Ecology and Management of Forest Soils [5 ed.] 9781119455714, 1119455715

Citation preview

Ecology and Management of Forest Soils

Ecology and Management of Forest Soils Fifth Edition

Dan Binkley Northern Arizona University Flagstaff, Arizona

Richard F. Fisher

This edition first published 2020 © 2020 John Wiley & Sons Ltd. 4e: Binkley & Fisher, 9780470979471, December 2012 Wiley‐Blackwell) All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Dan Binkley and Richard F. Fisher to be identified as the authors of this work has been asserted in accordance with law. Registered Office(s) John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Office The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging‐in‐Publication Data Names: Binkley, Dan, author. | Fisher, Richard F., author. Title: Ecology and management of forest soils / Dan Binkley, Northern Arizona University, Flagstaff, Arizona, Richard F. Fisher. Description: Fifth edition. | Hoboken, NJ : Wiley, [2019] | Includes bibliographical references and index. | Identifiers: LCCN 2019000082 (print) | LCCN 2019000419 (ebook) | ISBN 9781119455714 (Adobe PDF) | ISBN 9781119455721 (ePub) | ISBN 9781119455653 (paperback) Subjects: LCSH: Forest soils. | Soil ecology. | Soil management. | Forest management. Classification: LCC SD390 (ebook) | LCC SD390 .F56 2019 (print) | DDC 577.5/7–dc23 LC record available at https://lccn.loc.gov/2019000082 Cover Design: Wiley Cover Image: © Photos from Allan Bacon Set in 9/12pt Meridien by SPi Global, Pondicherry, India Printed in the UK by Bell & Bain Ltd, Glasgow 10 9 8 7 6 5 4 3 2 1

Contents Preface, vii Dedication, ix In Memoriam: Richard F. Fisher,  xi

9 Tree Species Influences,  205

1 Soil Foundations,  1

11 Soil Management: Harvesting, Site Preparation, Conversion and Drainage,  265

2 Forest Soils Across Space and Time,  13

10 Characterizing Soils Across Space and Time, 237

3 Minerals in Forest Soils,  45

12 Fire Influences,  293

4 Organic Matter in Forest Soils,  59

13 Nutrition Management,  319

5 Physics in Forest Soils: Soil Structure, Water, and Temperature, 83

14 Managing Forest Soils for Carbon Sequestration, 351

6 Life in Forest Soils,  109

15 Evidence‐Based Approaches,  373

7 Chemistry of Soil Surfaces and Solutions,  139 8 Biogeochemistry, 163

References,  393 Index,  435

v

Preface The sheer volume of information on forest soils has probably increased by at least 10‐fold since the year 2000. Much of the information added new numbers to our understanding of the typical rates of soil processes. Other changes have been more fundamental, such as the shift in the conceptualization of how organic matter is sequestered in soils: long‐lasting soil organic matter does not depend as much on chemical recalcitrance as on intricate soil environments and ecological interactions (including aggregate formation). A textbook cannot provide a systematic review for all the important topics addressed; authors need to pick and choose examples that have strong educational value. Compiling examples into representative tables and histograms is more powerful than any single example, but our compilations are intended to illustrate commonly observed situations rather than provide reliable conclusions about the population of all forest soils. This edition presents a large amount of new information, taking advantage of the opportunities presented by color illustration. This edition retained about two‐thirds of the literature used in the previous edition. The older case studies and research papers often provide as strong a foundation for

learning about forest soils are new studies. More importantly, historical studies have great value for insights about how our understanding of forest soils developed from our community of inquisitive colleagues over the decades. We’re sorry that a textbook format did not allow us to include all the great research from colleagues around the world. We hope the stories and examples we included form a foundation for building new understanding from additional study of the expanding worldwide literature. Forests and soils are whole systems, so it’s very arbitrary to pull out any single feature, such as soil chemistry, to examine in a single chapter when it’s relevant to every other chapter. For this reason, we developed chapter topics to be general themes, with diverse soil subjects entering each chapter, and with each topic arising in more than one chapter. For example, a chapter on soil organic matter relies, in part, on the subjects of soil chemistry and soil biota, but it also overlaps somewhat with a chapter on carbon sequestration. We hope readers will find the weaving of topics within and among the chapters to be helpful in stitching together a coherent, holistic understanding of forest soils. Dan Binkley

vii

Dedication

This book is dedicated to three groups of colleagues, and especially to those with membership in all three: • pioneering scientists who give us new ways to see inside forest soils; • persistent investigators who create and maintain the long‐term research sites (and databases) that

are so fundamental to seeing forest soils change; and • generous organizers of workshops and field trips that sustain the forest soils community we all cherish. Our friend, Bob Powers (1939–2013) is at the top of this Honor Roll.

ix

In Memoriam

Richard F. Fisher, Jr. 1941–2012

Dick Fisher grew up in Urbana, Illinois, where his ancestors were clever enough to realize that t­ reeless prairie soils could make great farm soils. He combined his undergraduate forestry major with a minor in the history and philosophy of science (B.Sc. 1964); this broad and deep curiosity about how things connect, and how things got to be the way they are, characterized Dick’s work with forest soils. He worked with Earl Stone at Cornell University for his PhD in soil science (1968), and again he picked up additional insights by minoring

in chemistry and geomorphology. Dr. Fisher then began a series of academic adventures, first an assistant professor at his alma mater, the University of Illinois, from 1969–1972. He moved to the University of Toronto as an associate professor (1972–1977), and then joined the University of Florida as a professor and the Director of the Cooperative Research in the Forest Fertilization program (1977–1982). Dick next headed west, to become a professor and head of the Department of Forest Science at Utah State University (1982–1990).

xi

xii

In Memoriam

His final academic positions were at Texas A&M University (1990–1999), where he was a p ­ rofessor, the head of the Forest Science Department, and the director of the Institute of Renewable Natural Resources. Retirement from A&M freed Dick’s time to dive into applied aspects of forest soils, which he supported by working as the director of research and development for Temple‐Inland Forest Products Corporation in Dibol, Texas (1999–2006). Dick’s career included a wide and deep set of contributions to his profession, students, and colleagues. He was an instructor in the Organization

for Tropical Studies field courses in Central America from 1970 through 1999; some experiments he established in Costa Rica are continuing to advance our understanding of the connections between tree species and soils. He served both the Soil Science Society of America (including chairing the Forest, Range and Wildland Soils Division) and the Society of American Foresters (he was elected a Fellow of the SAF). Dick authored and coauthored over 100 refereed publications, co‐authored three editions of this book, and served as co‐editor‐in‐chief of the journal Forest Ecology and Management for 18 years.

C ha pte r 1

Soil Foundations

OVERVIEW The forests of the world are shaped by the soils that support them, and the soils are in turn shaped by the trees. Soils change substantially across ­landscapes, in response to changes in the parent materials in which soils form, to differences in water flow, and the responses of plants and the rest of the ecosystem biota. Differences in soils lead to large differences in the species composition and growth rates of forests, largely mediated by the supply of water, nutrients, and sometimes oxygen in the soils. One of the most characteristic features of forest soils is the O horizon, the commonly occurring uppermost soil layer com­ posed of fresh litter, decaying ­litter, roots, and soil organisms. The tapestry of our current understanding of forest soil comes from threads that reach far back into nineteenth century, with great detail added from the advances of science in the twentieth century. The broad variety of soils and forests across dimensions of space and time can be investigated with three stand­ ard questions: what’s up with this soil, how did it get that way, and what will it likely be in the future?

EVERYTHING IN FORESTS AND FORESTRY CONNECTS TO SOILS A century ago, Aaltonen (1919) published a classic assessment of Finnish forests, and he concluded that understory regeneration of pine trees was

l­imited when overstory trees were present. The apparent inhibition of small trees by large trees extended too far away from the large trees for the suppression to be about shading, and Aaltonen concluded that the key issue must be about compe­ tition for soil resources rather than for light (Figure 1.1). If Aaltonen was right, then trenching small plots to cut off competition for soil water and nutrients with surrounding trees should stimulate the regenerating trees. This is exactly what hap­ pened in tests around the world: removing below­ ground competition for water and nutrients leads to large increases in growth of understory plants even when shaded by healthy overstory canopies. Competition for light is of course important for plants too, but the classic forestry idea of “shade tolerance” is often as much about soil resources as about light (Coomes and Grubb 2000). Ancient ideas about single resources limiting the growth of trees need to be left in the past. The production of most plants and ecosystems is limited by more than one resource, and almost always limited by one or more resources found in soils.

FOREST SOILS HERE ARE DIFFERENT FROM OVER THERE Across Union County, South Carolina in the south­ eastern United States, plantations of loblolly pine differ in growth rates by more than twofold. About 15% of the land can grow less than 5 m3 ha−1 year−1

Ecology and Management of Forest Soils, Fifth Edition. Dan Binkley and Richard F. Fisher. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

1

2

Chapter 1

Aaltonen’s 1919 representation of pine regeneration in Northern Finland:

Experimentation in a pine stand in North Carolina,1930s:

Experimentation in Bavaria in a Norway spruce forest

Figure 1.1  The regeneration of pine seedlings in Finland was inhibited by mature trees, at distances that were too great

to be competition for light. Aaltonen (1919, after the description from Kuuluvainen and Ylläsjärvi 2011) concluded that the issue must be competition for soil resources. Experimentation beneath full canopies of loblolly and shortleaf pine trees in Duke Forest showed that removing competition for soil resources by trenching plots led to huge growth responses by understory herbs and tree seedlings (center left, before trenching, center right, four years after trenching; after Korstian and Coile 1938). Shade is much deeper under mature canopies of Norway spruce, giving the impression there is only enough light to support an understory of mosses. However, trenching a small plot within the dark stand of Norway spruce in Bavaria, Germany, led to prolific growth of understory vascular plants (Source: lower photo, from Christian Ammer), just like the classic study of Fabricius (1927) who found trenching under a spruce stand more than tripled growth of understory oak and hornbeam seedlings.

