, 51 (34%) as moderate impact, and 29 (19%) as low impact. Considering the individual factors, 82% of new dams would represent a high or moderate fragmentation event, while 60% would cause the first major break in connectivity between protected Andean headwaters and the lowland Amazon. Deforestation would be a major issue for many dams, with 36% requiring new roads and 79% needing new transmission line routes. Eleven dams would directly impact a protected area (Finer and Jenkins 2012)

Fig. 2 Proliferation of hydroelectric dams in the Andean Amazon and implications for AndesAmazon connectivity (Finer and Jenkins (2012))

removal of trade barriers, the spread of crop species in forestry (Eucalypt sp.), agriculture (oil palm, soy etc.), and fisheries (Tilapia) each with their associated species (weeds, parasites etc.) in the tropics, has more than offset the improved controls in and through the west. There is a huge and growing body of literature on the evidence and impacts of alien fauna or flora in North America, Australia, New Zealand and South Africa, all former British colonies. Much less is known (and done about) in countries such as China for example where organisms are classified according to utility, not provenance. Nor have the west and science developed lasting solutions. While databases and information systems on alien and invasive organisms grow, attempts to contain them, in the few places where this is done (for example cane toad in tropical Australia) remain underfunded, inconsistent and largely ineffective. Although there are noteworthy exceptions in this general trend they are too far and between to interrupt what has become an exponential process. In the great majority of systems, unlike Kruger National Park, where park administration tries to unsuccessfully manage the invasion of the parks environment with countless plants (as they are dispersed by tourists) invasion continues unabated. This is not to say they have no impacts. Obiri (2008) analysing invasive plant species in Kenya and Tanzania demonstrated their impacts on livelihood systems. There is much less known how they affect ecology and wildlife. Invasive plant species are hazards that have shown negative environmental and socioeconomic impacts in East African drylands. They have degraded the environment and led to

How Environmental and Societal Changes Affect Wildlife in the Tropics

2183

serious impacts on human wellbeing such as reduced availability of goods and services for local communities, increased spread of diseases and reduced economic opportunities. This, in turn, has led to loss of livelihoods, and reduced food security. Among the most serious of cases is the Mesquite tree that has devastated social livelihoods of many dryland communities in Kenya and even led to constitutional court cases between local communities and the Kenyan government. In both Kenya and Tanzania key legislations (such as the EMCA, the Forest and Plant Protection Acts) and institutions such as NEMA and the MENR monitor and control invasive species, however their outcomes have not been successful. The invasive plants related disasters have risen as communities have progressively moved into the drylands and remained ill-prepared to cope with the hazards. For instance, in the Baringo area the population was 210,000 when the mesquite was introduced around 1986 but by 2006 it had risen to 540 000, meaning that more people were exposed to hazards and thus disasters were likely to occur

Urbanisation and the Growth of the Urban Sprawl Population projections suggest that by 2030 the urban population will have increased to almost 5 billion and that urban land cover will increase by 1.2 million square kilometers, nearly tripling the global urban land area circa 2000 (Seto et al. 2012). As many of the worlds cities are located on highly productive and diverse landscapes, Seto et al. (2012) suggest that under current trends a considerable loss of habitats in key biodiversity hotspots [will occur] with the highest rates of forecasted urban growth to take place in tropical regions that were relatively undisturbed by urban development in 2000: the Eastern Afro-montane, the Guinean Forests of West Africa, the Western Ghats of India and Sri Lanka. They further estimate that “within the pan-tropics, loss in vegetation biomass from areas with high probability of urban expansion is estimated to be 1.38 PgC (0.05 PgC year1), equal to 5 % of emissions from tropical deforestation and land-use change.” They conclude that only far-reaching policy changes to affect future growth trajectories could minimize global biodiversity and vegetation carbon losses’ by urbanisation. Studies like Aronson et al. (2014) suggest that the loss of plant and bird diversity during urban expansion amounts to some 90 %.

Mining and Oil Exploration As most of my colleagues I have had some experience with the impacts of mining on wildlife and biodiversity and the ways industry, policy makers and regulators, as well as scientists have tried to mitigate its impacts. Depending on the type of operation impacts can be localised and widespread, short-term and long-term, devastating or seemingly minor, hard to detect and highly controversial. I am talking here about Australia which prides itself on a highly regulated mining environment and where EIA has become the employment of an increasing number of wildlife biologists also. Few of these impact assessments will save the habitat of an endangered species, in particular now with the possibility of “biodiversity

2184

J. Bauer

offsets” where an endangered species occurrence on doomed “overburden” (the miners expression for forest ecosystems) as a mine development hindrance, can be simply swapped with another one where the species occurs and the mine assumes some’ responsibility with the magic wave of the legal wand. In most regions even such deeply flawed impact procedures are either not there or applied as for example described by Butler (on his website www.mongabay.com) or most recently in an article in Le Monde Diplomatique “Dirty water, dirtier practices” (LMD 2014) where HC Ospina describes how one of the world’s oil giants, destroys wildlife and communities across huge land areas in Ecuador with breathtaking indifference.

Oil Exploration and Rainforests With some of the world’s most promising reserves, exploration of oil in rainforests has become a major reality in many parts of South America and now Africa. Much of that exploration happens “under the radar” and it requires “watchdog organisations” to inform the world with what casual devastation forest communities and wildlife are treated. “The extraction of oil is now responsible for the deforestation, degradation, and environmental devastation of lands and communities across the globe. The oil extraction process results in the release of toxic drilling by-products into local rivers, while broken pipelines and leakage result in persistent oil spillage. In addition, the construction of roads for accessing remote oil sites opens remote lands to colonists and land developers. Some of the world’s most promising oil and gas deposits lie deep in tropical rainforests, especially in the Western Amazon. With oil at historically high prices, the incentive to develop oil resources has never been greater” (Butler 2012, Fig. 3).

A 2008 study published in Environmental Research Letters found that 41 percent of the Peruvian Amazon was covered by 52 active oil and gas concessions, nearly six times as much land as was covered in 2003 Oil and gas blocks in the western Amazon (Butler 2012)

Yellow indicates blocks already leased out to companies. Hashed yellow indicates proposed blocks or blocks still in the negotiation phase. Protected areas shown are those considered strictly protected by the IUCN (categories I to III). Oil and Gas pipelines

Fig. 3 Oil and gas projects in the Western Amazon: threats to wilderness, biodiversity, and indigenous peoples (Finer et al. 2008)

How Environmental and Societal Changes Affect Wildlife in the Tropics

2185

Caught in Between: Wildlife Between Political Instability and Warfare Wildlife can be negatively and positively affected by political borders and instability including warfare. At its most cynical, mine fields and danger zones, create no-man’s land where wildlife may thrive and the most impressive modern example is Chernobyl where the fascination of the return of wildlife in this no-man’s land has become subject to a global following. At its worst however, wildlife can be devastated by warfare. As happened in the US, and perpetrated again in Vietnam by this nation, it can be targeted to destroy a food resource for the enemy (Plains Buffalo in the US) or the means of transport (US targeting wild elephants in Vietnam). One can also destroy its habitat with herbicides (US employing Agent Orange in Vietnam), or draining it of water (Sadam Hussein draining the wetlands of Iraq to destroy Marsh Arab culture). In Africa the proliferation of automatic firearms, wars and political upheaval has affected and often devastated wildlife alongside humanity.

Environmental Chemicals and Wildlife: The Big Unknown? Ever since Rachel Carson’s inseminal book “Silent Spring” (1962) alerted her US country people to the rather horrific “side effects” of chemicals, the world has not stood still but added thousands of new ones with unknown effects to the worlds environments. While there are improved assessment procedures in place in some countries (not in the tropics generally), every textbook will tell us that these remain grossly simplistic and incapable of monitoring long-term effects within complex food chains and ecosystems. One of the new types of chemicals used, neonicotinoids are reviewed by “beyond pesticides” a North American watchdog organisation, as follows:

Neonicotinoids as an Emerging Threat to Beneficial Insects Neonicotinoids are a relatively new class of insecticides that share a common mode of action that affect the central nervous system of insects, resulting in paralysis and death (Beyond Pesticides nd). They include imidacloprid, acetamiprid, clothianidin, dinotefuran, nithiazine, thiacloprid and thiamethoxam. According to the EPA, uncertainties have been identified since their initial registration regarding the potential environmental fate and effects of neonicotinoid pesticides, particularly as they relate to pollinators. Studies conducted in the late 1990s suggest that neonicotinic residues can accumulate in pollen and nectar of treated plants and represent a potential risk to pollinators. There is major concern that neonicotinoid pesticides may play a role in recent pollinator declines. Neonicotinods can also be persistent in the environment, and when used as seed treatments, translocate to residues in pollen and nectar of treated plants. The potential for these residues to affect bees and other pollinators remain uncertain. Despite these uncertainties,

2186

J. Bauer

neonicotinoids are beginning to dominate the market place, putting pollinators at risk. As neonicotinoids have been linked to the collapse of honey bee colonies with lethal and sub-lethal effects threatening their crucially important role in agriculture, the European Union decided to ban their use in agriculture for two years. Late 2013 however, agrichemical giants Syngenta and Bayer announced that they would be suing the E.U. over its decision. There are similar concerns with others such as 2, 4-D Corn or soybean, where an Agent Orange related chemical is incorporated into a GM corn variant. There are also lessons from the lack of action for highly dangerous chemicals such as DDT, a pesticide used widely in the 1970s and 1980s (agricultural use was banned in most developed countries between 1968 and 1989) and the content of much of Rachel Carson’s book. It is still used in Disease Vector control to which it had been restricted by the Stockholm Convention of 2004. Ratified by more than 170 countries and endorsed by most environmental groups, this Convention recognises that total elimination in many malaria-prone countries is currently unfeasible because there are few affordable or effective alternatives. Public health use is exempt from the ban pending acceptable alternatives. Agricultural use of DDT continues in India, North Korea and possibly other countries. If one keeps in mind that DDT is not only toxic to a wide range of organisms but also leads to “Eggshell thinning” and magnified effects through bioaccumulation along the food chain towards apex predators (and is aware that many of these effects will be undiscovered in the tropics where few enforceable regulations exist), one can easily see the risks. There are two critical problems with current registration procedures and impact assessment methods for pesticides: the increasing reliance on industry-funded science and their dominance in even the review process of peer-reviewed studies and the inadequacy of current risk assessment procedures in particular to account for sub-lethal and accumulative effects of pesticides. Sub-lethal effects are rarely picked up in such studies although their ability to accumulate is deeply disturbing.

Changes and Trends in Human Wildlife Interactions There have been profound changes in how tropical societies interact with wildlife, what species and how many of those they can harvest and how those customs and landuse systems can survive in the modern world (Caldecott 1988; Robinson and Redford 1991; Horowitz 1998; Bauer 1993; Bauer and English 2011a, b). For once, many wildlife rich regions have become, legally at least, off limits to local communities (protected areas). At the same time the wildlife of these regions became the target of non –consumptive use, a term applied to remind us that wildlife watching from a tourist who pays for that privilege is akin to use and carries a cost also which can be to the wildlife (impacts of tourism) and to local communities. These might have to deal with large populations of protected species outside of protected areas, including wildlife which emerges to feed on their fields, or as might be the case with tigers or crocodiles, kill humans. They

How Environmental and Societal Changes Affect Wildlife in the Tropics

2187

might also have to deal with hordes of tourists as they descend from cruise ships using their land and scarce infrastructure. Wildlife trade is another impost on populations. There is the endless chain of middlemen as they collect and ship a vast range of wildlife and wildlife products to the cities around the world. I have chosen these three major uses of wildlife, by no means new, but massively expanded during globalisation, which have affected the relationship of modern society with wildlife.

Consumptive and Non-consumptive Wildlife Tourism The tourism industry, much of it around wildlife and nature has been subject to increased scrutiny from wildlife researchers as to its costs and benefits (e.g., Higginbottom 2004; Green and Jones 2005; Bauer and Giles 2001; Bauer and Herr 2004). The general message which emerges from these studies is the surprising importance of wildlife tourism and the benefit it might create for local and national economies. Figures for Australian whale watching tourism, well in excess of the A$ 200 million -then calculated values of enigmatic species as tourism drawing card (with the Koala estimated to be “worth” as a drawing card for the national economy of more than A$ one Billion), or the individual lion worth in excess of US $ 50,000 to the economy of an African lion country. These figures are impressive by any standard, yet all the more so as the proceeds can go (not always do) to local communities. There are also impacts associated with the “harmless” watching from elephant back for example as the case study from Royal Chitwan National Park, Nepal shows. In this example wildlife in habitats saturated with tourists, has to live in shrinking escape zones (because of ubiquitous tourism movement), loses its “natural behaviour,” becomes “habituated” in the “best case” scenario, extinct if it cannot handle that (Fig. 4).

Wildlife Trade and Wildlife Crime As a third major change in Human-Wildlife Interactions in the tropics has been the opening up of the global wildlife trade, legal and illegal, for pets, for medicine or as food item. With international transport and travel, international free trade agreements and internet as an oblique and save advertising and marketing tool, this multi-billion $ trade is taking a toll on wildlife which generally only declines after the supply lines dry up. In 2005, the Dalai Lama called on his Tibetan community to stop the trade with wildlife, which had emerged as a new threat to tigers, fuelled by an increasingly affluent Tibetan minority in Tibet. Before that, it had been the trade in tiger bones which went also from India over Nepal to China and before even that it had been tiger skins from Asia to the west mostly. The trade with wildlife and wildlife products, so it seems, not only has grown over time and with increasing globalisation. It has also been able to reinvent itself with diminishing wildlife resources, changing markets, fashions and ideas, always able to find a new market,

2188

Escape Zone Model

100 90 80 Percentage area left

Fig. 4 The impacts of wildlife tourism in Royal Chitwan National Park, Nepal. This modelling of escape zones (the areas animals are being displaced by prolonged and intensive tourism disturbance) as is the case in many wildlife viewing locations around the world, affects different species differently. The ones which are intolerant of humans (such as tiger here, blue line) experience further shrinkage of already reduced habitat. Species which can be more easily habituated (like one-horned rhino, red line) will be better able to co-exist with tourism (Curry et al. 2001; Cosgriff 1997)

J. Bauer

Small flush distance

70 60 50 40

Large flush distance

−

30 20 10 0 0

20

40 60 80 100 120 140 160 180 200 Number of tracks

outlet or a new product. Over the past decades there have been at least three major events, which have influenced the global market. – The emergence of Asia and particular China as new centres of affluence where a great appetite for wildlife either as food or as medicine has now been joined by the means to purchase it and an opening up of many poorly regulated regions to supply wildlife. – An increased understanding of the market, the establishment of a range of globally operating systems for both, monitoring it (TRAFFIC) and regulating it (CITES) – An increasing commitment, including from countries such as China to do something about it Trade with wildlife in a wider sense goes even further – there is another trade which has been hardly recognized yet remains the basis of the world’s gigantic pharmaceutic industry. The industry still depends to a large extent in its medicines on

How Environmental and Societal Changes Affect Wildlife in the Tropics

2189

natural compounds, mostly from rainforests. This trade, the biological information of complex organic molecules, often already selected through indigenous medicines, has added a new dimension to the value of the wildlife and its habitats. It promises huge gains, yet also requires huge investments. It is based on the biological information contained in species, which has become increasingly accessible with advances in molecular screening and replication technology. There is only a need to collect and trade limited quantities, essentially only one specimen is required.

Human-Wildlife Conflict An ancient interaction, the conflict between humans and wildlife, has received increasing modern attention over the past decade at least (e.g., Decker et al. 2002; Madden 2004). This is no doubt partly due to the fact that western, and increasingly non-western society, has become less tolerant to the needs of wildlife, in particular if that involves large, inconvenient and especially dangerous species. Another part of that attention however may be found in the increased separation between what human territories are and what should be wildlife’s place that we have determined it to be: our protected area system. As these “western” distinctions rarely hold in the real world we perceive increasing conflict in that fine balance. We also have, through modern conservation legislation, sometimes tended to support the rights of wildlife, neglecting that of local communities. This is the situation one may encounter if one is called into the house of a poor farmer family which grieves over the loss of their little daughter from a tiger or leopard last night (Maskey et al. 2001). Or the few hundred families who lose a family member to a crocodile in Africa every year. Or less threatening but just as real, as one encounters it from farmers in Asia, as they lament the loss of a crop to rhino, elephant, monkeys, deer, wild boar and so on (Boyd et al. 2002). There are long-term and invisible repercussions in this conflict. In Central Asia it leads to the prosecution of the few remaining snow leopards. With elephants in Sumatra or India it can become an unsurmountable problem for local communities. In brief: it is a situation which damages both, wildlife and communities.

Climate Change in the Tropics as It Affects Wildlife As described by Kaeslin et al. (2012) the major effects of climate change to wildlife include: – Ecosystem changes: These include geographical and altitudinal shifts, changes in seasonality and rates of disturbance, changes in species composition and a rapid increase in invasive species. – Species interactions: Impacts on wildlife species include changes in species distribution, abundance and interactions, for example through shifting phenology and mistiming.

2190

J. Bauer

– Human–wildlife conflicts: These are likely to increase as humans and wild species compete for the same dwindling resources. – Wildland fires: Increased drought, the drying out of previously wet forests as well as human interference and pressure are leading to more frequent and disastrous fires in ecosystems that are poorly adapted to such events. – Health and diseases: Wildlife, humans and livestock will be affected by the emergence and increased spread of pathogens, geographically and across species boundaries, due to climate, landscape and ecosystem changes. Most climate related changes in our wildlife populations simply happen with incremental gradual shifts, difficult to detect yet able to change the distribution of species, the composition of communities and the structure of ecosystems reported over a generation. What magnitude of changes can we expect? Current modelling predicts that even minimal global warming of one degree might already have significant effects. Kaeslin et al. (2012) conclude that “for scenarios of maximum expected climate change, 33 % (with dispersal) and 58 % (without dispersal) of species are expected to become extinct. For mid-range climate change scenarios, 19 % or 45 % (with or without dispersal) of species are expected to become extinct, and for minimum expected climate change 11 % or 34 % of species (again, with or without dispersal) are projected to become extinct.” Kaeslin et al. 2012: “According to the Intergovernmental Panel on Climate Change (IPCC) (Parry 2007), roughly 20–30 % of vascular plants and higher animals on the globe are estimated to be at an increasingly high risk of extinction as temperatures increase by 2–3  C above pre-industrial levels. The estimates for tropical forests exceed these global averages. It is very likely that even modest losses in biodiversity would cause consequential changes in ecosystem services (Parry 2007; Seppa¨la¨ et al. 2009).” Warming in the tropics will decrease the habitat of many endemic species of wildlife, which live in the cooler upland and montane rainforests, reducing their available habitat (threatened by many other factors) to isolated pockets of rainforest. For the Australian Wet Tropics in Queensland, a comparatively very small and therefore well researched rainforest area, it is predicted that seven frog species, five mammal species, three bird species and three skink species would lose over half of their present habitat with only a 1  C temperature increase. This would however only be the tip of an iceberg. As well as habitat changes, increased temperatures will physiologically affect some animals. Raised cloud levels are likely to change water cycles and affect some plants, frogs and skinks. Seasonal changes may change plant reproduction and fire regimes. Increased sea levels, cyclones and flooding may drastically affect coastal ecosystems through disturbance, altered water and fire regimes and an increase in vulnerability to exotic invasions and pathogens, one of the major threats to natural ecosystems in Australia. Most importantly ecosystems that are already fragmented, degraded and isolated (the increasing reality in many tropical regions) will be most affected. Across the pan-tropical world this threat is possibly most pronounced for coral reef systems where predicted increases in water temperatures from 2  C to 6  C will

How Environmental and Societal Changes Affect Wildlife in the Tropics

2191

have severe implications for the health of coral reefs, fisheries and entire coastal ecosystems and communities (see also WTMA 2008).

Future Projections for Mexican Faunas Under Global Climate Change Scenarios Global climates are changing rapidly, with unexpected consequences. Because elements of biodiversity respond intimately to climate as an important driving force of distributional limitation (Townsend et al. 2002). distributional shifts and biodiversity losses are expected. Nevertheless, in spite of modelling efforts focused on single species or entire ecosystems, a few preliminary surveys of fauna-wide effects and evidence of climate change-mediated shifts in several species, the likely effects of climate change on species’ distributions remain little known, and faunawide or community-level effects are almost completely unexplored. Using a genetic algorithm and museum specimen occurrence data, we Townsend et al. (2002) developed ecological niche models for 1870 species occurring in Mexico and projected them onto two climate surfaces modelled for 2055. Although extinctions and drastic range reductions were predicted to be relatively few, species turnover in some local communities were predicted to be high (>40 % of species), suggesting that severe ecological perturbations may result. African Lions Decimated by Climate-Influenced Pathogens Panthera leo (African lions) are now legally protected throughout their range, having been subjected to uncontrolled hunting in the past (Kaeslin et al. 2012). Their ecology is well studied and it is known that some populations thrive in certain protected areas of Africa. Lion numbers are, however, reported to be in decline in many areas, primarily due to the expansion of agriculture, ensuing control of problem animals, and, in some areas, poorly regulated sport hunting. Climate change brings new threats and exacerbates existing ones. In 1994, an epidemic of canine distemper virus (CDV) decimated the lion population in the Serengeti, causing the death of one-third of the resident population. This unusual die-off was followed by another event in 2001 in the nearby Ngorongoro Crater, the United Republic of Tanzania. A retrospective study was undertaken to understand these exceptional events, as CDV is an endemic disease in resident lion populations, but rarely causes mortality. In 1994 and 2001, analyses of blood samples of Serengeti lions detected unusually high levels of the tick-borne blood parasite Babesia leo. This parasite, among others, is usually detected at low levels in lion samples and ordinarily does not affect the health of the animal. The prevalence of this parasite was found to be at a very high level in prides suffering the highest mortality, while it was moderate in prides suffering no increase in mortality. This suggests that a co-infection with Babesia and the resulting lower immune status most likely was contributing to deaths caused by other pathogens among lion populations (Munson et al. 2008). Both of these CDV mortality events were linked to environmental conditions in 1994 and 2001, which were particularly dry and favored the propagation of ticks in

2192

J. Bauer

the Serengeti ecosystem. Tick (Ixodida spp.) levels on herbivores in the Serengeti were unusually high during these years, as extended droughts had weakened the animals. Lions feeding on this easily captured prey were very prone to high levels of infection by Babesia, due to the unusually large concentration of ticks present on the herbivores. Infection with Babesia triggered an immunosuppression, making lions more susceptible to the normally nonfatal CDV. Droughts and the resulting ecological conditions that led to these outbreaks are becoming more common in the Serengeti ecosystem. Munson et al. (2008) conclude that if extreme weather events become more frequent owing to climate change, mortality events caused by disruption of the ecological balance between hosts and pathogens are likely to become more common and to have devastating impacts on lion populations (Munson et al. 2008).

Conclusions We have seen in this chapter that wildlife in a rapidly changing world has declined, in so many regions, for so many reasons and to such an extent, that it has ceased to be a resource in particular for the poor and disadvantaged, often indigenous people. This negative trend for wildlife dependant forest people is further exacerbated by the development of a vast protected area system around the world and international and national legislation which make wildlife harvest and trade illegal. While there are examples how indigenous people have been granted special rights to continue their harvest (for example the harvest of marine turtles and dugong in tropical Australia) this general trend continues and despite of legislation and indigenous empowerment at the international stage. This chapter could only give a cursory glance of some of the major threats drawing the links to particular groups of wildlife. I refer the reader to the vast body of literature about the topics covered in this chapter, accessible to any internet search engine, but more specifically organisations (and websites) dedicated to them (see directory). On the other hand we can also see that many of these impacts continue, in fact accelerate, because of a rapidly growing tropical population, yet also a combination of ignorance, sheer carelessness and lack of commitment, mostly by industries in the various landuses, but also by the majority of politicians. Some things are changing however and there is increasing momentum to “clean up our act,” also when it comes to wildlife. In the next chapter I will describe the rather mindboggling number of initiatives, organisations, international conventions and activities to address the many impacts I have described.

References Aronson MFJ, La Sorte FA, Nilon CH, Katti M, Goddard MA, Lepzcyk CA, Warren PS et al (2014) A global analysis of the impacts of urbanization on bird and plant diversity reveals key anthropogenic drivers. Proceedings B http://dx.doi.org/10.1098/rspb.2013.3330. Accessed 11 June 2015

How Environmental and Societal Changes Affect Wildlife in the Tropics

2193

Bauer JJ (1993) Chapter 17: Wildlife Conservation and management. In: Pancel L (eds) Tropical forestry handbook, Springer, Heidelberg, New York, vol 2, 1st edn. pp 1059–1139 Bauer (2002) Bauer JJ, English T (2011a) Conservation through hunting – an environmental paradigm change in NSW, vol 1, Framing the game. Game Council NSW, NSW Government, Sydney Bauer JJ, English T (2011b) Conservation through hunting – an environmental paradigm change in NSW, vol 2, Raising the game. Game Council NSW, NSW Government, Sydney Bauer JJ, Giles J (2001) Recreational hunting: an international perspective. Wildlife Tourism research report series no. 13. Sustainable Tourism Cooperative Research Centre, Gold Coast Bauer JJ, Goldney D (2000) Extinction processes in a transitional agricultural landscape. In: Hobbs RJ, Yates CJ (eds) Temperate Eucalypt woodlands in Australia. Biology, conservation, management and restoration. Surrey Beatty and Sons Sydney Bauer J, Herr J (2004) Hunting and fishing tourism. In: Higginbottom K (ed) Wildlife tourism: impacts, management and planning. Common Ground, UK, Altona Bauer JJ, Bryant A, Goldney D, Schrader N, Costello D (2002a) Sustainable management and biodiversity conservation of semi-arid mixed Cypress-Eucalypt forests of NSW, Australia. NSW State Forests and Johnstone Centre, Charles Sturt University Bathurst, NSW Bauer JJ, Bryant A, Goldney D, Schrader N, Costello D (2002b) The vertebrate fauna of the Cypress (Callitris glaucophylla) forests of NSW Wales – impacts of forestry activities. NSW State Forests and Johnstone Centre, Charles Sturt University Bathurst, NSW Bauer JJ, Maskey T, Rast G (1995) River systems, hydrodevelopment and the species crisis in the Terai. In: Bhandari B, Shresta TB, McEachern J (eds) Safeguarding wetlands in Nepal. IUCNThe World Conservation Union, Heritage and Biodiversity Conservation Programme, Gland, pp 137–145 Beyond Pesticides (nd) Chemicals. http://www.beyondpesticides.org/pollinators/chemicals.php. Accessed 11 June 2015 Boyd M, Bauer JJ, Ren Z, Haohan W, Gadd L, DeLacy T (2002) Traditional ecological knowledge (TEK) of wildlife: implications for conservation and development in Wuyishan Nature Reserve. Fujian Province The International Program of the CRC for Sustainable Tourism, Griffith University, Green Globe Asia Pacific, Info Sheet 5 Bryan JE, Shearman PL, Asner GP, Knapp DE, Aoro G, Lokes B (2013) Extreme differences in forest degradation in Borneo: comparing practices in Sarawak, Sabah, and Brunei. PLoS ONE 8(7), e69679. doi:10.1371/journal.pone.0069679 Butler R (2012) Oil extraction: the impact oil production in the rainforest. Mongabay. http:// rainforests.mongabay.com/0806.htm. Accessed 11 June 2015 Caldecott J (1988) Hunting and wildlife management in Sarawak. IUCN, Gland/Cambridge, England Cosgriff K (1997) Wildlife Tourism in Royal Chitwan National Park, Nepal. Charles Sturt University, Charles Sturt University Albury, NSW Curry B, Moore W, Bauer J, Cosgriff K, Lipscombe N (2001) Modelling impacts of wildlife tourism on animal communities: a case study from Royal Chitwan National Park Nepal. J Sustain Tourism 9(6):514–529 Dasmann RF (1964) Wildlife biology. Wiley, New York Date L, Goldney D, Bauer JJ, Paull D, Bryant A (2000) The ecologically sustainable management of Callitris/Eucalyptus forests on the western slopes of NSW. In: Craig J, Saunders D (eds) Conservation in production landscapes. Surrey Beatty and Sons, Sydney Decker DJ, Lauber TB, Siemer WF (2002) Human-wildlife conflict management: a practitioner’s guide. NWDMC, Cornell University, Ithaca Dugan PJ (ed) (1990) Wetland conservation: a review of current issues and required action. IUCN, Montreux Finer M, Jenkins CN (2012) Proliferation of hydroelectric dams in the Andean Amazon and implications for Andes-Amazon connectivity. PLoS ONE 7(4), e35126. doi:10.1371/journal. pone.0035126