Soil Foundations

Percentage of area

25 20 15 10 5 0

3.8

4.6

5.3

6.1

6.9

7.6

8.4

9.1

9.9 >10.8

Loblolly pine productivity (m3 ha–1 yr–1)

Figure 1.2 Across 100,000 ha of Union County, South Carolina in the southeastern United States, the growth rate (productivity) of wood in loblolly pine plantations differ by more than twofold. The climate is the same for all locations, so the differences in growth rate result from differences in soils (Source: from data of the Natural Resource Conservation Service (2017)).

of wood, and about 15% can grow more than 9 m3  ha−1  year−1 (Figure  1.2). The climate doesn’t vary across the County, so what drives the differ­ ences in ecosystem productivity? Of course, soils provide the key explanation. The soils across the County vary in capabilities of supplying water and nutrients to support tree growth. Some of these variations result from factors that influence soil development over very long periods of time, such as the original parent material and position on slopes (near the top where water and nutrients are removed, or near the bottom where water and nutrients accumulate). As with most aspects of soils, changes in some factors (such as parent ­material or water supply) cascade with follow‐on changes in soil chemistry, the biotic community, and the growth of plants.

CURRENT FOREST GROWTH IS LIMITED BY SOIL RESOURCES ACROSS REGIONS Just as productivity differs across locations, the productivity within a single location may shift ­ upward or downward if soil conditions change.

3

At  the sub‐continent scale of the southeastern United States, the growth of loblolly pine planta­ tions differs by more than twofold from one region to another (Figure  1.3). The spatial pattern of growth differences connects to temporal patterns too: the ability of soils to support tree growth can change over time, in response to active or passive management. Estimates of the climate‐determined potential productivity across the region can be compared with the observed rates, and the differ­ ences between these two suggest that soil‐related factors restrain growth by 15–20% across the region. Silvicultural practices can move forest growth toward the climatic potential, or indeed reduce growth further if done poorly.

SOILS SHAPE FORESTS, AND FOREST SHAPE SOILS The two forests in Figure 1.4 are growing only 50 m apart, along the Mendocino Coast in California, USA. The tall forest has fast‐growing Douglas‐fir, redwood and Bishop pine trees. The short forest has slender stems of Bishop pine, Bolander pine, and pygmy cypress. What ecological factors explain such huge differences in vegetation? Differences in weather and climate would not be important, as the forests experience the same conditions. Some other possibilities that might come to mind include forest age (perhaps one type of forest is much older), the influence of recent fires (perhaps the short forest was a tall one before a fire?), or a legacy of logging. Given the focus of this book, it’s not sur­ prising that the key differences are in the soils. The tall forest grows on a deep soil with good properties for supplying oxygen and nutrients to the trees. The short forest is on a soil with a cemented layer that restricts water movement, and the high water table impairs the supply of oxygen for roots and the soil biotic community (Figure 1.5). But why are the soils so different? Have the soils been developing for different periods of time, or did they develop from different initial rock types? Or is it possible that they could be the same age, developed

4

Chapter 1

(m3 ha–1 at 25 years) 0 – 300 300 – 400 400 – 500 500 – 600 600 +

(m3 ha–1 at 25 years) 0 – 300 300 – 400 400 – 500 500 – 600 600 +

Figure 1.3  The accumulation of stemwood in loblolly pine plantations in the Southeastern United States ranges from

about 300 to 600 m3 ha−1 of wood volume at age 25 (upper graph). If soil fertility did not limit growth, many locations would show much higher productivity (middle graph). Across much of the region, current limits on soil fertility reduce forest stem biomass by about 60–90 m3 ha−1 (a reduction of 15–20%; lower graph). These graphs based on illustrative simulations using the 3PG model, provided by Jose Alvarez‐Munoz.

Soil Foundations

5

(m3 ha–1 at 25 years) 0 – 60 60 – 70 70 – 80 80 – 90 90 +

Figure 1.3 (Continued)

Figure 1.4  These two forests are part of the Ecological Staircase in the Jughandle State Park along the California coast. Differences in soils lead to differences in vegetation and tree sizes (Jenny 1973; Binkley 2015; Source: photo on left from Ron Amundson, on right from Sarah Bisbing).

6

Chapter 1

Figure 1.5  The soils under the tall trees (left) is well drained, with good supplies of oxygen for tree roots and the soil

biotic community. The soil that supports only short trees (right) developed an iron‐cemented layer that keep water perched high in the profile, restricting oxygen, root growth, and nutrient cycling rates (Jenny 1973; Binkley 2015; Source: photos by Ron Amundson).

from the same initial materials, and some other fac­ tor explains these differences? The development of soils always involves complex, interacting processes. An understanding of chemistry, physics, biology and ecology is required to understand how soils develop in different ways, and how these differences shape forests. In this case, the key feature that led to such different sorts of soils was the location relative to the edge of a terrace; water drained away better the loca­ tions supporting tall trees, and water flow was more restricted back from the terrace edge where only the short trees can grow. The comparison of these two adjacent sites also illustrates a key feature of forest soils, and a key

challenge in studying these systems. As the forests and soils began developing, the current rate of input of organic matter and leaching of minerals was prob­ ably rather similar. Only small differences in the transport of iron would have occurred in any single year. Yet the cumulative outcome of so many years with minor differences was the vastly different soils we see today. This theme comes up in other aspects of the study of forest soils. A short‐term study that lasts two or five years on the rates of decomposition of litter might conclude that the litter of decomposes more quickly at one site than another. This might be interesting, but the long‐term development of the soil would depend on the very small fraction of the

Soil Foundations litter material that enters long‐term pools that are protected from decomposition. We would not have a reason to expect (or evidence to support) that the initial rates of decomposition have any relationship to the important processes that determine long‐term C accumulation. Rates of rapid processes may be easy to measure, but important changes in soils gen­ erally develop from slow processes that have little effect in any single year. This book explores the fascinating array of the minerals, organic matter, and organisms that inhabit and shape soils, providing a foundation for understanding the forests that grow in theses soils – and how forests continue to shape soils.

SOILS ARE BOTH THE MOTHERS AND THE CHILDREN OF FORESTS Forests change at rates that are fast enough for us to see, over the course of a decade or a few decades. Soils change at about the same rate, but the changes aren’t so visible to our eyes. The changes in these components of the whole forest ecosystem result from mutual interactions between trees and the soils that support them. New forests commonly establish on soils that supported many generations of forests in the past. The soils provide reservoirs of water that trees depend on between storms, and stocks of all the major nutrients that form the building blocks of the trees’ biochemicals. While the soils are supporting the trees, the trees are changing the soils in subtle and dramatic ways. Trees pump high‐energy organic compounds into the soil, from both roots and aboveground tissues. They can also pump nutrients from one location in the soil to another, changing the nutrient profiles both laterally and vertically. The organic compounds (organic matter) fuel the phenomenal soil community of animals and microbes that transform the composition and structure of the soil. Understanding forest trees requires insights about soils, and understanding soils requires insights about trees and the rest of the biotic communities.

7

In the broadest sense, a forest soil is any soil that has developed primarily under the influence of trees. Trees may sink roots deeper than other types of plants, and they produce large amounts of complex organic compounds that take a long time to decompose. A key feature of most forest soils is the development of a layer of organic material at the top: the O horizon. Forest soils cover more than a third of the Earth’s land area, though in many locations the forests were removed and the soils converted to uses. Perhaps as much as one‐third of the Earth’s former forest soils are now devoted to ­agricultural, urban, or industrial use. Soils that develop in forests often differ strikingly from other types of ecosystems. Soils that are too dry to support trees typically lack a consistent O horizon, have high concentrations of salts, and hori­ zons cemented by calcium carbonate. Agricultural soils are routinely managed by plowing, homoge­ nizing the upper soil and removing the legacies of three‐dimensional structure that characterize forest soils. Some grasslands are present in areas that are wet enough to support trees, but fires or other pro­ cesses support the dominance of grasses. Grassland ecosystems tend to support d ­ ifferent sorts of biotic communities that lead to different soil composition and structure, including a top horizon rich in both organic and mineral materials. Many formerly forested soils are now in other land uses, and most of the soils that remain under forests have been modified to some extent by humans over recent millennia. The production of food has relied upon soils that were formerly under trees and grasses. The clearing of forests histori­ cally entailed “slash and burn” land use. Trees were felled and burned to provide fertile, weed‐ free fields for food crops. After several years, the invasion of other plant species (weeds) lowered the production of food, leading to abandonment and regrowth of trees. With the development of urban societies, soils could be harnessed perma­ nently for agriculture, as available labor could ­control weeds (and provide some nutrition man­ agement by recycling organic wastes). In more

8

Chapter 1

recent times, industrialization around the world consumed vast quantities of wood for fuel and building material. Harvested forests typically regrew, though in some notable situations the removal of trees led to long‐term deforestation. Wood from forests provided the fuel to develop civilization, and rates of wood removal for fuel and wood products did not plateau until the 1990s, when worldwide wood use stabilized at about 3.5 billion m3 ha−1 year−1 (Sutton 2014). When the first edition of this book appeared in the 1970s, more than half of that wood supply was taken from “natural” forests, where direct forest management was limited. Planted forests now provide over half of the annual wood harvested from the world’s for­ ests, as the area of plantations increased almost 50%. The most intensively managed forests account for just 1.5% of all forest area, but the soils of these plantations produce one‐third of the wood used for paper and solid wood products (INDUFOR 2012; Binkley et al. 2017). Humans have influenced forest soils even in some situations where harvesting has not occurred. Most forests experience fires, mild or severe, on time scales of a few years to several centuries or more. Humans have altered the occurrence of fires in many ways, including changing the structure of fuels in forests (by intensive grazing), and by providing the spark to light fires (more often than lightning does). Other human‐related influence on unharvested forests may be as large as the changes in fires. Tree species have been introduced into forests from around the globe, but even more important has been the spread of tree‐damaging insects and pathogens. The major tree species in the eastern part of North America was the American chestnut, but the exotic chestnut blight essentially removed chestnuts from the continent. American elm trees were codominant across much of the same region, but they have been largely removed by exotic Dutch elm disease. Human‐induced changes in the atmosphere also influence forests. Most forests of the planet receive much more nitrogen in rainfall than in historical times, and increased supplies of this often‐limiting nutrient can spur the growth of some trees at the

expense of others. Climbing concentrations of car­ bon dioxide in the atmosphere are warming the planet and shifting precipitation patterns. All these human‐related changes in forests and forest soils may seem to paint a dire future. Fortunately, both forests and forest soils are remarkably dynamic and resilient, and both will continue to support the planet’s future, though future forests and soils will not be the same as those in past eras.