2194

J. Bauer

Finer M, Jenkins CN, Pimm SL, Keane B et al (2008) Oil and gas projects in the Western Amazon: threats to wilderness, biodiversity, and indigenous peoples. PLoS ONE 3(8), e2932. doi:10.1371/journal.pone.0002932 Goldney D, Bauer JJ (1998) Conservation in an agricultural landscape – fact or fiction. In: Pratley J, Candrel G (eds) Agriculture and the environmental imperative. CSIRO Publishers, Melbourne Goldney D, Bauer J, Bryant H, Hodgkins D, Watson G (1995) Winning battles but losing the war: the education marketing imperative. In: Saunders DA, Craig J, Mattiske L, Saunders DA, Craig J, Mattiske L (eds) Nature conservation 4: the role of networks. Surrey Beatty & Sons, Chipping Norton, pp 547–588 Green C, Jones I (2005) Serious leisure, social identity and sport tourism. Sport Soc: Cult Commerce Media Politics 8(2):164–181 Groombridge B, Jenkins M (2000) Global biodiversity. Earth’s living resources in the 21st century. World Conservation Monitoring Centre, Cambridge, UK Higginbottom K (2004) Wildlife tourism: an introduction. In: Higginbottom H (ed) Wildlife tourism: impacts, management and planning. Common Ground Publishing, Altona, pp 1–14 Horowitz LS (1998) Integrating indigenous resource management with wildlife conservation: a case study of Batang Ai National Park, Sarawak, Malaysia. Hum Ecol 26(3):371–403 Kaeslin E, Redmond I, Dudley N (2012) Wildlife in a changing climate. FAO forestry paper 167, Rome LMD (2014) Dirty water, dirtier practices. Le Monde Diplomatique Madden F (2004) Creating coexistence between humans and wildlife: global perspectives on local efforts to address human–wildlife conflict. Hum Dimens Wildlife 9:247–257 Mainka SA, Mills DVM, Mills JA (1995) Wildlife and traditional Chinese medicine – supply and demand for wildlife species. J Zoo Wildlife Med 26(2):193–200 Maskey TM, Bauer J, Cosgriff K (2001) Village children, leopards and conservation. Patterns of loss of human live through leopards (Panthera pardus) in Nepal (Report). Department of National Parks and Wildlife Conservation/Sustainable Tourism CRC, Griffith University, Kathmandu/Goldcoast Mazoyer M, Roudart L (2006) A history of world agriculture – from the neolithic age to the current crisis. Monthly Review Press, New York, Translated from the French by Membrez JH Munson L, Terio KA, Kock R, Milengeya T, Roelke ME, Dubovi E, Summers B et al (2008) Climate extremes promote fatal co-infections during canine distemper epidemics in African lions. PLOS one. doi:10.1371/journal.pone.0002545 Obiri JF (2008) Invasive plant species and their disaster-effects in dry tropical forests and rangelands of Kenya and Tanzania. Masinde Muliro University of Science & Technology, Centre of Disaster Management & Humanitarian Assistance. http://acds.co.za/uploads/jamba/ vol3no2/obiri_3_2.pdf. Accessed 11 June 2015 Parry ML et al (2007) Technical summary. In: Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, Hanson CE (eds) Climate change 2007: impacts, adaptation and vulnerability. Contribution of Working Group II to the fourth assessment report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK, pp 23–78 Robinson JG, Redford KH (1991) Neotropical wildlife use and conservation. University of Chicago Press, Chicago Seppa¨la¨ R, Buck A, Katila P (eds) (2009) Adaptation of forests and people to climate change. A global assessment report, vol 22. IUFRO world series. IUFRO, Helsinki Seto KC, G€uneralp B, Hutyra LR (2012) Global forecasts of urban expansion to 2030 and direct impacts on biodiversity and carbon pools. www.pnas.org/cgi/doi/10.1073/pnas.1211658109 Thwaites R, DeLacy T, Furze B, Bauer J (1996) Xilingol Biosphere Reserve: planning issues. The Johnstone Centre, Albury Thwaites R, Bauer JJ, DeLacy T (2000) Towards a sustainable production environment. In: Craig J, Saunders D (eds) Conservation in production landscapes. Surrey Beatty and Sons, Sydney

How Environmental and Societal Changes Affect Wildlife in the Tropics

2195

Townsend PA, Ortega-Huerta MA, Bartley J, Sánchez-Cordero V, Sobero´n J, Buddemeier RH, Stockwell DRB (2002) Future projections for Mexican faunas under global climate change scenarios. Nature 416:626–629 Veiga JB, Tourrand JF, Poccard-Chapuis R, Piketty MG (2003) Cattle ranching in the Amazon rainforest. http://www.fao.org/docrep/ARTICLE/WFC/XII/0568-B1.HTM. Accessed 11 June 2015 Wescoat JL Jr, White GF (2003) Water for life: water management and environmental policy. Cambridge University Press, Cambridge WTMA (2008) Climate change in the wet tropics – impacts and responses. http://www.wettropics. gov.au/site/user-assets/docs/ClimateChangeBook2008.pdf. Accessed 11 June 2015

The Development of Wildlife Governance, Science, and Management Capacity in the Tropics Johannes Bauer

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Changing Aims and Changing Actors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . International Conventions and Commissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . International Thematic Monitoring Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . International Warning and Alert Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . International Scientific Platforms and Information Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . National Legislation and Governance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Civil Society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Growing Role of Charities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Growth in Wildlife User Organizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Development of National and Regional Capacity in W&BDM . . . . . . . . . . . . . . . . . . . . . The Changing Role of Zoos and Wildlife Institutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Global Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Global Web Action Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Changing Role of Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Emerging Role of Indigenous People and Their Organizations . . . . . . . . . . . . . . . . . . . . . . . . . The Growing Role of the Media and Celebrities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organized Illegal Wildlife Trade and Crime Syndicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . International Responses to Climate Change as a Game Changer . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2198 2200 2201 2201 2202 2204 2205 2206 2209 2209 2211 2212 2213 2214 2215 2216 2217 2219 2219 2220 2220 2222

Abstract

In 1964, Dasmann suggested that “outside assistance from the more fortunately situated lands” should be forthcoming to better protect and manage wildlife in the tropics. Since then, just this has happened and there are now far-reaching J. Bauer (*) Australian Carbon Co-operative Ltd., Bathurst, NSW, Australia e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_174

2197

2198

J. Bauer

international conventions which seek to categorize and track the state of the world’s wildlife and biodiversity, regulate and monitor its trade and, above all, protect it. These and many other relevant conventions are now complemented by national and regional regulations, along with countless networks, global warning systems and information platforms. This chapter describes the growing scale of this global venture around wildlife and biodiversity, alongside the shift towards a view which recognises the human rights and needs of hundreds of millions of forest-dependant people. It also shows that wildlife conservation is no longer owned by the West and that many tropical nations have taken leadership within their own borders. A diverse, at times clashing, often collaborative environment around wildlife and biodiversity management has developed, where the role of the nation-state is in general decline, where there are a multitude of international agreements and where industry vies with civil society for implementation (and control). This chapter concludes that while much of the above action is laudable, it does not appear to have been enough to halt the decline of wildlife and natural ecosystems around the tropical world. Hope though now lies with a potential game changer in the action around climate change and the creation of forest carbon markets. Whether this will be enough however is open to debate. Keywords

Bhutan’s protected area network 7 • Birdlife International (BI) • Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) • Ducks Unlimited • Elephant Trade Information System (ETIS) • Federation of Associations for Hunting and Conservation of the European Union • Hunter and fisher organizations • International scientific platforms & information networks • International thematic environmental monitoring programs • Kruger National Park • Ramsar Convention • Self-regulation in environmental management • United Nations Permanent Forum for Indigenous Issues (UBPFII) • Web Wildlife & Biodiversity networks • Wildlife charities • World Wildlife Day • Yanomani people • Zoonotic disease management

Introduction Since Dasmann (1964) suggested that “outside assistance from the more fortunately situated lands” should be forthcoming to better protect and manage wildlife in the tropics a great deal of just that has happened. There have been far-reaching international conventions which seek to categorize and track the state of the world’s wildlife, regulate and monitor its trade (the Convention on International Trade in Endangered Species of Wild Fauna and Flora, CITES), and, above all, protect it. There are comprehensive agreements for the protection of wetlands important for bird migrations (Ramsar Convention) and a convention on the protection of migratory species themselves (Convention on Migratory Species). Protected Area Commissions drive the development of an expanding global protected area network, and specialists around the world collaborate within a growing number of

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2199

specialist groups in IUCN’s Species Survival Commission (which researches, coordinates, and initiates action and obtains funding for a wide range of wildlife problems/topics). There is the Red List on Endangered Species, continuously updated by that organization along with many other global indicators collected by the World Conservation Monitoring Centre (WCMC), now part of UNEP. Several global programs try to promote understanding and action on the ubiquitous spread of invasive species, and there are “Elephant Trade Information Systems” which try to keep track of what happens to African elephants (there are also others on rhinos, for example). And perhaps most importantly, there have been the Convention on Biological Diversity (CBD), the Rio Declaration, and the UN Framework Convention for Climate Change (UNFCCC), which give the overarching frameworks and the planetary strategic aim for wildlife and biodiversity management, sustainable development, and climate change action. These and other relevant conventions are complemented by national and regional regulations along with countless networks, global warning systems, and information platforms. As for direct action, each tropical country has at last national offices, often with support from multilateral and bilateral development aid, for environmental and wildlife conservation (protected area administration, policies and legislation on protection and use of wildlife). There are hundreds if not thousands of national and international environmental and wildlife charities, some of which (WWF, NC, CI, F&FI, WCS) have become multibillion-dollar wildlife empires, with offices in most countries, operating thousands of projects and in some places all but replacing government departments. There are also species such as tiger and giant panda for which entire protected area networks have been established and hundreds of millions of dollars spent. Many of these terrestrial efforts now have expanded to freshwater and marine environments. Zoos, little more than animal menageries then, now form a global network to retain “ex situ” wildlife diversity when “in situ” has started to fail or has failed. Both approaches are supported, indeed driven now, by inclusion of articles in the Convention on Biological Diversity (CBD) itself, supported by veterinary sciences which have the many technological/medical advances available. And to demonstrate that technical advance, there are increasingly well-funded initiatives (rewilding and synthetic biology) where scientists and conservationists inspired by Hollywood’s “Jurassic Park” have started to resurrect from scratch what we have lost (e.g., the Spanish Ibex), intending to recreate extinct species and wild and vanished landscapes, for example, in the heart of Europe where abandoned farmland is plenty. And all of that is now, unimaginable for Dasmann, instantly linked with web-based direct and instant conservation action. This action can, within a day or so, exert enough pressure to stop the mass killing of Amur falcons in India (Birdlife International, see case study) or encourage the President of Indonesia into making good his promise of the protection of a critical orangutan habitat in Aceh province. There are entire villages in Indonesia where hundreds of volunteers from around the world rehabilitate and care for orangutan orphans, rescued from logging operations or the pet trade. There are even dedicated rooms in hotels in Panama where frog enthusiasts from around the world try to rescue its collapsing frog populations. All of this is truly a monumental proof of

2200

J. Bauer

how much we all care about wildlife, how far we are prepared to go, and what we can achieve. This chapter tries to summarize and synoptically view the responses of the international community to a growing list of environmental concerns, many of them with wildlife at their center.

Changing Aims and Changing Actors In order to better understand these actions we now have to ask ourselves the question of how it all works and fits together. Who manages it? Why does all that action happen? What role do various stakeholders and actors play? And how have those changed over time and with what consequences? I believe these are some of the most important questions for the type of wildlife management which has been unfolding in the tropics through western and multilateral and bilateral aid. These initiatives have evolved into a very diverse, highly connected international form of wildlife management. In this new world there is an uneasy alliance forming between the original players (International Aid), challenged by newly emerging states with their government departments. Further important players are the multinational dimension of wildlife organizations and commercial projects, which have the expertise to act, as long as the money is there. These wildlife management alliances work within the overriding land uses: mining, agriculture, forestry, water, and fisheries, each of them with their own policy and legislative frameworks. These are generally much stronger than wildlife management regulations and often are contradicting its aims. As we will see in this chapter action is not straightforward and has become embroiled in many contradicting aims, often not in the interest of wildlife. I will briefly assess the role of the main actors: – The government and government agencies (between the province, state, and nation) – Local communities and traditional land users and land use systems – Indigenous people – Industry and the corporate sector (including mining, logging, fishing, and tourism) – Local, national, and international nongovernmental organizations ((I)NGOs) – International frameworks, multilateral and bilateral aid and conventions – Organized amateur (citizen scientist) groups (birds, reptiles, insects, orchids, etc.) – International and corporate environmental industry (EIAs, Climate Change) – National and regional wildlife organizations – National and international wildlife use organizations (hunters, fishers, tourism) – Illegal trade and crime syndicates Few of these actors now work independently, and this increasing collaboration has become a welcome tendency in many ventures to improve outcomes and avoid repetition and competition.

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2201

Governance International Conventions and Commissions International conventions are the frameworks for action, led by the United Nations. I have chosen several International conventions, for example, the Convention on Biological Diversity (CBD), the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), and the Ramsar Convention, which have, as framework for wildlife conservation, become irreplaceable tools in the international effort to better manage the world’s wildlife. They are now supported by most governments and are driving an increasing number of actions to achieve their objectives. Significantly they have also started to connect in order to reduce overlap and increase effectiveness. In the case of CITES this convention has “spawned” a wide range of complementary (and sometimes competing) activities and programs. I invite the reader to have a look at their websites and sample some of their reports. They are impressive examples on just how far we have come in 30 years. Conventions and complementary programs: – The Convention on International Trade with Endangered Species (CITES) and complementary programs (TRAFFIC, UNODC, CAWT): http://www.cites.org; http://cites-dashboards.unep-wcmc.org; Wildlife and Forest Crime Analytic Toolkit Revised Edition: http://www.unodc.org/documents/Wildlife/Toolkit_e.pdf. – TRAFFIC network – www.traffic.org (in English); China – www.wwfchina.org/ english; Japan – www.trafficj.org (in Japanese); Taiwan – www.wow.org.tw (in Chinese); Russian Federation – www.wwf.ru/traffic (in Russian); Indochina – www.wwf.ru/traffic (in Russian); Indochina – www.wwfindochina.org/traffic. htm (in English); Mexico – http://www.wwf.org.mx/wwfmex/prog_traffic.php (in Spanish) Another convention, even more far reaching for wildlife, has been the Convention on Biological Diversity, formed 17 years later when it became clear that CITES was not quite adequate to protect the world’s declining wildlife. – The Convention on Biological Diversity as overarching framework for the preservation of wildlife and its habitats (ecosystems): http://www.cbd.int; http://www.cbd.int/convention/text/default.shtml Three further far-reaching international conventions have been established to protect wildlife, ecosystems, and processes they depend on: the Convention on Migratory Species (CMS), the World Commission on Protected Areas, and the Ramsar Convention on Wetlands. It is suggested that the reader looks up the websites on CBD, CMS, etc. if only to catch a glimpse of their impressive (and growing) number of programs and activities.

2202

J. Bauer

– Convention on Migratory Species (CMS): http://www.cms.int – The Global Protected Areas Program and the World Commission on Protected Areas of IUCN with its associated programs: http://www.iucn.org Significantly, there are two major international initiatives, the Man and the Biosphere Programme (M&BP) and the Ramsar Convention (RC), which have taken, from their inception, a more inclusive approach to conserve nature and ecosystems than national parks and other protected areas. These may include a wide range of human activities, including, in the case of Germany, former nuclear power facilities, or, as is the case with Ramsar wetland sites, fishing and waterfowl hunting. As such they have played an important and different model to protect nature and are in a dialectic, at times competitive, often constructive relationship with IUCN’s protected area model. – UNESCO’s Man and the Biosphere Programme: http://www.unesco.org/new/en/ natural-sciences/environment/ecological-sciences/man-and-biosphere-programme/ Yet another protected area initiative is the Ramsar Convention, which, established 43 years ago, is an example how an ecosystem-specific organization can provide a collaborative framework and access point for an international effort which has started to change the negative trend of some of the world’s most important ecosystems and associated wildlife. Its aims are described on its website and its many publications. – The Ramsar Convention on Wetlands: http://www.ramsar.org – The Ramsar Sites Database: http://ramsar.wetlands.org/Database/ AbouttheRamsarSitesDatabase/tabid/812/Default.aspx Most importantly, each of these international programs and frameworks is now supported by national departments which in some countries administer a third (and more) of the countries’ land area for conservation, like Bhutan or Tanzania. The commitment of the west has been far exceeded by many poor countries.

International Thematic Monitoring Programs The development of international global networks which can monitor (WCMC), warn, and respond to wildlife needs and threats to it (e.g., IUCN-SSC 130 SSG) has been one of the major achievements in environmental/wildlife governance. With national nodes and in regional networks (e.g., the European Wildlife Disease Network or the Australian Wildlife Health Network) a capacity is emerging, which, so one hopes, will be able to better and faster respond to the outbreak of epidemics or even pandemics. This capacity is on the one hand struggling with

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2203

newly emerging zoonosis (HIV, SARS, bird flu, etc.) but has also continuously improved technology at its disposal. Any practitioner can instantly access information and report new incidences and threats (see websites below). We can also see that such “citizen scientist” information (e.g., bird watching) is one of our most powerful and reliable tools to detect climate change–related shifts in populations (or show, as below, the importance of protected areas as buffer zones). Over the past 20 years there have been development trends in the world’s tropical regions which worked mostly against but also for the natural environment. While the destruction of tropical environments has continued unabated, accelerating in some regions (e.g., the Congo), declining in others (Amazonia since 2004), some tropical timber-producing nations have destroyed their timber resources and dropped out of the markets, with others entering it (Sarawak). Now, however, we have started to know what we are losing while that happens. While in 1993 IUCN and WWF had only just established a small World Conservation Monitoring Centre (WCMC) whose task was to monitor species status in its RED LIST, this Centre has now become part of UNEP, has multiplied its tasks, is aided by dramatically improved computer technology, and collaborates with dozens of other organizations including UNCBD and UNCSD (RIO+20). From the huge resources available for free we can, at a fingertip, find the answer to many questions as to what ails which species where and what is being done about it. We can also see from these growing lists, however, that things are not going well and that while we know much more about what is happening, our ability for change has not improved at the same rate. And if we are working in that field as I have for decades, we are also only too painfully aware of the shortcomings of such lists and how little they actually often capture what is really going on. Also, the countless projects and associated scientific papers and reports are, while showing the amount of resources spent, not an indication of success. I would venture a guess that most projects fail in their objective to successfully sustain wildlife and its habitats. – UNEP’s World Conservation Monitoring Centre (WCMC): http://www.unepwcmc.org/ The news we get is mixed and at times difficult to interpret. We know that many forests are being destroyed (and this is being monitored across the world with Brazil with its own National System as the world leader), but we also know that many of those destroyed forests have started to grow back, while others have been all but irreversibly incorporated into the world’s growing agricultural estate (often as oil palm and soya plantations or cattle pastures). – World Database on Protected Areas (WDPA): http://www.unep-wcmc.org/ world-database-on-protected-areas – Protected Planet # UNEP-WCMC: http://www.protectedplanet.

2204

J. Bauer

International Warning and Alert Systems Disease emergence and spread do not respect geographic boundaries, and this is particularly the case with wildlife, where migrating species can transport pathogens across the globe. Zoonotic disease management therefore requires an integrated approach that involves different sectors: mainly human, livestock, wildlife, and food. Efficient early warning forecasting of zoonotic disease trends through functional surveillance systems is the key to effective containment and control. Another key is the development of global response networks, which can be put in place quickly, also in countries which have no capacity to do so. Early intervention during a disease epidemic often leads to better outcomes with reduced disease burden and associated economic impact. – Global Early Warning System for Major Animal Diseases, including Zoonoses (GLEWS), World Health Organization (WHO): http://www.who.int/zoonoses/ outbreaks/glews/en/index.html – Multidisciplinary disease analysis: http://www.glews.net – The European Wildlife Disease Association: http://www.ewda.org/ – Master of Wildlife Health and Population Management: http://sydney.edu.au/ courses/Master-of-Wildlife-Health-and-Population-Management; http://sydney. edu.au/vetscience/wildlife_masters/ During the dramatic transition toward a human-dominated world, humanenvironment (wildlife) interactions have drastically changed. While overall, most wildlife has been depleted, often destroyed, there is now a protected area network, which has in some places reached or even exceeded what optimists could have planned or hoped for (Chapter 5). There has also been, aided by countless development projects from United Nations and in particular the Global Environmental Facility (GEF) from World Bank, the development of wildlife and biodiversity policy and legislation in most tropical regions. The same can be said for the wildlife trade. We know that this trade has, along with growing incomes in Asia (the origin but increasingly the destination of that trade), grown dramatically over the past 20 years. We also know, however, that special organizations have been established during that time to monitor that trade (TRAFFIC), cooperating and funded by the United Nations Office on Crime and Drugs (UNOCD), the Convention on International Trade of Endangered Species (CITES), or Interpol. And if we know that elephant populations have dramatically declined over that period (e.g., the African Pygmy Elephant (formerly Loxodonta africana cyclops, now Loxodonta cyclops, a distinct species) has declined by 67 % over the past 10 years). Why do we know that? Because we have especially dedicated programs to monitor it, e.g., the Elephant Trade Information System (ETIS). Such monitoring efforts are not confined to governments and international agencies in our world of instant social media. If we want to see, for example, how the rhinos in South Africa are faring, we can find out that they do very bad indeed from a website and information network established by a group of

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2205

South African independent journalists (www.oxpecker.org). It is also sadly evident, however, that, while we are perfecting the counting and the monitoring, the “wildlife management” activity by itself was far less successful. Being cynical one might say that while we have become reasonably good in counting and measuring we have mostly failed in action.

International Scientific Platforms and Information Networks Along with scientific progress and the development of information and communication technology there has been a dramatic growth of scientific and information platforms and networks, which provide much needed information instantly. One of those is IUCN’s Species Survival Commission with a still growing number of Species Specialist Groups (130). Other ones focus on human-wildlife conflict, the emergence of zoonoses, particular ecosystems and landuses. There has been the emergence of thematic action networks which have become forces that challenge, e.g., wetland development and hydrodevelopment with their huge, indirect, and mostly unaccounted long-term environmental costs (Lamarque et al. 2009). What goes on in the world’s freshwater systems is, for example, monitored by the organizations listed below. These are just some of the bigger and global ones. For saltwater, each major ecosystem has a similar list. The list is large. Websites: FADA WCD AR GWP IRN Rivernet WI DU

Freshwater Animal Diversity Assessment: http://fada.biodiversity.be World Commission on Dams (WCD): www.dams.org American Rivers: www.amrivers.org Global Water Partnership: www.gwpforum.org International Rivers Network: www.irn.org Rivernet: www.rivernet.org Wetlands International Ducks Unlimited

Along with these databases, information networks, and warning systems there has been a wide and sustained effort to increase the knowledge and make that knowledge accessible through the Internet. – International information networks and scientific platforms to track wildlife populations and wildlife themes – IUCNs RED LIST and Species Survival Commission from its website: http:// www.iucnredlist.org – Invasive Species Specialist Group: www.issg.org. – The Global Invasive Species Information Network (GISIN): http://www.gisin. org

2206

J. Bauer

National Legislation and Governance National policy, legislation, governance, and capacity for implementation are the most important frameworks for wildlife management which, with the exception of large parts of the oceans, happens at national levels. The development of environmental policy and legislation in wildlife conservation, sustainable development, and climate change in tropical countries with greatly differing governance has been one of the greatest achievements of the international community and respective nations over the past 30 years. Many national moves were guided by international conventions and the current State of the Environment reporting (SoE), national reporting to the United Nations Convention on Biological Diversity or national communication on GHG emissions to UNFCCC are examples of national “housekeeping” through compliance to international commitments. Always a work in progress, professionals working in offices and departments within each country are in constant communication with many of the governance and management bodies described in this chapter. Conversely there is an intense and instant level of scrutiny on what happens in nations through such membership bodies and the media. While this type of national regulatory capacity developed, there has however been a level of foreign large-scale natural resource use/exploitation (mining and oil, forests, agricultural land, coastal fisheries) investment, which has increasingly challenged the role and capacities of the nations, at times with disastrous results. In one extreme case, PNG, such massive foreign investment in logging, mining, gas, and oil exploration and marine fisheries has pitted a central government with poor and compromised capacity against its own communities and undermined a socially harmonized development. By no means restricted to PNG but widespread throughout the tropical world the power and function of the state to act as regulatory and harmonizing body for national development has been made beholden to foreign interest. It is often up to bodies such as the UN, World Bank, or other large donor organizations to balance, often unsuccessfully, this destructive power. Bhutan’s Protected Area and Wildlife Corridor Network Bhutan has, along with many other tropical countries such as Tanzania, China, or Costa Rica, made a major investment in its National Protected Area system connected by a network of corridors. As this system in other tropical countries has been superimposed on its production landscape it will need to maintain sustainable land use and wildlife while generating income opportunities, mostly from tourism, for its local communities. The map of this protected area network (as in most tropical places) shows that its major content is the integration of wildlife needs with the livelihood needs and aspirations of local communities (Image Tshering 2008, Fig. 1). Most government departments now have sections dedicated to protected area and wildlife management, sometimes as divisions in the Ministries of Forestry. Many of those departments are the national representatives of international agreements and conventions (CBD, CITES, RAMSAR) and work closely with multilateral and bilateral funding bodies which either advise and/or provide

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2207

Fig. 1 This depiction of Bhutan’s Network of Protected Areas and Biodiversity & Wildlife corridors by the Ministry of Agriculture and Forestry shows well its Buddhist philosophy as it embraces the communities of people, plants and animals it looks after. (Image by courtesy of K. Tshering, MoAaF, Bhutan)

capacity and development funding. Many such departments are also in charge of environmental management (including mining, forestry, and development) and work closely with nongovernmental organizations (such as WWF, CI, FFI, and BLI).