FOREST SOILS DIFFER IN MANY WAYS FROM CULTIVATED SOILS Soils are much more than just the basement of for­ ests, and more than a provider of water and nutri­ ents. Soils are dynamic systems comprised of thousands‐to‐millions of types of organisms, taking in energy and matter and releasing heat, gasses, and chemical byproducts that remain in the soil. These dynamic systems exist in four dimensions: the usual two dimensions associated with the Earth’s surface, the depth below the surface, and the dimension of time over which soil‐shaping processes occur. The distinctions between wildland forest soils and soils that have been strongly influenced by cultiva­ tion might be illustrated best with a pair of pictures (Figure 1.6). The legacies of agricultural cropping of soils go far beyond issues of maintaining soil organic matter and nutrients. Forest soils typically show strong structural differences across small distances, as long‐term patterns of soil roots, animal burrowing, and water percolation develop cumulative legacies that may become larger, perhaps being reset partially when soils are stirred by the blowdown of trees with large root systems (for example, Figure 6.2).

FOREST SOIL SCIENCE IS AS OLD AS SOIL SCIENCE ITSELF The circular nature of the interactions between trees and soils may seem obvious today, but insights about these intimate interactions were rather slow in

Soil Foundations

Figure 1.6 These soil profiles come from pits that are only 25 m apart in the Duke Forest in North Carolina, USA. A few centuries ago, the profiles would have been the same. Both were cleared and cropped with corn and tobacco in the 1700s and 1800s. The soil on the right was planted with pine trees about 1900, and the soil on the left was maintained in grass/forb vegetation (as part of a powerline right of way). The pine trees restored the 3‐dimensional structure of the soil, including legacies of root channels. Areas with higher concentrations of oxy­ gen in the soil atmosphere are orange, and those with lower concentrations are gray. Channels formed by roots can provide for rapid transport of both air, water, and dis­ solved chemicals, influencing soil chemistry and ecology at scales of millimeters to tens of meters (Source: photo from Allan Bacon).

­eveloping over past centuries. Early insights into d soils would have largely been based on empirical experiences with little ability to explain why the fertil­ ity of one soil exceeded another. Early natural ­philosophers in the West who pondered the connec­ tions between plants and soils included Aristotle (384–322  bce), Theophrastus (372–297  bce), Cato (234–149 bce), Varo (116–27 bce), and Virgil (70–19 bce). Pliny the Elder (23–79  ce) gave the most complete account of the ancients’ understanding of soil as a

9

medium for plant growth, including crops and forests. He chronicled folk wisdom and other empirically derived information. Many farmers recognized that certain crops were more productive on certain sites or soils, and that legumes may improve soil fertility. Not much insight about soils developed over the next mil­ lennium in the West. Scientific approaches finally began to take seed, and in 1563 Bernard de Palissy published a treatise On Various Salts in Agriculture, in which he stated that soil is the source of mineral nutrients for plants. However, his work had little impact on other natural philosophers. The period from 1630 to 1750 saw the great search for the “principle of vegetation.” During this time, natural philosophers conceived of matter as com­ prised of four (or five) “elements”: fire, air, water, and earth (and perhaps niter [potassium nitrate]). Van Helmont (1577–1644) conducted a classic experi­ ment with a willow (Salix) tree: he grew 164 pounds of willow tree in 200 pounds of soil in a pot, and at the end of the experiment he measured that the soil in the pot had lost only 2 oz. Since only water had been added during the experiment, he concluded that the 164 pounds of willow had come from water alone. Van Helmont may not have been the first to conduct this experiment and draw the wrong con­ clusion. A similar experiment was conducted and similar conclusions were also drawn by Nicholas of Cusa (1401–1446). Robert Boyle later repeated the experiment with Cucurbita, obtained similar results, and reached the same mistaken conclusion. Despite the apparently definitive experiments of Van Helmont and Boyle, the quest for the principle of vegetation continued. John Woodward (1699) grew Mentha in rainwater, in water from the River Thames water, in sewage Hyde Park, or in sewage mixed with soil. He found that the plants grew ­better as the amount of “sediment” increased, and concluded that “certain peculiar terrestrial matter” was the principle of growth. Frances Home experi­ mented and reached similar conclusions in the 1750s. Home went on to note that “exhausted soils recovered from exposure to the air alone” and therefore concluded that the air must be the ­ultimate source of the essential materials.

10

Chapter 1

It was not until 1804 that de Saussure success­ fully explained the origin of the material compris­ ing plants. He found that most of the mass of the plant was carbon derived from the carbon dioxide of the air. Interestingly, we have found in our teaching that fewer than half of the undergraduate and graduate students taking basic and advanced courses in ecology understand that the majority of the mass of plants derives from the air, not from the soil or water! The work of de Saussure and Boussingault in France in the early nineteenth century began a period of rapid scientific advancement. In 1840, Justus von Liebig in Germany published Chemistry Applied to Agriculture and Physiology, and modern soil science began. Liebig helped dispel the theory that plants obtained their carbon from the soil and developed the concept that mineral elements from the soil and added manure are essential for plant growth. However, Liebig erroneously believed that plants received their nitrogen from the air rather than from within soils (as demonstrated by de Saussure). Lawes and Gilbert put these European theories to test at the now‐famous Rothamsted Experiment Station in England and found them to be generally sound. Following the work of these chemists, scientists in many different fields including geology (Doku­ chaev, Hilgard, Glinka), microbiology (Beijerinck, Winogradsky), and forestry (Grebe, Ebermayer, Muller, Gedroiz) contributed to the development of what today is termed “soil science.” Most of the early research on soils was directed to its use for agricultural purposes because Europe had a critical food shortage. The importance of soils to forest ecosystems was recognized by several early scientists. In 1840, Grebe, a German forester, stated: “In short, almost all of the forest characteristics depend on the soil, and hence, intelligent silviculture can only be based upon a careful study of the site conditions.” Pfeil echoed this thought in 1860, and it became a cen­ tral theme of European forestry (Fernow 1907). Gilbert (1877) examined how soils developed in relation to geology across landscapes that varied in

parent material and topography. Hilgard (1906) recognized a relationship between vegetation and soils in North America similar to the relationship that had been noted in Europe. The work of Grebe, Pfeil, Hilgard, and others, perhaps unintentionally, laid the foundations of forest ecology. In North America, early ecologists such as Merriam (1898), Cowles (1899), and Clements (1916) knew that soil was important in vegetation dynamics, but none of them understood soils well. Toumey (1916) noted the importance of soils in American silviculture for the first time. Early research in forest soils was dominated by basic scientific studies. In fact, some very important early soil science research was done on forest soils. Ebermayer’s work on forest litter and soil organic matter (1876) had a strong influence on soil science and agriculture. Muller’s work on humus forms (1878) marked the beginning of the study of soil biology and biochemistry. Gedroiz (1912) pio­ neered work on soil colloids, suggesting that soils performed important exchange reactions, and laid the groundwork for modern soil chemistry. These scientists were interested not only in soil properties, but also in the processes that led to the existence of the properties. Scientific curiosity about forest soils developed more application late in the nineteenth century as wood production became a pressing problem in Europe, and the restoration of degraded forestlands abandoned after agriculture became a necessity in North America. Forest soils research in the twenti­ eth century was dominated by studies focusing on species selection for reforestation, methods of site preparation for reforestation, methods for estimat­ ing the productive capacity of forest lands, tree nutrition and response to fertilization, and soil changes under intensive forest management. Late in the twentieth century, new interests developed in the areas of impact of deposition of acidity and nutrients from air pollution, and sustainability of forests facing intensive harvesting (for biofuels and short‐rotation forestry). Education about forest soils developed in parallel with advances in scientific understanding. In

Soil Foundations Germany, H. Cotta emphasized that soils were a key to forest production as early as 1809. Forestry textbooks began to include information on soils, including Grebe (1852) and B. Cotta (1852). The first text devoted to forest soils was published in 1893 by Raman, and the first French text by Henry (1908), with continuing production of forest soils texts across Europe (including Remezov and Progrebnyak 1965). Forest soil science developed more slowly in North America, perhaps of the lack of concern was because forests seemed to be inexhaustible. Only after World War I did the ideas of managing selected forests for sustained yields, reforestation of abandoned farmlands, and the establishment of shelterbelts in the Midwest begin to take hold. The first soils textbooks to appear in North America came along only after World War II, from Wilde (1946 and 1958) and Lutz and Chandler (1946). Coile (1952) provided an overview of the connection between soil properties and tree growth (for more background on the develop­ ment of forest soils in North America, see Gessel and Harrison 1995). A hiatus in forest soils texts in North America lasted until the late 1970s, when Armson (1977) published a text in Canada, and Pritchett (1979) published the first edition of this text.

THREE FUNDAMENTAL QUESTIONS APPLY TO EVERY FOREST SOIL Each forest soil is a unique, incredibly complex sys­ tem. Indeed, the idea of any single, discrete forest soil is an oversimplification of the scales of space and time where many rapid and slow processes shape and reshape the soil system. This complexity may be daunting, and we certainly cannot under­ stand all the intricacies of forest soils. Fortunately we can examine and understand many key aspects, with a combination of clear thinking and appropri­ ate measurement techniques. Three questions might be helpful when trying to understand the soil for any particular forest or location within a forest:

11

1. What’s up with this forest soil? This question leads us to characterize the structure of the soil, the biotic community shaping the soil, and the chemistry of interactions. 2. How did this forest soil get that way? All forest soils have developed over varying periods of time, under the influence of many processes. Some processes and factors may be much more important than others in determining the devel­ opment of soils over short terms and very long periods. Identifying the important factors that shaped the history of a soil may provide the foundation for understanding the soil’s current condition and launch into the third question. 3. What’s next for this forest soil? Sustaining or enhancing the ability of soils to grow plants is a nearly universal goal for people who work in forestry (and agriculture). The future trajectory of a soil depends on its current state (Question 1), and insights about the soils development (Question 2) are crucial. The future of a forest soil depends very heavily on what events and impacts will happen, and these are especially important for managed forests.

IDEAS, MEASUREMENTS, AND CLEAR THINKING ABOUT EVIDENCE CAN ANSWER THE THREE QUESTIONS The application of the three questions will always entail a combination of ideas we have learned from other people and other soils, and direct information about each soil. Some of these ideas are likely to be fairly solid representations of how the world really works, and as in all sciences some ideas are likely to fail to capture the real story. Similarly, some of the “direct” information we have about a soil may be biased (for example, emphasizing the top‐most soil and missing key features deeper below the surface) or extrapolated imperfectly (such as applying a single soil texture across a hectare, glossing over variations within the hectare). The most robust insights and

12

Chapter 1

management of forest soils may depend on taking an “evidence‐based” approach. Evidence‐based approaches (whether formal or informal) carefully spell out how much an idea is based on reasoning and how much on the results of measurements and experiments. These approaches also explicitly rec­ ognize that evidence and insight gained for one soil or landscape will not necessarily fit perfectly with

another soil or landscape. General knowledge of forest soils can be applied most usefully to specific forest soils when the strength of evidence is articu­ lated clearly. The following chapters develop the ideas needed to understand the composition and processes that comprise forest soils, and how evidence‐based approaches continually improve ­ our insights about fascinating forest soils.

C ha pte r 2

Forest Soils Across Space and Time

OVERVIEW Soils development differs with climate, topography and parent material. These driving variables influence the composition and processes in soils, and these within‐soil state variables interact as soils influence organisms and organisms influence soils. All of these features take place across the dimensions of space and time. Some features of soils change slowly, such as the mineralogy of soil clays, and some change more rapidly, such as organic matter pools. These changes in time are matched by changes in space; moving across landscapes entails moving from one soil pedon or tessera to another, along with subtle or major changes in soils. Around the forests of the world, some generalizations are possible, but not many. On average, soils are more fertile at the bottom of slopes than on ridges, and tropical soils are deeper and store more C than temperate or boreal forest soils. Variations within regions are typically larger than general differences in the averages across regions. Local soil details are always important.

SOILS ARE REAL, BUT WE DEFINE THE SOILS WE STUDY A forest soil exists in a very real location at a very real time. When we come to ask questions about soils, such as how fertile soils may be, or how soils respond

to land use practices, we start by ­acknowledging that the “soil” we study is something of an abstract idea as well as a tangible object. The complex composition and processes of forest soils take place over a great range of dimensions of space and time. At the scale of a single bacterium (about 1–10 μm), the “soil” is a very tiny neighborhood indeed. Within that neighborhood, there are chemical gradients of oxygen, acidity, and nutrients, as well as exposed surfaces of minerals and organic matter and enclosed spaces within aggregates (peds). The rates of change in the tiny regions can be quick for some processes, with some ions moving from the adsorption to a surface to dissolved in soil solution to uptake by a bacterium in a matter of seconds (Figure 2.1). Soils are of course the foundations of forests. They provide sustained supplies of water, nutrients, and physical anchoring to allow trees to withstand the forces of deploying massive canopy sails high in the air. Soils develop from parent materials of widely varying mineral composition arrayed across landscapes, under the wide range of climatic drivers. The importance of these scales of space and time become even more intertwined across longer distances and times. The “scale” for the engagement of a fine root of a tree with the soil is much larger than tiny neighborhood of a bacterium, and the time frame is likely much longer (months to years). Issues of space are also complicated by the fact that the interface of the root and the soil is heavily influenced by the connections to the tree’s canopy, which can be 10 or even 100 m distant. The canopy

Ecology and Management of Forest Soils, Fifth Edition. Dan Binkley and Richard F. Fisher. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

13

14

Chapter 2

(a) Models

Data nm

m

100 m

Mm

Models Data sec

day

year

century

Millennium

(b)

Figure 2.1  The composition and dynamics of forest soils vary across orders of magnitude of space and time, and the most

powerful scale for investigating soils always depends on the questions being asked. Measurements to obtain data are straightforward at some scales, but insights are limited (depending on hopefully useful model representations) at other scales (Source: Paul 2015; Hinckley et al. 2015, used with permission).

Forest Soils Across Space and Time of course provides the carbon compounds that are both the building blocks and energy source for the fine root. The canopy also provides the opportunity for soil water to be “pulled” into the fine root. Leaves exposed to dry air can develop very low water potentials that cause water to be pulled to the roots from the surrounding soil. Scaling up to the size and lifespan of trees brings in a large array of considerations, including browsing of plants by animals (and the unavoidable production of animal waste), the processing of dead organic matter by the soil biotic community (recycling nutrients), and perhaps major events such as intense windstorms and wildfires. The soil beneath trees is very real, but all of these aspects of dynamics in time and space show why our study of soils needs to be carefully defined. If the boundaries of a defined study system of soils are drawn poorly, the risks are high that important driving factors will be left outside the system, limiting the insights that can be pulled from the study.

15

IMPORTANT CHANGES IN SOILS OFTEN OCCUR IN AREAS SMALLER THAN FOREST STANDS Forest management has classically been practiced with the idea of a stand, with a stand defined as an area that is more similar in species composition, tree size, and age than areas beyond the boundary of a stand. The idea of a stand might apply to some unmanaged forests, where changes in time since a major event, or soil type, lead to apparent natural units of ecosystems. Soils often change substantially over notably smaller areas the scale of common forest management. For example, this plantation of Eucalyptus trees in Brazil covered an area of dozens of hectares, but important soil differences gave more than a twofold difference in the growth of trees from the upper slopes to lower slopes (Figure 2.2). Changes along a hillslope may seem obvious, but spatial changes in soils are common in almost all situations, including many soils

Figure 2.2  Over a distance of a few dozen m, the soils differed by more than twofold in ability to grow Eucalyptus trees.

The arrow indicates the size of the seven‐year‐old trees on the lower slope, and the same‐size arrow contrasts with the size of trees higher on the slope (Source: photo provided by J. L. Gonçalves and J. L. Stape, used with permission).

Chapter 2

O horizon C (kg m–2)

16

2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

2.0 1.5 1.0 0.5 0.0 0

0–15 soil C (kg m–2)

0 10 10 20 20 30 Xd 30 40 ista m) ( 40 e nce 50 50 tanc 60 60 (m Y dis )

3.5 3.0 2.5 2.0 1.5 0

X

10

dis

20 10 0 30 20 tan 40 30 ce 40 ) 50 m 50 e( (m tanc ) 60 60 Y dis

3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8

Figure 2.3  The soil in a 40‐year‐old, 0.24 ha plot of white oak trees in Texas was very flat, but minor differences in drain-

age of this wet site led to substantial variation in the C accumulated in the O horizon and the 0–15 cm A horizon of the mineral soil (Source: data analysis provided by Suzanne Boyden).

that are apparently flat and uniform. A white oak forest with no slope showed twofold differences in soil C of twofold over distances of a few meters (Figure 2.3; see Chapter 10).

Driving variables:

State variables:

Climate

Vegetation, animals, microbes

Parent material

DRIVING VARIABLES CAUSE SOILS TO DIFFER ACROSS DIMENSIONS OF SPACE AND TIME Differences in structure and processes in forest soils can develop from differences in driving variables. A driving variable (or factor) affects the state of a complex system (such as a soil), but the variable itself doesn’t change as a result of the new state of the system. Driving variables for soils include climate, parent material, topographic features (landscape position, slope and aspect), and land use (Figure  2.4). A dry climate leads to very different sorts of soil development than occurs in wet climates, and Figure 2.2 illustrates how landscape (or

Topography

Soil structure, chemistry, fertility

Dimensions through which ecosystems develop:

Time, space

Land use

Figure 2.4  The various components of soils (state variables) develop from the influences of driving variables, and from interactions with other state variables. The interactions occur over the dimension of time and space, and overall outcomes of these interactions differ from one location to another, and from one point in time to another.

topographic) position can drive differences in soil development. The characteristics of soils can be considered state variables. A state variable changes in response to driving variables, and also in response to interactions

Forest Soils Across Space and Time with other state variables. The future condition of a state variable results from past legacies of driving variables and interactions with state variables. Key state variables in soils are the biotic community, the chemistry of soil solutions, and the 3‐dimensional structure of the soil profile. Soil development occurs from the influence of driving variables on state variables, as well as the mutual interactions among state variables. All of these dynamics require the dimension of time; short dimensions of time do not allow the types of soil development that can occur over very long periods. Time is not a driving variable or a state variable; time is a dimension like space. We would not refer to “space” as driving soil development, though soil development may vary across a spatial gradient for many reasons. For the same reason, it would not be very useful to refer to time as “driving” soil development, though of course many factors do require time to be able to change soils. This “systems” view of explaining differences among soils developed from a rich history of soil science over the past one‐and‐a‐half centuries. In 1883, the Russian geologist Vasily Dokuchaev remarked with enthusiasm: The still young discipline of these relations is of an exceptional inspiring scientific interest and meaning. Each year it makes greater and greater strides and conquests; gains daily more and more of active and energetic followers, eager to devote themselves to its study with the passionate love and enthusiasm of adepts. (quoted by Jenny 1961a)

Dokuchaev developed ideas about how various factors would interact to develop the soils found across landscapes and regions, and developed an equation for soil formation:

S

f cl, o, p to



where S = soil, cl = climate, o = organisms, p = geologic substrate, and to is a measure of relative age (this version of the equation was related by Jenny 1961b). A somewhat similar approach was ­proposed

17

independently by an American soils professor, Charles Shaw (1930):