Wildlife Management in the Global Protected Area Network Much of the world’s management capacity for wildlife focuses on protected areas. Protected by legislation WM in these zones should be straightforward and “unencumbered” by communities, which generally lose their user rights – or so the theory goes. In the real world, however, where many internal and more external influences continue to degrade ecosystems (e.g., hydrodevelopment affecting its aquatic systems) and where populations of animals have become so small often that they require support, wildlife management becomes a critical task. I have chosen Kruger NP as an example how extensive – and expensive – reintroduction programs are part of that wildlife management. My own summary and evaluation of that effort, however, also shows how limited the success of such ventures often is, if one delves deeper, even in world-famous and well-resourced areas such as KNP. Restoring Wildlife Communities: A Kruger Experience Kruger National Park, whose overall very successful management history is recorded in Du Toit et al.’s (2004) milestone book, provides an impressive example of the fate of

2208

J. Bauer

reintroduction programs, even in seemingly well-managed and resourced environments. Attempts to restore not only locally extinct populations but previously depleted mammal communities in Kruger National Park go back to 1918 (Freitag-Ronaldson and Foxcroft 2003), when the Game Reserves Commission made provision for reintroductions of species which had disappeared before protection measures had proven to be effective. A reintroduction policy vacuum in 1918–1948, which resulted in serious considerations to introduce exotics (thank God never implemented), was followed by a reintroduction surge between 1962 and 1989. During this time 16 species of mammals (not counting a guinea fowl release in 1930 –probably successful) were released during more than 30 release events (not recorded consistently for two species of rhino for which above authors comment on “numerous”). I have attempted to interpret the above author’s slightly ambiguous data (no doubt mostly caused by very poor recording of events) and taken the liberty to call, e.g., attempts which resulted in heavy mortality with four surviving animals “unsuccessful” and also assumed that “unknown” cases suggest that the animals have not done well. It is obvious that most of these efforts took place in the 1970s and involved more than 850 animals, in the case of the reedbuck 340 individuals (Fig. 2). The success of these operations as recorded by Freitag-Ronaldson and Foxcroft (2003) is not encouraging. It seems that only three species, reintroduced either in very large numbers (370 mountain reedbucks) or during “numerous” (FreitagRonaldson and Foxcroft 2003) release events (Black and White Rhino), have successfully established good populations. A fourth species, the Nyala antelope, seems to have done reasonably well (“regularly seen along a road,” whatever that means) while for a fifth species, the sable antelope, the authors only record that “64 were released in 1976” with no other information (thus either to be categorized as unknown or even more likely unsuccessful). The fate of all the other expensive attempts is either unknown or unsuccessful. The reasons for this rather worrying failure to restore original diversity in KNP are several: Oribi and grey reebok have seemingly failed to reestablish because of unsuitable habitat. High tick burden, anthrax, and drought were seen as a cause for failure of the Roan antelope while the limited success of the Eland is being attributed to ticks. The only limited success of Lichtenstein’s hartebeest and the suni seems to be due to poor captive breeding performance. For the great majority of cases the reasons for failure are unknown. There are several lessons to be learnt from the Kruger experience. For once it shows how poorly and ad hoc reintroductions were handled, even within a park which aspires to become a world model and even up to relatively recently. It also shows how difficult reintroductions are, even within very natural, extensive environments not plagued by exotic predators such as is the case in Australia. And lastly it shows that even very large efforts have been compromised by very poor follow-up and monitoring. In the case of KNP this is particularly surprising, as monitoring was one of its great strengths. Considering the huge size of the area (more than 20,000 km2) and the complexity of the animal community, however, suggests that this would have been a very difficult task.

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2209

Number of Reintroductions 6 5 4 3 2 1 0 1961- 1966- 1971- 1976- 1981- 1986- 1991- 199665 70 75 1980 85 90 95 2000 Not Known 24%

Unsuccessful 28%

Successful 16%

Limited Success 32%

Fig. 2 Reintroductions of mammal species in Kruger National Park, SA. There are several lessons to be learnt from this Kruger Experience. For once it shows how ad hoc reintroductions were handled, even within a park which is a world model of scientific management and even up to relatively recently. It also shows how difficult reintroductions are, even within very natural, extensive environments not plagued by exotic predators such as is the case on many islands and in Australia. And lastly it shows, that even very large efforts have been compromised by poor follow-up and monitoring. In the case of KNP this is particularly surprising, as monitoring was one of its great strengths. Considering the huge size of the area (more than 20 000 Sq.km.s) and the complexity of the animal community however suggests that this would have been an almost impossible task (data based on park records as recorded in du Toit et al. 2003)

Civil Society The Growing Role of Charities The largest growth sector in wildlife (including animal rights as they apply to nondomesticated species) and environmental action has happened in civil society, driven first from western cities where a large number of formal and informal action

2210

J. Bauer

networks have culminated in what is most appropriately described as a charity industry based on wildlife. Already substantial in 1993, it is now dominated by a dozen large organizations, mostly US and UK based, with billion-dollar budgets engaging in an increasingly diverse – and ambitious – range of portfolios (In organizations such as WWF wildlife now encompasses nature and environment, biodiversity, climate change, etc.) and increasingly guided (and funded) by international development agendas with their many social goals (Millennium Development Goals, Rio Declaration, Convention on Climate Change). These “general” wildlife charities are complemented by user and advocacy organizations, which have included many of their agendas and increasingly collaborate with charities. 7a Conservation Organizations (i) Wildlife Charities (BLI, FFI, WCS, FZS) (ii) Conservation Charities (WWF, WCS, CI, NC) 7b User and Advocacy Organizations (iii) Hunting Organizations ( FACE, INDABA, DU, SCI) (iv) Animal Rights and Welfare Organizations (v) Zoos and Wildlife Sanctuaries Including Private Industry (WCS,FZS) NGOs, many of them international, have started to play an unprecedented role in the management of wildlife around the world. Most of them have developed their own scientific expertise and are very closely aligned with scientists. Many of them, however, are also closely aligned to governments, and sometimes industry, both of them major funding bodies. Others such as Greenpeace maintain their independence by electing to remain membership based and funded. Global scrutiny of their ventures, for example, through journalism or the USA-based Charity Navigator (www.charitynavigator.org) has considerably added to accountability. There are also individual and independent websites and watchdog organizations such as Amazonwatch, Mongabay, or Chris Lang’s REDD-Monitor, which have started to play an important role for transparency, discussion, and especially dissent. I suggest that the reader visit some of the websites of wildlife charities (see directory) in order to gain an appreciation of the scope of their activities and programs. While most of these NGOs as described above have evolved along the lines of natural sciences there are notable exceptions (e.g., WSPA and PETA) where the concerns of a growing number of people with the treatment of livestock and pets have spilled over to wildlife. The western and urbanized agendas of these organizations are often at odds with realities in the (tropical) “wild world” or of natural sciences including wildlife management principles. But then, they also provide important and independent voices which deserve to be heard and are particularly important as a countervoice to a science which has all too often become compromised by neoliberal economics and industry/government interest.

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2211

– Animal welfare groups: World Society for the Protection of Animals: http:// www.wspa-international.org – PETA – People for the Ethical treatment of Animals: http://www.peta.org/aboutpeta/

The Growth in Wildlife User Organizations Hunting and Fishing in the modern world have two faces. On the one side they are the land use of many rural and indigenous people. On the other they have become recreational activities in western countries which have created multibillion-dollar industries and are, as part of tourism, an important factor also in the ecology/ economy of large tropical species (Bauer and Giles 2002; Bauer and Herr 2004). In continental Europe and North America hunters and fishers, pay much money to be able to hunt and fish and have formed powerful organizations, which continue to play a major role in the management of wildlife populations. The largest organization of its kind in the world is the Federation of Associations for Hunting and Conservation of the European Union (www.face.eu), which with its seven million members is providing a unique force in wildlife management, at no cost to government and with great financial benefits. Other organizations such as Ducks Unlimited, founded as a response of hunters and fishers to the destruction of wetland in North America, have become world leaders in waterfowl and wetland management. – Ducks Unlimited as a North American Wetland and Waterbird Force: www. ducks.org Complementing (and at times dwarfing) similar conservation projects and commitments from mainstream conservation efforts and NGOs, DU has become a highly successful and competent body to conserve and restore wetland habitats in North America (Canada, USA, Mexico). As of 2014 it has protected and restored more than 50,000 km2 of wetlands and is influencing (improved management) another ~ 400,000 km2 through its 700,000+ members. Similar efforts are advancing across Europe, including Russia, often supported by old, traditional, and wellorganized hunting organizations which have, for many years (e.g., the “office national de la chasse et de la faune sauvage (ONC)” in France, the “Bundesjagdschutzverband (BJV)” in Germany, or the more recent “Game Conservancy (GC)” in the UK) worked closely with governments (at times implementing its policies and laws) to protect wildlife and its habitats. These organizations have also developed century-old systems of game statistics which are unique in their ecological information content. Controversial for many years, hunter and fisher organizations, in particular the International Council for Game and Wildlife (CIC), have now been successful in defining and bringing across their powerful conservation message. Accepted and cooperating with the world’s major conservation bodies (such as UNEP, IUCN)

2212

J. Bauer

they have started to form global alliances to diversify conservation approaches which include the legitimate sustainable harvest and use of wildlife. – Collaborative Partnership on Sustainable Wildlife Management: http://www. fao.org/forestry/wildlife-partnership/en/

UN General Assembly Proclaims 3 March as World Wildlife Day 27 December 2013, Geneva, 23 December 2013 – On 20 December 2013, the sixtyeighth session of the United Nations General Assembly decided to proclaim 3 March, the day of the adoption of the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), as World Wildlife Day, to celebrate and raise awareness of the world’s wild fauna and flora. Celebrating World Wildlife Day 3 March 2014 – The Collaborative Partnership on Sustainable Wildlife Management welcomes the United Nations General Assembly decision to proclaim 3 March as World Wildlife Day, as a means of celebrating the importance of the world’s flora and fauna, strengthening efforts to conserve biodiversity, and stepping up the fight against the illegal trade in wildlife. This start of collaboration instead of much conflict in the past is a major improvement of policy conditions for wildlife management/biodiversity conservation and will pave the way for a new era where money is being used for joint action instead of wasted for futile, costly, and socially destructive conflict. It has also cemented the position of other international hunting advocacy groups and organizations (CIC, FACE, Indaba) in a much-needed renegotiation toward wildlife management in the lesser-developed nations in particular the tropics. – African Indaba: http://www.africanindaba.co.za/ – International Council for Game and Wildlife Conservation: www.cic-wildlife.org

The Development of National and Regional Capacity in W&BDM Wildlife and Biodiversity Management and Conservation in the tropics, conceived and still driven by the west through charities and multilateral and bilateral development agendas has developed own regional institutions which, like the Wildlife Institute of India (WII) in Dehra Dun, might have become governmental institutions while the College of African Wildlife Management (CAWM) in Tanzania, established by a US charity, is now funded by multiple donors. I have chosen these two institutions, both with considerable regional outreach as examples, how tropical regions have started their own approaches. – African Wildlife Foundation and Mweka College: http://www.awf.org and http://www.mwekawildlife.org – Wildlife Institute of India in Dehra Dun, Uttar Pradesh: http://www.wii.gov.in

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2213

In the same country, a civil society and conservation leader in Asia, an Indian organization was formed by Elephant Authority Prof Sukumar to form an umbrella group for the conservation of the approximately 22,000 Asian elephants remaining in increasingly shrunk and fragmented habitat within a growing sea of humanity. The premise of this organization is that the elephant’s needs will cover most of the needs of biodiversity. – Asian Nature Foundation: http://www.asiannature.org – From Amateur Groups to Citizen science networks: http://www.birds.cornell. edu/citscitoolkit/about/definition Amateur wildlife enthusiasts, now more appropriately termed “citizen scientists,” have and continue to make a major contribution to science and the knowledge and management of wildlife. Much of our knowledge of taxonomy (and evolution) is based on what groups of amateurs collected and organized over the past centuries and that applies to insects as much as to birds, reptiles, fishes, or plants. Organized amateur and advocacy groups (led by birdwatchers, hunters, and fishers) have now become major actors in wildlife conservation with bird watchers, a multimillion-member community around the world and with centers in most nations, having developed organizations such as Birdlife International. One of the outstanding examples in member-driven wildlife management, birdwatchers have become the most sophisticated and universal environmental monitoring system in the world, based on organized, collective, and scientifically planned and analyzed regular global surveys by millions of members at no cost. There remains much development potential in that sector which is even attractive for the younger generation in particular through themebased volunteerism. There are organizations such as Friends of the Earth International (FoEI) who make extensive use of “conservation volunteers” which assist scientists in their projects around the world, often ending up as scientists themselves. – Birdlife International (BI), “the global partnership for nature and people”: http:// www.birdlife.org If one looks at BI’s website and sees the large number of logos, each from a national organization one cannot help but be impressed how ’bird lovers’, a group of people one finds in every country, have managed to become so organized that they are now a powerful organization in the world which plays an important role in bird conservation. Nor is it the only representative speaking up for birds. Wetlands International (grown out of bird organizations) and others make similar contributions.

The Changing Role of Zoos and Wildlife Institutions There is a new understanding and changing function of zoos and wildlife parks as they change from wild animal menagerie to a global network of powerful institutions. They have managed to combine income generation through tourism as one of

2214 Fig. 3 The viewing of wildlife, one of the great contents of many forms of tourism targets animals in various forms of captivity and in the wild. There is a continuum of costs, profits and impacts as these experiences are sought

J. Bauer

KEEPING COSTS ATTRACTIVENESS OF VIEWING ENVIRONMENT EASE OF VIEWING IMPACTS OF TOURISTS OPERATIONAL COMPLEXITY VIEWING

TOURIST NUMBERS PROFITS CAPTIVE (Zoo)

SEMI-CAPTIVE Halfway (Exclosure)

WILD Protected Area

the entertainment hubs of big cities with nowadays astonishing zoo research/ science and even involvement in “in situ projects. This role and extent, e.g., described by Tribe (2001), has however further grown. Some of them like the three New York Zoos and the Frankfurt Zoo have created large wildlife organizations (Wildlife Conservation Society (WCS) and Frankfurt Zoological Society) with numerous projects around the world (Fig. 3). Viewed as a complementary system zoos and protected areas are in a unique situation to collect and funnel large resources from western societies and the urban rich into wildlife and protected areas, all as part of extended wildlife tourism.

Global Media Much has been written about the responsibility – or lack thereof – of the global media in presenting, and trying to influence, the general public in matters of environment and wildlife. The emergence of wildlife TV shows and programs in the wake of David Attenborough is testament to the huge role media plays in shaping our responses to wildlife. Although the growing number of wildlife shows has played a major role in garnering the support of the western society in wildlife and environmental action in the tropics, there are now more specialized types of journalism which have opted for direct action. Investigative journalism and action as a transparency mechanism: – Oxpecker – Investigative Journalists against Wildlife Trade: http://oxpeckers.org – Amazonwatch: http://amazonwatch.org/ Nor is investigative journalism confined nowadays to disseminate stories of environmental and wildlife abuse around the world. Rhett Butler, the founder

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2215

of Mongabay, has developed, with NASA, nothing less than a global deforestation monitoring tool. – Mongabay (http://news.mongabay.com) Deforestation tracking tool – Read more at http://news.mongabay.com/2014/0319-katerva-award.html#bgf S4d7m4382I9Ue.99 Similarly Chris Lang’s REDD Monitor (www.redd-monitor.org) has become the most informed voice in the growing field of forest carbon (trading). Nor are they the only journalists or civil rights activists who have started to play an important role in what happens in wildlife and environmental management around the world. There is a group of less specialized but all the more connected action networks which operate also in matters of wildlife. They are getting big, and they continue to grow.

Global Web Action Networks Perhaps as a culmination of the above, there are now social media activities which have started to target specific problems around the world, sometimes within hours or days and with astonishing success. Avaaz.org, founded by the Canadian Ricken Patel, is currently emerging as the largest of its kind (Avaaz, e.g., has now 32 million members). – AVAAZ The World in Action: http://www.avaaz.org Such information and action networks play a growing role to – Identify and share information on wildlife violation – Identify and connect partner organizations which can act – Address and if necessary exert pressure on governments and stakeholders through “global shaming” – Mobilize resources for action – Implement and set in place education programs – Implement local institutions to own, oversee, and monitor/govern future activities/violations Such web-based action has been increasingly popular and successful as it is applied by many wildlife charities. A recent post on the trapping of Amur falcons for food in India gives an impressive example how fast and effectively such instantaneous networks can work in the modern world where “big brother watches” and media are on constant alert for such stories. Action for Amur Falcons Brings Hope for an End to Hunting in Nagaland Last year’s news of the massacre of Amur falcons in India shocked the world. BirdLife’s Indian partner BNHS moved immediately to mobilize a response. The trapping was

2216

J. Bauer

stopped, nets destroyed, and arrests made, although not before terrible damage had been done. This year, the generous response to our international appeal has enabled BNHS, with the support of the BirdLife Partnership, to organize a comprehensive program to keep the falcons safe around the Doyang reservoir, where they roost during their stopover. The program has mainly been implemented by a local NGO, Nagaland Wildlife and Biodiversity Conservation Trust, working with the Nagaland Forest Department. As a result, not a single Amur falcon was trapped during the 2013 autumn migration. Attitudes have changed so much in the space of a single year that the Amur falcons are now treated, in the words of Nagaland’s Chief Minister, as “esteemed guests.” A year ago we brought you the shocking news of a hunting massacre taking place in Nagaland, India, which BNHS (BirdLife in India) had been alerted to by colleagues from the campaigning NGO – Conservation India. Tens of thousands of migrating Amur falcons () were being illegally trapped on the roost at a reservoir at Doyang and then being taken to local markets alive, or killed and smoked, for sale as food. Online news articles and a graphic video of the atrocity were quick to spread via social media. Policing, Local Law Enforcement, and Education were implemented swiftly. As one suspects that this annual Amur falcon migration provided important food for people one might also expect/hope that steps to compensate for that loss of a wildlife species as food source are being taken.

The Changing Role of Science Although in this chapter I treat science as only one (if a crucially important one) of the many stakeholders and drivers in the evolution of W&BDM, it has become the underlying logic and in some ways even framework for most action. Science not only underpins what policy and legislation comes up with. It also buttresses the rationale of international conventions. Much of the “on-ground action” by NGOs and land user groups is nowadays based on science. Most of the larger NGOs have their own science departments. And every minister responsible for a particular portfolio has her/his national scientific advisory body. Perhaps most importantly, however, science and scientists have developed their own large stakes around the environment including wildlife conservation and are supported by their own powerful organizations. National research bodies ranging from universities to research institutes depend on grants from industry and government. Many of them have become part of the industry or developed industries around it. And all of them have shaped science to suit them, described by Davies (2004) with rare and often disturbing insight. There has been a wide range of activities to introduce industry standards which complement and go beyond those set by the International Standardisation Organisation (ISO). Further, with many billion dollars dedicated to environmental and social development in the tropics a large industry around this has developed. An environmental industry has grown around “Environmental and Social Impact Assessments” in particular in mining and oil exploration/extraction. It is not

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2217

surprising that the industries themselves have gone into that sector to exert control of the process (they pay for it), recapture money otherwise lost to the process, and demonstrate social and environmental commitment. The problems lie in the many details. In a book called Rethinking Voluntary Approaches in Environmental Policy, Rory Sullivan, the director of the UK-based company of investor responsibility, Insight Investment, in London, had a closer look at voluntary approaches such as Corporate Responsibility and Codes of Conduct as alternatives to traditional approaches of environmental regulation. He did this in order to examine the rather widespread cynicism in the general public toward that “self-regulation” and analyzed three initiatives: Environmental Management Systems (1), the Australian Greenhouse Challenge (2), and the Australian mining industry’s Code for Environmental Management (3). In particular he examined the narrow putative advantages of such self-regulation (reduced cost, increased flexibility) and the wider multiple environmental goals such policy needs to be part of. Sullivan (2005) concludes that neither opponents nor proponents of voluntary approaches to self-regulation of industry are very convincing. He suggests that while the three case studies provided some evidence of effectiveness and benefits, the reported environmental outcomes “lack dependability,” could probably have been achieved from other approaches (e.g., regulation, I imagine), and none have “been designed to gather substantive and credible environmental, economic or other performance data” (Sullivan 2005). The history of the International Timber Trade Organization (ITTO) may also stand as an allegory for the limitations of industry responses. Sullivan’s (2005) analysis suggests that while at least for Environmental Management Systems as specified by ISO 14001 and the Greenhouse Challenge there have been benefits in environmental performance, it was unclear whether these incurred because of code compliance or simply as a response to a better regulatory environment. More specifically for the Australian Mining Industry, it was also clear that, while environmental performance had increased, the level of that improvement remained uncertain and poorly monitored, and grave abuses, especially in mines overseas, occurred whether they adopted any code of conduct or not. Each of the industries connected directly with wildlife and biodiversity conservation: forestry (e.g., Forest Stewardship Council (FSC)), hunting, fishing (various ertification schemes), and adherence to operational improvements, for example, “Turtle Exclusion Devices” (TED) and tourism (Green Globe 21, Green Leaf, etc.) have undertaken attempts for better environmental performance as response to the depletion of its resource, a more regulatory environment, and pressure from consumer groups. Most industries are now at pains to point out their approaches to achieve sustainability, helped by the vague definition of the term.

The Emerging Role of Indigenous People and Their Organizations Ignored in international development agendas, persecuted by loggers, miners, farmers, and governments, and forgotten in modern societies, indigenous people around the world have in few places been able to maintain or reclaim (some) of their

2218

J. Bauer

rights. While indigenous nations such as PNG are the great exception (communities have “managed” to retain 98 % of their land (e.g., Sakulas et al. 2013) and are led by an indigenous government), Canadian, Brazilian, Australian, or New Zealand First People have been able to reclaim some of their rights and land. Many others, in particular in the tropical regions, have not. Perhaps most importantly, however, many indigenous people have been prevented from development as their own land use, mostly collecting, hunting, and fishing, was not recognized as a “legitimate” land use, often inaccessible because of protection by national or international decree. As much of that land remains in remote and less accessible regions, many of those indigenous people have either lost the entire or parts of their land use as their land is protected under western and not entirely well-guided protected area legislation. The protected area system in the tropics has often allowed national governments to exert their dominance. There are only few states (e.g., the state of Sarawak in Malaysia) where wildlife use and management has been promoted by the state and in the interest of indigenous people. In other places such as Nepal’s south, indigenous people (there the Tharu people) have lost most of their indigenous hunting and fishing rights while having to compete with many newcomers (e.g., illegal “squatters” from India). The International organization Survival International gives much insight into the extent of such displacement. There are also a growing range of activities under the auspices of UN to support indigenous people and their emerging organizations. Not a turnaround but a crucial benchmark for indigenous people was the establishment of the United Nations Permanent Forum for Indigenous Issues (UBPFII), which has provided momentum for the gathering indigenous voice. – The United Nations Permanent Forum for Indigenous Issues (UNPFII or PFII) There are now independent and instant communication pathways to support indigenous people in their ways of life which support wildlife conservation. – Survival International as indigenous advocacy group: http://www. survivalinternational.org/tribes/yanomami For example, the Yanomami-Mining, ranching, and health care chaos threaten Yanomami. For thousands of years, the Yanomami have thrived in the rainforests of South America. As indigenous people in the modern world continue to be disowned, displaced, murdered at times, if they stand in the way of development (see www. survivalinternationl.org), they have also, in some places, managed not only to reassert their rights but also gained some significant successes. Australia’s tropical north may well serve as a case study how that may proceed but also what kind of problems indigenous people and their land management (which often is wildlife, NOT agriculture) still face. Far from being a success story, it is a cautionary tale on things which work and on things which do not and above anything else of the

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2219

remaining tensions between the state and the indigenous peoples (e.g., Bauer et al. 2009). Theirs is the oldest continuous land use history in the world and one which has, not for reasons of “primitiveness” but for very sophisticated ones, “stuck to” the sophisticated management of a very diverse and rich resource, wildlife. NAILSMA in northern Australia is an example how indigenous people are recovering their heritage. – NAILSMA and the Northern Land Council in Australia’s Indigenous Tropics: http://www.nailsma.org.au/

The Growing Role of the Media and Celebrities On 26 March 2014, David Beckham visited the Yanomami tribe in Brazil in the run-up to the 2014 World Cup and met their most prominent spokesperson Davi Kopenawa, known as the “Dalai Lama of the Rainforest” (http://www. survivalinternational.org/news/10099). According to SI’s website, “Beckham and Davi” talked about the problems that the Yanomami face, especially the illegal gold mining on their land. The Yanomani people “liked David’s visit a lot because he was very interested in the problems in the Yanomami reserve. He saw that there are many threats to the environment and to our culture. He showed he was concerned about the Yanomami people.” Millions of Beckham fans followed that visit on his social media sites. There have been many visits like Beckham’s to indigenous people and around endangered wildlife, as both are popular as “worthy causes” for the world of the superstars and for many other reasons. Some of them have sought (and become) UN ambassadors for a worthy cause; others have even done courses on how to engage with that world. Powerful in attracting mass media attention, distributed around their millions to mostly young followers through Facebook, Twitter, YouTube, and other social media, such brief attentions are now able to sway governments.

Organized Illegal Wildlife Trade and Crime Syndicates Unfortunately there is also action in matters of wildlife which is on “the dark side.” As soon as traditional land and wildlife users are denied access to a valuable wildlife resource either by legislation or wildlife protection activities, the illegal sector may take over and flourish (while legal activities cease and the development of sustainable practices is prevented). By their very nature they have no official websites and operate however extensively on Dark Internet (dark address, deep web, dark net). A “dark force” in wildlife management, they have the capacity to offset a wide range of activities of those who act in the interest of the global community, the environment, and wildlife.

2220

J. Bauer

International Responses to Climate Change as a Game Changer Starting long before the Kyoto Protocol (with many scientific studies, technological advances, working groups, and meetings providing increasing knowledge and momentum) growing concern about a changing world climate through human actions, and the impact of that on the environment and human societies (land use, fisheries, settlements, frequency of catastrophic climate events, etc.), has in 2014 become the biggest global environmental concern. Growing understanding on the process and causes of climate change has made climate change the center of environmental attention, negotiation, and action. This is reflected in the contents of international multilateral and bilateral programs, government policy, and civil society. Many environmental contents and programs from most sectors including science itself are currently renegotiated around climate change. Wildlife programs set up with something else in mind are now newly scrutinized in the context of a changing climate. A study by researchers from the UK as reported in the Climate News Network may serve as an example how many past activities, here buffer zones, created for mammals originally, are finding “new uses” for birds in a changing climate. Birds are responding to climate change and land degradation threats by using nature reserves as stepping stones to cross Africa and find new habitats that provide refuge against extinction (Brown 2013). As can be deduced from the researcher’s comments, climate change has not only vindicated past efforts in wildlife conservation but also demonstrates how such studies suddenly gain additional importance to guide future efforts in adapting to or mitigating climate change impacts on natural systems, including our land uses. They also provide now a multiple array of new opportunities (and funding) for wildlife/biodiversity studies and projects which have been enthusiastically taken up by every old and many new stakeholders. As for the first time in human history these programs and activities have also gained a huge commercial element (with the elevation of CO2 as the global environmental currency) it is justified to say that, with increasing urgency of action to mitigate/adapt to climate change, there will be a massive surge in environmental programs (and funding and action) around wildlife, biodiversity, ecosystem protection, and so on. This process continues to gain momentum and will present the biggest opportunity wildlife conservation has ever had. Much of that, however, will be driven by new markets with their many problems if applied to the environment.

Conclusions I have attempted in Chapter 3 to give a brief overview of the growing list of activities, which have developed around the world to ensure that wildlife remains for the next generations to enjoy and to value for all kinds of reasons and around different value systems. For each of these international activities there are hundreds of programs and tens of thousands of projects around the tropical world supported

The Development of Wildlife Governance, Science, and Management Capacity in. . .

2221

Fig. 4 A generalized and simplified presentation on the changing role of the various actors involved in wildlife management in tropical, “generally” less developed or developing nations which have adopted the general trends in wildlife management. After a dramatic decline of indigenous traditional users, followed by the state, a diverse management environment has developed where national players increasingly compete with well-established international consortiums and where a general trend in protection has greatly increased the role of illegal users. There is, however, also a re-emergence of community and collaborative models, supported by states and international community, in particular in the less commercial fishing sector

by the above (and many other) players. Many of these activities are around its uses; others such as “The Great Ape Project” are about nothing less than human rights for the five species of apes (http://www.utilitarianism.net/singer/by/200605–.htm) with several organizations for each ape (Orangutan Foundation International http:// www.orangutan.org/). Many of the projects realize that wildlife in the tropics must be seen in the context of many communities of forest-dependent people, numbering more than one billion. They also give testimony that most of us see wildlife as humanity’s joint heritage to care for but also as an extremely valuable good of the international community. Also, interest is not only confined to the western world. This is not the case any longer, if ever it was, as countries ranging from China and India to PNG, Costa Rica, Tanzania, Bhutan, Thailand, Nepal, or Mauritius have developed commitment and own conservation approaches. These are in terms of resources, ingenuity, and impact often on par or larger than those from the west. In Brazil, it was not the programs from the “west” which reduced deforestation in the Amazon dramatically between 2004 and 2013 but its own state-of-the-art Satellite Monitoring System along with legislation and law enforcement (Fig. 4).