S

M C V T D



where S = soil, M = parent material, C = climate, V  =  vegetation, T  =  time and D  =  deposition or erosion. Does it make sense to put these ideas into the format of an equation? To a mathematician, an equation is created to be parameterized and solved. Hans Jenny (1941) popularized these ideas in the form of an equation that he hoped could someday be solved quantitatively. Jenny’s approach assumed that soil‐forming factors worked independently, and could be represented with a set of independent (non‐interacting) equations. No one ever solved this equation, and it’s clear that interactions are fundamental between factors in the equations. So was it a good idea to try and describe the incredible complexity of soil formation in a simple equation? What numbers could be plugged in to represent red pine trees versus red maple trees, or the species identity of earthworms? What units would there be for the dependent variable, “soils”? The equation idea would be a poor one if people were encouraged to view soils as being simple, linear systems. The idea could be good, however, if it stimulated thinking on soil formation and how to deal quantitatively with soil formation (Bryant and Arnold 1994). Jenny (1958, 1980) also proposed the idea of a forest “tessera” as a useful mental model for thinking about how soils (and forests) develop. A tessera is a small stone used to make mosaic pictures; each stone is separate, but in combination creates the whole image. A tessera for Jenny was something of a 3‐dimentional system of soil and vegetation that interacted strongly (Figure  2.5); moving one or several meters away one would enter a new tessera with its own internal interactions. Water and materials would flow between adjacent tesserae, but interactions between tesserae would be less important than those within. Landscapes would be comprised of thousands or millions of tesserae. This

18

Chapter 2

CLIMATE IS NOT THE MOST IMPORTANT SOIL FORMING FACTOR

Figure 2.5  Each of these three polygons would represent

a forest + soil tessara (Jenny 1958, 1980), where interactions within the unit are greater than interactions across boundaries with the adjacent tessarae. The tesserae would be 3‐dimensional, covering an area of ground reaching as high as the trees and as deeply belowground as the soil reaches. The tessera can also be 4‐dimensional when changes over time are included (a “tesseract” is a 3‐ dimensional representation of a 4‐dimensional cube).

mental model may be useful for thinking about interactions that occur within one location in a landscape, and how that location connects with flows of water and materials to the rest of the landscape. The dynamics of tesserae through time provides a 4‐dimensional way to think about soils and how they change. The tessera idea may not translate to actual forests and soils in a concrete way, as the real world may not have the sort of discrete boundaries that are so evident in artistic mosaics built of tesserae.

Back in the nineteenth and early twentieth centuries, soil scientists were relatively free to speculate about soils and the processes shaping them. In the absence of strong evidence and experiments, there was no way to rule out any conceptualization that was logically plausible. Domains without much evidence can be fertile fields for imagination and endless debate. The idea of “zonal” soils was widely accepted (and debated): soils generally come into equilibrium with the climate, because climate’s influences inexorably drove soil formation from diverse initial states to climatically determined equilibrium states. For example, some soil scientists thought that grassland soils in the middle of the North American continent were zonal, in equilibrium with the climate. Hans Jenny was skeptical, noting that the soils were only 10,000 years old, and he wondered what the soils would be like in another 100,000 years (Jenny 1989). The idea of climate dominance and the image of zonal soils was logically plausible, and without quantitative evidence the idea could not be supported or refuted with confidence. Widespread quantitative evidence became available in the middle of the twentieth century, providing a solid basis for comparing ideas of soil patterns over time and across landscapes (Cline 1961; Richter and Yaalon 2011). We now have the benefit of another half‐ century of evidence in the form of huge databases that allow us to test classic (and new) ideas about soils. For example, we can say with very high confidence the average carbon content of the O horizon plus 0–100 cm depth of forest soils across Europe is precisely 130 Mg ha−1, based on almost 5,000 locations (De Vos et  al. 2015). Does climate explain most of the variation around this average? No. Statistical analyses showed somewhat strong relations between soil C accumulation and humus form, tree species, taxonomic classification of the soils, and parent material. The statistical insight available by knowing the precipitation and ­temperature of a forest soil was quite small.

Forest Soils Across Space and Time

SOIL‐FORMING PROCESSES OPERATE SIMULTANEOUSLY AT VARYING RATES Soils develop as plants add organic material, as weathering processes transform rocks, stones, and finer mineral grains, and as transport process remove and add materials. The initial mineralogy is important, and the rate of change as well as the accumulation of byproducts depends heavily on climate and the ecology of soil organisms. Soils typically develop vertical layers that differ in structure, mineral composition, chemistry, and biology. Sometimes the horizons are horizontally rather uniform across forest stands, but often the variation in depth and continuity is dramatic. These soil profiles also show the influences in many locations of multiple parent materials; the dynamic surface of the Earth includes landslides, floods, and other mass‐movement events that lead to more than one type of parent material for many forest soils.

How fast does soil formation operate? The answer of course is specific to each location of the Earth’s surface, and specific to each feature of soil development. The carbon and nitrogen contents of forest soils often increase dramatically over periods of a few centuries or millennia from the time of initial development of parent material, and other features such as the gradual removal of metals present in the original minerals take tens or hundreds of millennia to develop. A general approach to defining a timespan of soil development might be the time it takes for bedrock to become soil‐like in texture such that it becomes eligible for water‐transport erosion. One such estimate of “soil production” comes from an escarpment in a landscape formed on granitic parent materials in southeastern Australia. The age of materials near the Earth surface can be estimated using isotopes produced by cosmic rays bombarding nuclei. The patterns of 10Be and 26Al indicated that weathering slows rapidly with increasing depth of the soil, from about 50 mm in 1,000 years with little soil present, to less than 10 mm in a 1,000 years when soil accumulation has reached 80 cm (Heimsath et al. 2000, Figure 2.6).

80

Soil production rate (mm 1000 yr–1)

So if evidence does not indicate that climate is the dominant soil forming factor, which factor merits the title of being most important? This is an example of a question that sounds sensible, but has no sensible answer. As described in this chapter and through the book, the influence of parent material never seems to weather away completely. The presence (or absence) of particular tree species clearly alters soils in major ways, at times scales of decades. Soils are often different across topographic gradients, but “topography” is really just a collective word for a diverse set of interacting processes (such as hydrology) and components (changing parent material) that generate differences without leading to simple patterns. Soils are always changing so time is always important, but changes over time do not necessarily erase the influences of processes. Figure 2.4 gives an abstract representation of soils as complex systems with lots of boxes, arrows, and interactions, and simple answers from complex systems usually arise only in extreme conditions (where one feature manages to be of overriding influence).

19

70 60 50 40 30 20 10 0

0

20

40

60

80

100

Soil depth (cm)

Figure 2.6 The rate of soil production declined as the depth of the accumulated soil increased across a landscape in southeastern Australia (Source: from data of Heimsath et al. 2000).

Chapter 2

Oak-hickory forest

S

OR

UWR ~50m

0

Depth below soil surface (m)

The vertical structure of forest soils most commonly includes five major types of horizons. The top layer is usually an O horizon, composed primarily of recent plant detritus from leaves (and often roots), and sometimes physically mixed by soil animals with some mineral soil from below. The next layer (A horizon) often may be high in mineral content but enriched in organic matter, either from material directly deposited as plant detritus, transferred by soil animals, or precipitated from compounds dissolved in the O horizon above. Below the A horizon, some soils have a zone where organic matter or clays have been depleted (moved lower into the profile) in a process called eluviation, giving the designation E (or Ae) horizon. As water moves farther into the profile, materials (organic compounds, iron, aluminum, and clay particles) may be deposited in a B horizon. Deeper yet into the profile, underlying parent material may remain substantially unaffected by the soil forming process, with a designation of a C horizon. There are many, many variations on the characteristics of these horizons, with dozens of names for classification purposes (van Breemen and Buuman 2002; Schaetzl and Anderson 2005; Buol et aI. 2011). The idea that the C horizon is largely unaffected by biological and other soil forming factors makes sense from the context of a current forest. From the perspective of long‐term soil development, however, the “subsoil” actually undergoes substantial alteration as a result of biological processes (Richter and Markewitz 1995), physical strain patterns with depth and across landscapes (St. Clair et al. 2015), and chemical weathering of parent material (Oh and Richter 2005). The loss of silica from three soil series in North Carolina, United States, was substantial, with 10–80% of the silica lost in the upper part of the subsoil (Figure 2.7). The loss of calcium in the subsoil was even more extensive, with the bedrock for two of the soil series losing most of the Ca in the top 6 m. Trees that send roots deep into the C horizon and underlying saprolite (weathering bedrock) may tap reserves of water and nutrients beyond what is present in the “soil” ­

–2 –4 –6 –8

Tarrus series (phyllite) Cecil series (granitic gneiss) Enon series (diabase)

–10 –12 –14 –16 –18 –100

–80

–60

–40

–20

0

–20

0

Silica lost (% of original)

0 Depth below soil surface (m)

20

–2 –4 –6 –8 –10 –12 –14 –16 –18 –100

Tarrus series (phyllite) Cecil series (granitic gneiss) Enon series (diabase)

–80

–60

–40

Calcium lost (% of original)

Figure 2.7  Weathering of minerals in soils occurs both in the soil (A, B, and C horizons) and in the subsoil (saprolite, weathering bedrock). The picture shows a granitic gneiss profile with a Cecil soil developing at the top (S is saprolite, OR is oxidizing/weathering bedrock, and UWR is unweathered bedrock). The graphs show the loss of silica (middle) and calcium (lower) from the original parent material of phyllite, granitic gneiss, or diabase (Source: after Oh and Richter (2005); picture used with permission of Elsevier).

Forest Soils Across Space and Time (see  Callesen et  al. 2016, Chapters 5, 15). The expanded view of connecting soils and the weathering material beneath them is termed the “critical zone” and is an area of substantial research interest (Richter and Billings 2015).