2222

J. Bauer

If we try to see the development of these various sectors over the past 50 years in concert, something like Fig. 3 might emerge: a diverse, at times contradictory, and often collaborative environment, where the role of the state is in general decline, where there is a multitude of international agreements, and where industry vies with civil society for implementation (and control). While none of all this action seems to have been enough to halt the decline of wildlife and natural ecosystems around the tropical world, a potential “game changer” has arrived with the action around climate change and the creation of forest carbon markets. Before I discuss why climate change has become a “game changer” especially in forestry (Chapter 5), I will, however, introduce a “reality check” (chapter 4). For that I will take a step back and will show why all these efforts might not be enough, even be misdirected, in our global quest to save Earth’s wildlife.

References Bauer JJ, Giles J (2002) Recreational hunting – an international perspective, CRC monograph series. Griffith University, Goldcoast, Queensland, Australia Bauer J, Herr A (2004) Hunting and fishing tourism. In: Higginbottom K (eds) Wildlife tourism. Common Ground, UK, pp 57–78 Bauer J, Birckhead J, Priestley M, Greiner R (2009) Scoping a Feral Animal Control Program – NT northern region (pigs) (report). Natural Resource Management Board, Northern Territory, Australia Brown P (2013) The role of bufferzones. Climate News Network. [email protected], London Dasmann RF (1964) Wildlife biology. Wiley, New York Davies G (2004) Economia: New economic systems to empower people and support the living world, ABC Books, Sydney NSW du Toit JT, Rogers KH, Biggs HC (2003) The kruger experience-ecology and management of savanna heterogeneity. Island press,Washington Freitag-Ronaldson S, Foxcroft LC (2003) Anthropogenic influences at the ecosystem level. In: Du Toit JT, Rogers KH, Biggs HC (eds) The Kruger Experience: ecology and management of savanna heterogeneity. Island Press, Washington, pp 391–421 Lamarque FJ, Anderson R, Fergusson M, Lagrange Y, Osei-Owusu L, Bakker (2009) Humanwildlife conflict in Africa – causes, consequences and management strategies. FAO forestry paper 157. Rome Sakulas HW, Bauer J, Birckhead J (2013) Community participation in biodiversity conservation and development projects: a Papua New Guinean perspective. Environ Papua New Guinea 2(1):1–13 Sullivan R (2005) Rethinking voluntary approaches in environmental policy. Edward Elgar, Cheltenham, UK Tribe A (2001) Status assessment of wildlife tourism in Australia series Wildlife tourism research report series; No. 14. CRC for Sustainable Tourism, Griffith University PMB 50, Gold Coast Mail Centre QLD 9726 Tshering K (2008) Walking the middle path: opportunities for biodiversity conservation and cultural preservation through sustainable tourism development in the protected areas of Bhutan. PhD Thesis, University of Sydney

Modern Adverse Trends Which Affect the Wildlife Management Efforts Johannes Bauer

Contents Introduction: Allow Me to Shake Your Faith! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Poverty, Land, and Wildlife Rights: The Elephant in the Room . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wildlife in the “Awkward Indigenous Space” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The “Tragedy of the Commons” Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ownership as a Neglected Key Factor in Wildlife Management . . . . . . . . . . . . . . . . . . . . . . . . . The “Government’s Animals” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Loss of Wildlife Harvest Traditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Corporate Land Use Shift (CLUS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2224 2225 2226 2227 2228 2229 2231 2233 2235 2236

Abstract

In this chapter I will reach beyond the conventional in wildlife management and ask some inconvenient questions which have plagued wildlife and biodiversity ecologists for some time now, with unfortunately few answers so far. Many of us, as we struggle with Dasmann’s premise, have started to ask these uncomfortable questions about western understanding and scientific concepts as we apply them around the developing and tropical world. We ask questions about the production systems we, and this includes scientists, promote, the governance arrangements we help to put in place, and the stakeholders we support. We know well that we often fail to reach wildlife and wildlife-dependent communities alike. We also know that our favourite systems we like to promote do no work in the real world and that we are losing the middle ground (e.g. wildlife which can be sustainably harvested) of productive and healthy ecosystems. But we also have countries, places and projects where approaches have started again to reflect ethnic and J. Bauer (*) Australian Carbon Co-operative Ltd., Bathurst, NSW, Australia e-mail: [email protected] # Springer-Verlag Berlin Heidelberg 2016 L. Pancel, M. Ko¨hl (eds.), Tropical Forestry Handbook, DOI 10.1007/978-3-642-54601-3_175

2223

2224

J. Bauer

national identities and do things better than western approaches. It will be clear that few of the trends driving the loss of the middle ground I describe in this chapter are reversible in our world. What can happen, however, is that communities and traditional landowners, the real guardians of wildlife, can claw some of the lost middle ground of productive wildlife management back. Keywords

Alternative wildstock-livestock scenarios • Corporate landuse shift (CLUS) • Japanese deer hunting culture • Kangaroo harvest • Land and property rights • Land and Wildlife Ownership • The government’s animals • The new western and scientific commons • Poverty, land and wildlife rights • Loss of wildlife harvest traditions • Dasman’s Premise • The Tragedy of the ‘Good People’

Introduction: Allow Me to Shake Your Faith! In this chapter I will reach beyond the conventional in wildlife management and ask some “inconvenient questions” (with few answers so far) which have plagued wildlife and biodiversity ecologists for some time now. Many of us, as we struggle with Dasmann’s premise, have started to ask these uncomfortable questions about western/northern and scientific concepts as we apply them around the developing and tropical world, the production systems we (and this includes science) promote, the governance arrangements we (help) put in place, and the stakeholders we support. We know well that we often fail to reach wildlife and wildlife-dependent communities alike. We might also suspect that we misapply science with its western contexts and jeopardize the growing efforts around the world to find alternative ways. And we also see too many examples where our growing number of responses is too much costly talk and too little action, most of them deeply compromised by what we fail to act upon. And not unimportantly, we well know that the ways we measure our successes (money spent, papers written, projects “finalized”) are a very poor, often misleading, measure of success. And last but not least, we are surrounded by a growing plethora of “arrangements” where the state fails to regulate where it should (often the corporate sector), yet also prevents markets to develop where it should (the communities) support them. Before I do this, however, it is crucial to frame all that around the “factors” which need to be addressed in the tropics as they develop: poverty, growing inequality, abused women and children, and an increasing industrialization of agriculture, forestry, and fisheries where valuable resources go to the powerful and often corporate, while the dregs are left for communities. Driven by underregulated multinational corporations from the developed world the latter has started to replace what regulation there was and appropriated what was community land and goods. These industries and organizations have been savvy to access and exploit the increasing amounts of money spent by the international community, also for conservation, which they capture amid gaining environmental credibility – and at great profits. Increasingly we see those players in dominant positions at

Modern Adverse Trends Which Affect the Wildlife Management Efforts

2225

international forums, where they can manipulate development agendas, including those of United Nations. Nowhere is this more evident than in climate change action, where the promises and examples of huge carbon profits beckon. I have chosen some examples for this chapter where I believe things are going badly wrong, not because we do not spend enough money or because our premises including those from science are so bad. Many things in matters wildlife go badly wrong because there are trends (ecological, social, economical) at play, poorly recognized and acknowledged, which work against our good intentions, trends which operate at such scales, at times outside of what we can target, at others in places we do not want to touch (land ownership), that our efforts need to be reevaluated and redirected.

Poverty, Land, and Wildlife Rights: The Elephant in the Room Poverty has, since the Brundtland report, been recognized as one of the major causes for environmental (and wildlife) destruction in the Third World. Not so surprisingly this “discovery” has experienced some polarization and a great global debate about whether it should be the markets, free ones preferably, to correct that “trend” (the “trickle-down effect”) or whether there is some intervention required where “the west” stretches its helping hand to wildlife and to disadvantaged people, many of those living with no access to clean water, medical care, sufficient food (see Millennium Development Goals of UN http://www.un.org/millenniumgoals), and are often the hapless victims of political conflict that includes the “landgrabs” of global industries (around mineral, fish, agricultural, soil, and wildlife resources (ivory, pets, rhino horn, shark fins – the list is endless)). In the eyes of two of the world’s leading agronomists (Mazoyer and Roudart 2006) these “trends” are destroying the livelihoods (and rich cultural and often wildlife environments) of two billion farmers. Not a minor matter that, but one which should be cause for alarm. If not all, most wildlife management texts have come from the western world, many of those from the USA, and we have seen how Dasmann’s views have more or less foreshadowed (if they could not fully comprehend the full scale of) the interventions I have described in Chapter 3 from “the more fortunately situated lands.” Conversely, many texts on conservation and sustainability focus on developing nations, the poor countries where inequity and lack of resources prevail, where warfare is ripe, and where the fate of wildlife seems a minor concern – at least to the national leaders. In a book about wildlife management in the often disadvantaged tropics it would therefore seem one has to be very careful to take these differences into account, or in fact, the book has to be written very differently. So what would these differences be? Quite obviously we will have differences which will relate to the culture, to religion, to the general environment (forest, grassland, wetland), to the abundance of particular resources, to the state of the livestock industry, to the range of available options, and to the wealth of the nations. Each of these will have a fundamental role to play in how national conservation and

2226

J. Bauer

wildlife attitudes and forms of management (or the lack thereof) are being applied. More importantly, however, than any of that is, I suggest, the question of land rights, forest rights, and wildlife rights. Only if these are solved can we talk about the responsibilities. As we see even World Bank has recognized that crucial question and discussed its implications, yet again in a ritual of sorts, with a policy paper 2014 (see GLF Committee 2014). This “new” focus on land and property rights, if hardly new, is of great significance. Yet how will it be implemented? As land rights are so closely connected to mineral rights, carbon rights nowadays (and perhaps water and species rights), most governments have tried to keep them away from “landowners,” as they seek their “resource rent,” especially if these have “only” indigenous or “traditional” land tenure. Access and rights to forests and what lives in them (wildlife, NTFPs, biodiversity) are the elephant in that “land rights” room, and it will be essential to make sure that land rights are looked at in the context of wildlife (we have seen that people in Tanzania call wildlife the “government’s animals”; one could say the same thing in China, Bhutan, or here in Australia) and protected area legislation accompanied with greatly improved support to improve access and sustainability of these resources, many of which have moved beyond the reach of communities. Once we assume that land and wildlife rights would be given back to communities we would run into problems. We would realize that wildlife has often become inaccessible to communities, including some 350 million indigenous people for whom it is their “land use,” because of what the west has made out of it, often protected areas or a new “scientific (western) common” where access to wildlife resources is determined by some charity in London or a scientific expert group from Gene`ve. Communities which can recapture their traditional/indigenous uses would need to regain ownership of what has become a “western commons.” This trend has been greatly exacerbated by a growing vast tourism industry, often around wildlife, which has been able to capture the proceeds of much western conservation action, while contributing almost nothing. It has also been exacerbated by the progressive depletion of wildlife on nonprotected land, often overcompensating for lost protected wildlife. And if that should not have been enough this protection includes species which are, as is the case of wild boar in Bhutan or China, at “superabundance,” through the depletion of megapredators, (western) animal protection legislation, and agricultural changes (e.g., Bauer et al. 2005; Boyd et al. 2003).

Wildlife in the “Awkward Indigenous Space” Nowhere are such poorly defined or absent land and wildlife rights more important than in the “awkward indigenous space.” Many indigenous people have not developed what the west has called agriculture and forestry within increasingly modified food and fiber systems. They continue to rely on wildlife to meet their daily needs in sophisticated wildlife use systems. While these systems have been considered inferior as they produce less quantity with more effort they have, because of their reliance on diversity, managed to coexist with what the west has called “wild,”

Modern Adverse Trends Which Affect the Wildlife Management Efforts

2227

“wilderness,” and “wildlife.” More ominously such systems have led to lingering concepts of “TERRA NULLIUS” which have, here in Australia, or in the Amazon, for example, extirpated indigenous land and wildlife rights. The only modern “landuse” – and the west still hesitates to call it that – which the west has been able to develop around wildlife is either recreational fishing and hunting or tourism, mostly once the land has been “put under protection.” Much of the tragedy of the world’s indigenous nations, and the wildlife they depend on, lies in this modern divide where cared for and loved community land has joined the ‘global (scientific, western) commons, that “awkward space” which is at odds with the modern world and with the ways the west has discarded “wildstock” in favor of a selected few (livestock). Much has been written about that (e.g., Birckhead et al. 2000; Sakulas et al. 2013; Bauer and English 2011a, b), and there are even changes. In Canada, New Zealand, and the Australian North (see Northern Land Council and NAILSMA) some displaced indigenous groups have regained their land (or at least some of it) and have started, with support from the (inter)national community and the state, to exercise their land rights and own land use systems which were mostly hunting and fishing. There is a great need that other indigenous people in the tropical world are given these rights also, if only as an act of climate justice while the world implements REDD+.

The “Tragedy of the Commons” Revisited In our relationship to the sea and its many goods, we are currently standing at a crossroad. We can either turn right and go down the path of agricultural production (as we have in our terrestrial environments) and attempt to simplify the sea, replace wild fish life with “domesticated fish,” and try to control the processes around that. We can also turn left and manage marine systems with the better understanding and, more importantly, a more humble attitude toward the complexity of food chains and our (in)ability to manage them “scientifically” (population ecology handles two or three species models well (at least over some years) but fails to predict community ecology). We would do this because the left turn is the sensible way to go if we want to avoid all the terrestrial repercussions of terrestrial agriculture repeated in the sea. One could even argue that while the terrestrial environmental “side effects” we created around agriculture are already taxing us to the limit, the management of marine shifts (which will be the consequence of the right turn) will be well beyond current and future ecological, political, and social management capabilities (even more so with a changing atmospheric and marine climate). Rather than going down the path of aquaculture as response to the depletion of marine environments, as happened in the terrestrial (including freshwater) ones, there is still time to choose the wise path which carefully manages and restores what we still have and, ever so carefully, supplements it, not replaces it, with more sophisticated systems as our understanding grows. Not with some quick fixes we just happen to think of, because we have ruined the other ones. Much of that shift will happen, however, not because of whatever ecological decisions we might make but because of ownership (Fig. 1).

2228

J. Bauer

TERRESTRIAL ECOSYSTEMS

TRENDS

WILD STOCK

LIVESTOCK

MARINE ECOSYSTEMS

CURRENT TRENDS WILD STOCK

LIVE STOCK

ALTERNATIVE SCENARIO WILD STOCK

LIVE STOCK

1950ies

2000

FAIT ACCOMPLY

Fig. 1 Alternative wildstock-livestock scenarios for terrestrial/freshwater and marine production environments

Ownership as a Neglected Key Factor in Wildlife Management Ever since Hardy coined the term “Tragedy of the Commons” there has been a clear and logical economic and social rationale why lack of ownership of a resource, in that case FISH, led to poor governance, management, and eventually depletion. This logic can be applied to the forests in Nepal, which, once taken under the control of the government, away from communities, started to disappear (they had turned into commons where nobody had responsibility and could benefit); to fishes in many ocean zones where, due to lack of legislation, everybody (in particular foreign fishing vessels) tried to catch them before the competitor did; but also to the Queen Alexandra Bird Wing (the largest butterfly in the world) in Oro Province of PNG, which, despite being sold for almost US$ 10,000 a pair (illegally?), disappears because local landowners cannot sell it and breed it in captivity. The reason for that is in Annex 2 of CITES and cannot be sold legally while rainforest land falls to the axe or oil palms. The list of such “worthless” wildlife by national or international decree is endless. Because of this lack of ownership, people do not plant trees (why should they if that just costs money while the future resource value goes somewhere else; moreover, if the government does this they might lose their land once forests are established as many landowners feared in Nepal). In each of these three cases even the economic rationale for community ownership is clear, in fact overwhelming. It is also workable (community forestry in Nepal now thrives; many fisheries under community control are, remain, or have become again highly productive and sustainable) and benefits the wider community instead of often a

Modern Adverse Trends Which Affect the Wildlife Management Efforts

2229

foreign fishing industry or illegal gangs. The Oro birdwing can be very successfully bred in captivity (for sale and reintroduction of wild populations). And perhaps even more importantly, due to this lack of ownership and the capacity to make an income it is the illegal trade (often in the shape of organized crime which thrives with that incentive) which finds such legislation an incentive while for the wider good population base it becomes impossible to develop their own management (which would drive the criminal gangs out especially if supported by police). This “tragedy of the good people” has now become firmly entrenched around a growing number of species as they join CITES Annexes.

The “Government’s Animals” Another form of “tragedy of the commons” occurs when the government seeks to search its resource rent at the expense of communities and landowners. Wildlife in western (and in particular Anglo-Saxon legislation on which much postcolonial legislation of tropical countries in now modeled), whether on private or on public (protected), land often belongs to the State. While the intent of that legal step was often the protection of wildlife that purpose does not work so well any longer. For many communities it has led to a “western or scientific (science often supports that) commons,” where wildlife has become either of no interest to the community (landowner) any longer or, worse, is now only accessible to either illegal markets, to the “nonconsumptive use” of tourism (which rarely pays a resource rent), industrial bioprospectors (which do neither mostly), or industries (often foreign with government support) which were able to develop in that “commons” (common in marine fisheries). An ambiguous role is played by science which mostly works against communities and landowners as only the state and industry can afford it. To me the most telling case of such an unjust and divisive system is the NSW kangaroo harvest where, with the support of government, a corporate kangaroo harvesting industry with only several large companies has been able to appropriate the (state-owned?) kangaroos on private land, with landowners receiving nothing. Much has been written about that deeply flawed and divisive system, few independent (not industry-paid) scientists would defend it, yet it has been able, because of its (paid by the industry?) government and scientific support (policies, legislation), to persist and even grow (e.g., Grigg 1988; Bauer and English 2011a, b). Thanks to “its scientists” it is even flouted in international and misguided circles as “best management practice.” In this case “kangaroo rights” have been denied to landowners by the state, yet given to an industry which uses its clout (including its ability to employ scientists (landowners cannot afford them)) to exclude landowners to what grows on their land yet could be a major resource for them (e.g., Grigg 1988; Bauer and English 2011a, b). This remains so, even after serious concerns about the management itself (scientific data collection and analysis, population estimates) have emerged (Mjadwesch 2011). Similar trends are evident in many parts of the tropical world with valuable wildlife resources. Such wildlife, in Tanzania called “the Government’s animals,” might benefit from protection because of tourism and legislation

2230

J. Bauer

or through scientific harvesting plans (developed by the scientists for the industry). It has, however, lost its value to landowners which can (and will) plan around and against it. In most cases this is more adverse to wildlife than the lack of a scientific study. As we can see in the case study below this is not restricted to developing countries or wildlife harvest but also the case in Australia where communities are excluded from participating in species rehabilitation programs by government agencies. Today, wildlife often remains a responsibility for the state and its agencies (“agency animals”) which have insufficient resources – and intent (many contradictory interests) – to manage it, while others (tourism, for example) pay little or nothing. Landowners, indigenous and farmers alike, but also those who would seek legitimate income, and a vast number of people with goodwill, remain excluded. In this no-win situation the gap between reality and intent grows wider by the day. Wildlife remains without value to most, landowners are deprived of an income source, the state has a responsibility it cannot carry, the public is excluded with all its concerns (and resources) – and wildlife declines. Although one could argue that NGOs have broken that culture of nonparticipation, the reality is that many government arrangements of this type suit them all too well (having gained many similarities with corporate industry) in protecting their own role, incomes, and resources.

The Bridled Nailtail Wallaby, Brush-Tailed Bettong, Bilby, and Hairy-Nosed Wombat as Doomed “Agency Animals”

Some years ago I was involved in an (unsuccessful) reintroduction of nailtail wallabies (NTW) in NSW, Australia. The bridled nailtail wallaby (left), believed extinct, was rediscovered in 1972 at the town of Dingo on a private property in Queensland. The property owner was bought out by the government (some 12,000 ha), a national park established at the site, and a NTW Recovery Plan written by the country’s experts (Lundie-Jenkins and Lowry 2005). Some 30 years later the population continues to decline, there are few more than the 300 animals left then, while Western Plains Zoo gave up its very easy and successful breeding because lion and Black rhino programs were more popular and lucrative (government support, visitor dollars, prestige). For landowners (supported by WPZ) breeding of the above endangered native species (and others) would have been easy. Landowners could have been guaranteed a market for surplus animals and for a fraction of the costs. Viable populations of many thousands of animals could have been established on many private properties with the (happy) property owners deriving income from their management (and from tourism ventures). Nothing happened. The state insisted that only IT could manage endangered species with everybody being losers (Based on Bauer and Cameron 2001; Bauer et al. 2002; King 2006) (Fig. 2).

Modern Adverse Trends Which Affect the Wildlife Management Efforts

2231

Fig. 2 The captive breeding release of endangered species in Australia (From left to right: bridled nailtail wallaby, brush-tailed bettong, bilby, and hairy-nosed wombat) is now as before something dear to the public. Easily done (from a professional point of view especially if supported by zoo infrastructure and staff) yet generally unsuccessful from saving a species. The exclusion of communities, landowners and the private sector from conservation work is one of the major reasons why this is the case

The Loss of Wildlife Harvest Traditions The decline of traditional and indigenous harvest systems, as has happened around the world in countless examples over the past century, has been, and this is often overlooked, synonymous and often causative to the erosion and loss of traditional and indigenous knowledge systems around wildlife and natural systems in particular. The Anglo-Saxon distinction between protection and production, distributed around the world first through colonialism, later through the environmental movement with its many agencies and projects around a growing number of protected areas, has affected many traditional wildlife harvest management systems in particular in the nondeveloped tropical world. Here, where seemingly no organized groups were around (recreational hunters, for example) to stop that (indigenous people also come to mind) these systems proliferated and progressively displaced indigenous people and their land use. There are countless stories of this disownment around the tropical world. While often unable to protect their target species in these projects, local, traditional, and indigenous harvesting systems, often sustained over centuries previously (and well capable of “modernization”), were destroyed along with the land use, culture, lifestyle, and value systems. Local wildlife knowledge and harvest systems often decline together, to be replaced by a value vacuum around wildlife which is in nobody’s interest. I have given two examples from my own experience. The decline of indigenous/traditional knowledge and harvest systems in Wuyishan Biosphere, Fujian Province, China (see Boyd et al. 2003), and the current deer management dilemma in Japan, also based on the loss of a hunting culture along with modernity. Both show a modern stalemate in wildlife management and utilization which works against local communities and wildlife.

2232

J. Bauer

The Loss of Japanese Deer Hunting Culture

In February 2011 I was invited with a colleague to offer advice in a deer management project between a university and Japanese land use authorities. Over the 10 days we spent in Japan we gained a unique perspective of a deer management problem which seemed as intriguing as it was absurd. It showed a modern Japanese society which had, while continuing to insist on wild harvested whale meat, lost its taste for millions of native deer, Cervus nippon, which lived and multiplied in its forests (which they damaged greatly) and mountains where only few were prepared to hunt it, sell it, and, above all, eat it. To change this problem and perception among the public, the Japanese government was far from idle, and our first “lesson” consisted in the attendance of a deer preparation ritual by one of Japan’s most famous chefs of French cuisine (as applied in Japan, French cuisine enjoys a very high reputation). In order to promote venison consumption, he instructed the Japanese public in a TV show how to butcher the deer and how to cook it, following a French recipe. While my colleague and I were watching in fascination we were served beautifully prepared little parcels of deer sushi, obviously catering for the more traditional-minded of the hundred or so attending journalists. Afterward we went for a tour around Japan, mindfully interrupted at times with a French deer meal, to show us the extent of the problem. What we learnt was that, along with parts of Europe and the USA, deer had in Japan also benefited from modern forestry with its abundance of cover and food, the cessation of agriculture in other parts of the landscape, and a legislative environment which did not seem to encourage hunting. There had also been a very pronounced trend among young Japanese not to join hunting clubs and the declining fraternity of hunters. It also became clear that our Japanese friends were at a loss what to do about it. There did not seem to be an attempt (as we first thought) to farm deer (misguided in any way as it would have been a distraction from the management of the wild population, essential to reduce their impact on forest regrowth and quality); our friends much rather thought that one could catch deer in large numbers, to kill – and bury – them. There was little thought so it seemed to us who would do that (nothing less than a national hunting system) and that it would have to involve firearms and a large number of trained hunters – judging from the size of Japan, half a million at least. While this deer overpopulation seemed to be a grave problem across much of Japan, there were attempts, for example, in Hokkaido, to develop hunting tourism around Japanese deer, as well as some other unique and abundant species such as the Japanese serow, but nothing which was suitable to effectively reduce deer number and revive the use of a healthy, humane, and, for many, delicious supply of prime meat (I would estimate that the national harvest could be in excess of one million animals (~40,000 t of prime meat). A final visit to Nara, famous for its huge wooden (continued)

Modern Adverse Trends Which Affect the Wildlife Management Efforts

2233

temple and its resident population of deer (some 7000 + which share roads, parks, and sometimes restaurants with the local populace and the many Japanese tourists who visit for that very reason) seemed only fitting to impress on us the complexity of this modern relationship with deer which has developed in modern Japan. Japan is not alone with that problem. It is shared by an increasing part of the western world and more and more countries in the tropical world also, where western value systems, conservation legislation, and protected area systems have driven hunting – the harvest of wildlife – more or less into the underground as an activity, not a legitimate land use any longer, condemned to be legally treated often as “bushmeat” while western meat, cattle and sheep (for which huge areas of forests are being cleared), are being offered as THE alternative. As we can see from this example this vacuum in wildlife use (including fisheries) through (the prevented) lack of ownership but also an inability to modernize (resources, lack of science, etc.), for example, in agriculture, plays into the hands of a land use shift away from traditional owners toward those who can afford it, and who have the support and power from governments. It is not a minor shift but one which currently transforms the tropical world.

The Corporate Land Use Shift (CLUS) What I have called “The Corporate Land Use Shift (CLUS)” is about the acquisition of increasing tracts of valuable land and crops by increasingly large, technologically highly advanced corporate and multinational entities, including countries now (also see landgrab). This process, as described by Mazoyer and Roudart (2006), is destroying the livelihood of farmers around the world and is leading to the homogenization of agricultural production in terms of wildlife/biodiversity loss, see Figs. 4 and 5. After the great diminishment of wildlife and ecosystems a new phase has commenced where much traditional agricultural land is bought by foreigners, multinational firms, but also countries (such as China) which want to secure food production. The direct/indirect/cumulative impacts of this new wave of land use change are mostly unknown, yet the combination of pesticide and GM typical for these systems will lead to a second wave of diminishment of wildlife which will make irreversible past losses and negate what might be gained in conservation efforts. Whether in forestry or in hydrodevelopment or agriculture the corporate approach comes at a large cost which is paid by the public, the rural communities, and ecosystems with their wildlife and their many unaccounted essential environmental services these provide for free for the thankless users. The most astonishing feature of that way of “managing” the environment is the sheer fact that the enabler and indeed power behind much of that approach is provided by SCIENCE. Not so

2234

J. Bauer

Fig. 3 A Japanese lady posing with one of the 7000+ Japanese deer (Cervus Nippon) which roam the streets, parks – and templesof the historic city of Nara in Japan

Fig. 4 A global process, the acquisition of the valuable assets of communities, often through mining and oil exploitation, landgrabs, and valuable wildlife

Modern Adverse Trends Which Affect the Wildlife Management Efforts Natural Ecosystems

The two main steps of wildlife/biodiversity loss during agricultural & forestry development in the tropics

Wildlife/ Biodiversity Hunting/ fishing/ gathering

STEP I Logging/ Clearing Colonial hunting

2235

Traditional Agricultural Landscapes Wildlife/ Biodiversity

STEP II Monocultures/ Pesticides/GM

Before and after Colonialism (