SOME SOILS HAVE BIRTHDAYS AND DEATHDAYS, BUT MOST JUST KEEP FORMING AND ERODING The idea of soils changing over time has fascinated soil scientists for at least a century and a half. A wide variety of studies have tackled this idea by finding soils that have beginning dates, such as those following volcanic eruptions, glacial retreat, and terrace deposition along rivers. It also makes sense that soils can be extinguished, with a death date set by being buried by new volcanic eruptions, and by mass wasting that moves what was once soil downhill to become sediment in a new system. Research on “birthday” soils has shown than soil development from bare rock to well‐ developed soil may take more than 1,000 years, though development can be much faster when the parent material has fine texture as in flood deposits, and moraines formed below glaciers. The time required to develop a modal Hapludalf (gray‐ brown podzolic) soil was estimated to be more than 1,000 years after deglaciation in the northeastern United States. Development of distinctive profile characteristics can be extremely slow in arid regions and in some local areas where drainage or other conditions may slow the weathering

21

processes. In short, there is no absolute time scale for soil development (Table 2.1). The extent of research on the development of soils with birthdays doesn’t match the extent of these soils on the Earth’s surface. Most forest soils do not have birthdays, and soil changes reflect long‐term, on‐going process of addition of new material (from geologic uplift, transport from uphill locations, and wind‐blown material), the transformation of parent material, and the loss of mass from chemical dissolution and erosion. The dynamics of soils in a hilly area near the coast of Oregon in the United States (Anderson et al. 2002) gives some insights about what may go on with forest soils occurring across much of the planet. The bedrock of graywacke sandstone rises at a rate of about 10 mm every 1,000 years, and overall removal of material from the landscape (denudation) roughly matches the rate of rise. Soil formation includes the interacting array of processes described in this chapter, including chemical weathering of the minerals. About half of the mineral weathering in the system occurs in the soil (0.7 m thick), and about half in the weathering bedrock of the subsoil (about 5 m thick; deeper rocks show no sign of weathering). The soil has a bulk density (mass per unit volume) that is about half that of the bedrock, so the transformation of rock into soil essentially doubles the volume of the original material. The well‐formed soil has lost about 50% of the Ca originally in the rock, along with about 25% of the Mg and Na. The loss of these elements (and Si) adds up to a loss of about 100 kg m−2 of mass as the soil forms. The ­weathering

Table 2.1  Timescales for changes in aspects of soils.

Rapid

Dynamic, 1–10 years

Slowly dynamic (decades‐century)

Persistent (millennia)

Water content Redox potential Nutrient supply

Carbon pools Nutrient pools Acidity

Iron and aluminum oxides Clay accumulation Occlusion of organic matter and nutrients in aggregates (occlusion)

Texture Rock volume Cemented soil layers

Temperature

Bulk density, porosity

Source: Modified from Richter (2007).

22

Chapter 2

period for the 5  m of bedrock lasts about 50,000 years, and then the material is close to the surface and transformed into soil that lasts about 2,000–4,000 years. The material in the original bedrock spends about 20 times longer as weathering bedrock than it does as soil. Physical erosion processes are responsible for a large portion of the denudation of the landscape. Overall, the surface of the Earth rises and is lowered by erosion, and rocks weather into soils through the interactions of biology, chemistry and physics; and then soil erodes at rates that are generally similar to the rates of uplift. The Oregon soils described above have an average “lifespan” of 2,000–4,000  years, but no birthdays. Another example from a hillslope in southeastern Australia showed that the combination of weathering and transport led to soil lifespans (or residence time) of about 20,000 years on ridges and 35,000 years on slopes. The longer residence time on slopes results from slower rates of soil production coupled with thicker accumulation of soil. Where topography is too flat to allow erosion through water movement or mass flow, the “lifespan” of soils can be millions of years. Bacon et al. (2012) estimated that the lifespan of a soil on flat topography in the Calhoun Experimental Forest in South Carolina, United States. The Ultisol was about 6 m deep, underlain by 24 m of weathering saprolite atop granite gneiss. Denudation lowers the landscape by about 10–20 m in a million years (unless matched by uplift), and the residence time (lifespan) of the soil is on the order of 1–3 million years. Such ancient soils have experienced millions of years’ worth of change, yet still support productive forests as a result of a wide range of processes that can sustain soil fertility (uplift, dust deposition, and others).

OVERALL, SOIL FORMATION HAS FOUR COMPONENTS An overall description of soil formation (at the development of horizons) includes aspects of:

1. addition of organic and mineral materials to the soil as solids, liquids, and gases; 2. losses of these materials from a given horizon or from the entire soil profile; 3. movement (translocation) of materials from one point to another in the soil; or 4. transformations of mineral and organic matter within the soil. Some scientists might subdivide these into more classes, perhaps as many as 17 processes that account for soil change (Bockheim and Gennadiyev 2000, 2009). Additions to the soil occur in both organic and inorganic forms. Wind and rain bring in dust particles and aerosols, sometimes from continents across oceans. These additions are generally small on an annual timescale, but may be fundamentally important for sustaining soil fertility over very long periods of time. For example, the Hawaiian Islands in the middle of the Pacific Ocean may receive less dust than almost anywhere else on Earth, yet dust inputs can replace a third to a half of the material lost from the weathering of parent material (Porder et  al. 2007). Forests in the Sierra Nevada of California receive more P annually from wind‐blown dust than from weathering, and about one‐third of the dust comes all the way from Asia (Aciego et al. 2017). Some of the world’s most productive forests occur in warm, sunny regions where abundant water near streams and rivers supports tree growth. These “riparian” settings are often hotspots of deposition of sediments from upstream landscapes, and erosion during later floods. Riparian soils may be very dynamic on the timescales of three generations. The 250‐year‐old cottonwood tree in Figure 2.8 established on flood sediments that were more than 1 m below the current terrace level. Subsequent meandering of the river then eroded the terrace (and toppling the tree into the river) at the same time that new sediments were being deposited on the sandbar across the river. Organic matter is added to soils continually, as roots pump organic sugars and acids into the ­rhizosphere (the soil immediately next to roots),

Forest Soils Across Space and Time

23

Figure 2.8  Repeated flooding of riparian forests (left) leads to accumulation of sediments that raise the height of the soil

surface above the level where trees originally establish. This 250‐year‐old Fremont cottonwood (right) is about a meter in diameter, and it established on flood deposits that were more than a meter lower than the current surface of the soil (evidenced by the proliferation of lateral roots that formed in what was the original A horizon). Multiple floods deposited new sediments over the following centuries (with deposition increased by the “drag” of tree stems on the velocity of water flow). Meandering of the Yampa River then cut into the terrace (while depositing new sediments on the other side of the river), eventually toppling the tree into the river.

and intermittently as roots, leaves, and entire plants die. In some cases, soil animals mix surface horizons thoroughly, bringing organic materials from aboveground and belowground plant tissues into direct contact with soil minerals. In other cases, O horizons remain more distinct from underlying mineral horizons, but even in these cases about 10–20% of the material added to the O horizon may be dissolved and transported into mineral horizons (Qualls et al. 1991). Most of this input of organic matter occurs near the top of the soil, so this is also the zone of maximum biological activity driven by the opportunities for animals and microbes to combine high‐energy carbon compounds with oxygen to release energy, carbon dioxide, and other lower‐energy compounds. Losses of materials from horizons involve oxidation of organic compounds (forming carbon dioxide), and leaching of soluble carbon compounds

and mobile inorganic ions such as calcium, magnesium, potassium, and sulfate. Surface erosion can also include the physical removal of soil material, by water or wind, landslides, and other forms of “mass wasting” (Reiners and Driese 2004). Dissolved losses of materials are routine in soils whenever there is enough water available to leach the soil. Surface erosion is largely unimportant in forest soils until fires, slope failures, or earthquakes bring sudden changes. Primary minerals in rocks form under conditions of high temperature and pressure in the Earth’s crust, and when volcanic eruptions where hot material rapidly cool and solidify. The minerals are largely unstable in the low‐pressure environment of the Earth’s surface, where exposure to oxygen, water, and acid leaching combine to weather the primary minerals into dissolved ions and secondary minerals. Long‐term soil development involves

24

Chapter 2

weathering of primary minerals into lower‐energy compounds, and the products of weathering may accumulate in soils (as secondary minerals) or be leached from the soil. The major set of processes involved in complete losses of material from primary minerals is desilication (or ferralization). Warm areas with high rainfall can have rapid dissolution of silica from soil minerals, and fast loss of silica (and other elements such as potassium, magnesium, and calcium) leading to residual soils that are largely the residual iron and aluminum sesquioxides (the name comes from the molecules containing 1.5 oxygen atoms for each iron or aluminum atom). In a well‐drained environment, this process leads to the formation of an oxic horizon, a zone of low‐activity clay lacking in weatherable primary minerals, resistant to further alteration (Figure 2.9). In a soil with a fluctuating water table, this process leads to the formation of plinthite, which may occur as scattered cells of material or may form a continuous zone in the soil. When exposed to repeated wetting and drying, plinthite becomes indurated to ironstone, which has also been termed laterite (Alexander and Cady 1962; Soil Survey Staff 1999; Eswaran et al. 2003). The dominant transport processes within forest soils result from (i) the direct input of organic materials from roots within soils; (ii) physical movement of material by animals; (iii) compounds suspended or dissolved in water moving through soils; and (iv) major events such as uprooting of trees. Eluviation is the movement of material out of a portion of a soil profile, playing a major role in the A and E (or Ae) horizons. The materials being eluviated can include soluble organic compounds, iron and aluminum bound with organic compounds, and relatively mobile cations such as potassium. The opposite process, illuviation, is the movement of material into a portion of a soil profile. Illuviation is a key aspect of the forming of B horizons in forests, as changes in soil structure or chemical environment lead to the deposition or precipitation of materials from higher in the profile. B horizons are often distinguished by an accumulation of organic matter, clay, and iron and aluminum.

Figure 2.9 Very long periods of soil development (100,000 years to more than 1,000,000 years) in warm, wet environments can lead to the weathering and leaching of almost all primary soil minerals, leaving behind a soil matrix composed mostly of aluminum and iron sesquioxides as in this Oxisol from Brazil. Despite the ancient age of this soil, it may be very productive for agriculture and forestry when fertilized.

A major set of processes dealing with redistribution of mineral materials in forest soils is podzolization (Figure 2.10). Podzolization is a set of processes that brings about the chemical migration of iron and aluminum (typically in association with organic compounds). The mobilized iron and aluminum may not go far, as changes in the B horizon promote dissolution and precipitation of organic compounds, iron, and aluminum. Lundström et  al. 2000 summarized the podzolization process: 1. Soil biota (particularly ectomycorrhizal fungi) release organic acids; citric, shikimic, oxalic and fumaric acids that comprise about 1–5% of dissolved organic C in soil solution, accounting for up to 15% of solution acidity;

Forest Soils Across Space and Time 2. Organic acids weather soil minerals, and released iron and aluminum are transported out of the upper soil (E horizon), bound with soluble organic molecules; 3. Organic molecules are biologically oxidized in the B horizon, leaving iron and aluminum to precipitate on surfaces. The net effect of podzolization is an E horizon that retains high amounts of silica, and B horizons high in sesquioxides (Stobbe and Wright 1959; Ponomareva 1964; Brown 1995). Podzolization is a major soil‐forming process in the boreal climatic zone, but it also occurs in warm areas. The southeastern coastal plain of North America has extensive areas of aquods (groundwa-

25

ter podzols), and “giant podzols” with extremely thick E and B horizons occurring in the tropics (Van der Eyk 1957; Klinge 1976). The processes leading to podzolization are intensified by certain types of vegetation, especially hemlocks (Tsuga), spruces (Picea), pines (Pinus), and heath (Calluna vulgaris) in northern latitudes and kauri (Agathis australis) in New Zealand, but it is important to note that podzolization can also occur under hardwood forests. Organic compounds leached from the foliage and litter (Bloomfield 1954) or produced during litter and soil organic matter decomposition (Hintikka and Näykki 1968; Fisher 1972) lead to rapid mineral weathering (Herbauts 1982) and translocation of organic complexes downward in the soil profile (Ugolini et al. 1977; van Breemen et al. 2000b).

Figure 2.10  A podzolized soil beneath a balsam fir stand in the Montmorency Forest in Quebec, Canada. On the right side, the O horizon grades abruptly into a gray horizon where the iron and aluminum have been leached from the sandy parent material by dissolved organic acids (forming an Ae, or an E horizon). Precipitation of organic matter is evident at the top of the B horizon that is enriched in iron and aluminum. To the left side of the picture, the windthrow of a tree partially mixed the soil, burying the former E (or Ae) horizon with material from the B horizon, with a new E horizon beginning to form beneath the current O horizon. The spatial variation in this soil illustrates the challenge of sampling and quantitatively describing some soils with complex horizons and uneven boundaries. Sampling this soil would be even harder than it would appear in this picture, as large rocks were removed in the digging of the pit. This podzolized soil is classified as a Ferro‐humic Podzol in the Canadian system of soil taxonomy, an Orthic Podzol in the United Nations Food and Agriculture Organization (UNFAO) system, and a Humic Haplorthod in the US system.

26

Chapter 2

Figure 2.11  The extent of physical mixing by animals is especially evident where northern pocket gophers tunnel in the

soil in winter, plugging tunnels in the snow with soil materials that are then left in piles and ridges on the soil surface after snowmelt. The width of the ridges is about 8–10 cm.

Weakly podzolized layers developed under hemlock within 100 years after fields in New England were removed from cultivation. These layers have also formed within relatively short periods in local depressions resulting from tree uprootings in hemlock forests. Recognizable mineral soil development under black spruce, and perhaps red spruce and balsam fir, can also be fairly rapid (Lyford 1952). Buol et al. (2003) point out instances in which forest soils on glacial moraines in Alaska developed a litter layer in 15 years, a brownish A horizon in silt loam in 250 years, and a thick spodosol (podzolized) profile in 1,000 years. In addition to the within‐profile redistribution of water‐borne materials, physical mixing is important in many forest soils. Physical mixing (pedoturbation) can be a dominant feature of soil development where soil animals (such as worms and gophers) churn soils (Figure  2.11). Some soil mixing can also happen from purely physical processes (Bockheim et  al. 1997; Schaetzl and Anderson 2005). Freezing and thawing of soils can lift and move soil particles, leading to accumulations of stones on soil surfaces (which could be mistaken for erosional loss of fine materials that were once mixed with the stones). Some soils have clays that expand and contract with wetting and drying,

also physically moving soil particles. In many f­orests, windthrow of mature trees causes significant mixing of soil horizons (see Figure  6.2). Discontinuous horizons are often developed in this process, and soil horizons may even be inverted in some cases. A characteristic “pit and mound” surface condition is also often developed (Lyford and MacLean 1966). It’s important to keep in mind that some major changes in soils can result from changing chemical gradients within the soil, without any substantial movement of particles or chemicals from one location to another. Referring back to Figure  1.3, it’s evident that major differences in soil development occurred at a scale of mm to cm, in response to interactions of plant roots and soil biota with local gradients in water and oxygen.

PARENT MATERIAL CAN BE MINERAL OR ORGANIC Parent materials consist of mineral material, organic matter, or a mixture of both. The organic portion is generally dead and decaying plant remains and is the major pool of accumulation for nitrogen in soil development. Mineral matter is the predominant

Forest Soils Across Space and Time

SOILS DEVELOP ACROSS TOPOGRAPHY OVER TIME Topography can have a profound local influence on soil development, over‐shadowing that of climate (Daniels and Hammer 1992). Topography affects development mainly through its influence on soil moisture and temperature regimes and on leaching and erosion. For example, in most coastal plains, small variations in elevation can have a pronounced effect on drainage. Water in soils with restricted drainage often becomes stagnant, and plant roots and the soil community consume dissolved oxygen faster than it can be restored. The ensuing anaerobic conditions have large effects on the chemistry and mobility of iron, aluminum, manganese, and some other heavy metals. Under anaerobic (reducing) conditions, these metals move in the soil solution until they eventually precipitate or oxidize and precipitate. Such precipitation produces gleyed conditions that are characterized by a gray to ­grayish‐brown matrix sprinkled with yellow, brown, or red mottles or concretions in the B horizon. Subsoil colors are generally reliable guides to ­drainage conditions, ranging from dull bluish‐gray for very poorly drained, reduced locations to bright yellows and reds for better‐drained areas. Topography also has long‐term influences. Highly weathered soils in Puerto Rico show similar exchangeable cation stocks in soils developed on ridges from volcanic and quartz‐diorite parent

materials, but higher cation stocks on slopes (Figure  2.12). Indeed, much of the exchangeable Ca present on the ridge soils came from deposition of dust (Porder et al. 2015). Topography also affects the accumulation and cycling of N. A tropical forest in Costa Rica showed higher stocks of N on flat ridges and where the ridges transition to slopes (shoulders) than on slopes (Weintraub et  al. 2015). The decline of N stocks from the soil surface downward was much steeper on slopes, and nitrate production was much lower. The slopes probably had higher rates of N loss from erosion than the ridges, and this expectation was consistent with a strong trend in the ratios of 15N to 14N. Erosion would not discriminate between the isotopes, as moving soil particles would carry away both isotopes. Processes such as nitrate leaching and denitrification can remove a disproportionate amount of the lighter isotope, leaving more 15N behind. The flat ridge positions had δ15N values of about six, compared with less than five for the slopes (Figure 2.13). 7

Exchangeable cations (Mg ha–1)

type of parent material, and it generally contains a large number of different rock‐forming minerals in either a consolidated state (granite, conglomerate, mica‐schist, etc.) or an unconsolidated state (glacial till, marine sand, loess, transported sediment, etc.). The original mineralogy of the parent material is referred to as primary minerals, and new minerals that form during soil development are secondary minerals. As soils age, the influence of parent material on their properties and management may decline, but it doesn’t disappear (Schaetzl and Anderson 2005).

27

6

K Mg

5

Ca

4 3 2 1 0

Volcanic

Quartz-diorite ridge

Quartz-diorite slope

Figure 2.12  Millions of years of weathering created deep

soils in Puerto Rico, with similar stocks of exchangeable cations for ridge locations of soils weathered from volcanic and quartz‐diorite parent materials. Soils developed from quartz‐diorite parent material on slopes had much greater stocks, as a result of gradual erosion replenishing the near‐surface presence of unweathered rock (Source: from data in Porder et al. 2015).

Chapter 2

28

removal from soil via transpiration. Roots penetrate deep into the profile, taking up nutrients and returning them to the upper soil in leaf fall and root death. Because of the close relationship between soils and plant communities and the influence of each on the formation of the other, an examination of the process of vegetation development should promote understanding of the influence of the soil on these processes (see later chapters).

7 r2 = 0.58

Soil δ15N

6

5

Ridge 1.4 Mg N ha–1

4

3

0

10

Shoulder 1.2 Mg N ha–1

20

Slope 0.8 Mg N ha–1 30

40

Slope (degrees) Figure 2.13  The N content of forest soils (to 1 m depth) in a Costa Rican landscape was greater on flat ridges than or shoulders or slopes. The δ15N values declined as slope increased, indicating relatively greater losses of N by erosion (which does not discriminate between isotopes) on slopes (Source: from data of Weintraub et al. 2015).

BIOLOGY ALSO DRIVES SOIL FORMATION The roles of biota in soil development can hardly be overemphasized. Tree roots grow into fissures and aid in the breakdown of bedrock. They penetrate some compacted layers and improve aeration, soil structure, water infiltration and retention, and nutrient‐supplying capacity. In addition to tree roots, a multitude of living organisms in the soil are responsible for organic matter transformations, translocations, decomposition, accretion of nitrogen, and structural stability. Bacteria and fungi are intermediaries in many of the chemical reactions in the soil, and animals play a vital role by physically mixing soil constituents. Furthermore, the soil may be protected from erosion by a vegetative ground cover, and soils on steep slopes can be held in place by strong networks of roots. Forest canopies substantially modify the temperature and moisture conditions of the soil below, reducing incoming light and precipitation (as some water that lands on the canopy evaporates), and increasing water

VEGETATION AND SOILS DEVELOP TOGETHER, BUT AT DIFFERENT RATES The Earth was formed more than 4.5 billion years ago. As recently as 200 million years ago the major land masses were all joined in a single supercontinent. The supercontinent broke into fragments that largely define today’s continents, and the fragments began slow, ponderous voyages around the planet (Frisch and Meschede 2010). North America and Europe have been separate for about 80  million years, and they’re still moving apart at a rate equal to a person’s height in a lifetime. Massive continental‐scale glaciers scoured the northern part of North America and the whole of northwest Europe as recently as 10,000–20,000 year ago. The vegetation of many of the warmer areas of the world has apparently had a rather constant composition for millennia. Revegetation of the abandoned ice fields and coastal areas inundated during intraglacial periods has taken place relatively recently. Tree species evolved and simultaneously colonized the European, Asian, and North American continents while these land masses were joined in the pre‐Cretaceous period. Glaciation in recent millennia shifted the distributions of some species southward in Asia and North America, but this option wasn’t available in Europe where the Alps limited southern opportunities. This is a common explanation given for the low diversity of tree species in Europe, and the greater similarity and total diversity of tree species in Asia and North America.

Forest Soils Across Space and Time

FORESTS CHANGE AT TWO RATES Forests change over time, either so slowly we may not notice for decades, or so quickly that we’re surprised. Some changes are relatively predictable, such as the number of trees that fit into a hectare inevitably declines as average tree sizes get larger. Other changes depend on contingent events, such as droughts that combine with fires or beetles to restructure species composition across landscapes. Parent material may dictate the plants that initially colonize an area, and it can affect the species composition of stable communities. This is quite evident in glaciated boreal regions. Finer‐textured deposits are often imperfectly drained and are invaded by larch and alders. Over time, internal drainage improves as soil structure develops, and these areas may (or may not) become dominated by spruce and fir trees. In the same region, coarser deposits are initially occupied by open stands of aspen and birch or pine and shrubs. Eventually, all but the coarsest of these deposits develop a finer‐ textured B horizon and gradually become stable spruce‐fir forests. However, the coarsest sand may remain in a series of pine forests. Plant species differ in their effects on soils (see Chapter  9), so shifts in species composition over time can change soils (Fisher and Stone 1969; Fisher and Eastburn 1974).

SOIL PROPERTIES INFLUENCE VEGETATION DEVELOPMENT Five easily observable properties of soil (texture, structure, color, depth, and stoniness) can be used to infer a great deal about how a particular soil influences plant growth. Texture, whether the soil is a sandy loam or a clay loam for example, largely determines the water‐holding capacity of a soil, and often relates to nutrient supply as well. Soil structure includes possible descriptors of massive, single grain, or blocky, and structure modifies the influence of texture on water‐holding capacity

29

and its corollary, oxygen availability. Color can relate to the amount of organic matter in the soil, which is generally well correlated with fertility. Depth tells us how much water can be held and what volume of soil is exploitable for nutrients, and stoniness tells us how much non‐soil objects dilute soil volume. If we know something about the climate, these properties can go a long way toward allowing us to determine the type and vigor of the plant community that any given soil might support. Many of these properties are strongly influenced by parent material but they change during soil development, and the ability of an area to support vegetation changes as the soil matures. The type of parent material from which a soil is derived can influence its ability to supply nutrient cations such as potassium, magnesium, and calcium. These are often called “base” cations for historical reasons, even though they are not bases in the sense of acids and bases. Changes in vegetation type often coincide with changes in underlying parent rock. For example, soils derived from sandstone are generally acidic and coarse‐textured, producing conditions under which pines and some other conifers may have a competitive advantage. On the other hand, limestone may weather under the same climatic conditions into finer‐textured, fertile soils that support demanding hardwoods and such conifers as red cedar. In the Appalachian Mountains of eastern North America, it is not unusual to find a band of pine growing on sandstone‐derived soils parallel to a strip of hardwoods growing on limestone soils. Although not all limestone‐derived soils are productive, this parent material exerts an important influence on the distribution and growth of forest trees. The influence of parent material may last essentially forever, as millions of years of soil development do not remove the legacies of initial parent material. For example, soils in the Blue Mountains of Australia which developed on volcanic parent material support far greater forest production and biomass than those derived from sandstone, even after millions of years (Figure 2.14).

30

Chapter 2

Figure 2.14  After millions of years of development, soils derived from volcanic parent materials (a Tertiary basalt from a volcanic plug) in the Blue Mountains of southeastern Australia support tall (>30 m) forests of brown barrel (Eucalyptus fastigata). Nearby soils developed in Triassic sandstone parent material supports short ( granite and quartz. The “objective” of the weathering action would not be obtaining Ca, as the Ca precipitates with oxalate. The benefit is the concomitant release of other nutrients (especially P and maybe K) which the fungi take up and sends to

the tree roots. The scientists concluded that the fungal production of oxalic acid likely increased mineral weathering by 10‐ to 100‐fold.

MINERALOGY MATTERS FOREVER Early ideas about soil formation often gave a dominant role to climate, expecting that over time locations that varied in other characteristics would become similar after enough time had passed for climate to moderated initial differences (Chapter 2). This reasonable idea turned out not to be so strong. For example, the mineralogic composition of granites varies depending on the nature of the original magma body from which the granites formed. Some magma originates from melting of former igneous rocks, and some from melting of former sedimentary rocks. A comparison of soils developed on these two types of granitic parent materials in Pernambuco in eastern Brazil showed major differences (Silva et  al. 2016). The evergreen

60

250

Exchangeable cation (kmolc ha–1 to 1.8 m)

K

Soil C (Mg ha–1 to 1.8 m)

200

150

100

50

0

50

40

30

20

10

0 I-type

S-type

Mg Ca

I-type

S-type

Figure 3.13  Soils formed under warm, wet tropical rain forests differ substantially depending on the type of granite

(I‐type formed from igneous precursors, S‐type from sedimentary precursors; Source: from data in Silva et al. 2016).

Minerals in Forest Soils broadleaved forest receives about 1,800 mm rain year−1, with a warm equatorial climate. The age of the rocks is on the order of 0.5–1.0 billion years, and the soils have formed in place over an unknown period of time that would likely be many millions of years. Soils developed in granite derived

57

from igneous precursor rocks had higher clay contents, higher pools of organic matter, and more than double the pools of exchangeable cation nutrients (Figure 3.13). The legacies of the mineralogies of parent materials seem to last as long as soils last.

C ha pte r   4

Organic Matter in Forest Soils

OVERVIEW Soil organic matter forms a small fraction of the mass of most forest soils, but it profoundly influences the physical and chemical properties of the whole soil. Soil organic matter plays important roles in maintaining site productivity and is a major sink for atmospheric carbon. Soil organic matter in the principal source and sink of plant nutrients in both natural and managed forests. The O horizon is a distinctive feature of most forest soils, and O horizons often contain a substantial proportion of the entire soil organic matter. These O horizons have distinctive structure and are generally classified by their structure. Organic matter within the mineral soil horizons also drive soil development and influence the chemistry of soil solutions. The amount of carbon stored belowground as soil organic matter may roughly match that stored aboveground in the forest vegetation. The average turnover time for soil organic matter may be shorter than the lifespan of dominant trees in the forest, yet the average age of the carbon that remains behind in soils is usually much older than the trees. This seeming paradox results from the combination of a very large flow of carbon through the soil (relatively rapidly), and a very small proportion of the flow that accumulates in long‐term, stabilized soil organic matter.

Soil organic matter plays fundamental roles in soils, comprising the energy source for most biogeochemical

reactions, a binding agent that fosters soil aggregation and structure, and a reservoir for ­storing water and exchangeable ions. Carbon comprises about 45% of the mass of soil organic matter, and soils are one of the most important pools in the global C budget (see Chapter 14). Soil organic matter comprises the majority of the O horizon (forest floor) of soils, often constituting 1–15% of upper mineral soil (and B) horizons. A focus on soil organic matter and its role in soil productivity is important because of its dynamics of inputs, outputs, and net accumulation or loss in response to climate, topography, vegetation, and disturbance regime. We generally measure soil organic matter by a one‐time determination of soil organic carbon, but this is akin to determining a person’s wealth by looking at his assets and disregarding his liabilities. Ecosystems produce new organic matter and decompose old organic matter every year. When these inputs and outputs stabilize, soil organic m ­ atter also stabilizes, but if we decrease inputs or increase outputs through cultural activities, we alter the ­balance and soil organic matter may change dramatically. This could, in turn, decrease productive potential through a complex and poorly understood chain of events; the exact role of soil organic matter in soil productivity declines is unknown. Soil organic matter consists of all of the carbon‐ containing substances in the soil, except inorganic carbonates. It is a mixture of plant and animal residues in various stages of decomposition, the bodies of living and dead microorganisms, and substances synthesized from breakdown products of all of

Ecology and Management of Forest Soils, Fifth Edition. Dan Binkley and Richard F. Fisher. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

59

60

Chapter 4

these. Soil organic matter occurs in solid, colloidal, and soluble states, in relatively distinct O horizons at the top of the soil and mixed intimately with mineral horizons.

THE HIGHEST SOIL ORGANIC MATTER MAY BE IN THE CANOPY Rain forests around the world characteristically support large communities of epiphytic plants (and animals and microbes), and the accumulation of dead organic matter on large branches can proceed over centuries, leading to well‐developed histosols high above the soil surface (Nadkarni et al. 2002). Canopy organic matter can develop to depths of a meter in some situations, with well‐developed profiles grading from fibrous remnants of plant material to humified sapric material (Enloe et al. 2006). Although most forest soils begin on the ground, it is fascinating to remember that soil forming processes occur anywhere conditions are suitable.

FORESTS SOIL PROFILES TYPICALLY BEGIN WITH AN O HORIZON The O horizon a distinctive feature of most forest soils, often accounting for 20% or more of total soil C. Typical forest O horizons have masses of 1–4 kg m−2 (10–40 Mg ha−1; Figure  4.1), though it’s not unusual for O horizons to accumulate more than 5 kg m−2 in cold or wet locations. In general, O horizons are deepest where environments are cool and wet and where mixing by soil fauna is slight. Tropical forests (both plantations and natural forests) tend to accumulate 5–15 Mg ha−1, equal to about one to two years of aboveground litterfall inputs (reviewed by O’Connell and Sankaran 1997). Temperate forests commonly accumulate 20–100 Mg ha−1 of O horizon. Are decaying logs part of the soil? To a soil scientist, absolutely. The rate of organic matter input to the soil from fallen logs typically rivals the rate of leaf litterfall (Sollins 1982), though the influence of

decaying logs on underlying soil organic matter appears limited (Spears et  al. 2003; Krzyszowska‐ Waitkus et al. 2006). Decaying logs contain dynamic communities of microbes and soil animals, processing organic and inorganic compounds just as in O horizons and mineral soils. Decaying woody material is commonly exploited by plant roots, and contributes organic matter to lower horizons, via leaching (with reports up to 150 kg C ha−1  year−1, Magnússon et  al. 2016) and physical mixing. Woody material also contains large (slowly available) pools of nutrients; less than 5% of the annual soil N supply to trees likely comes from coarse woody material (Hart 1999; Laiho and Prescott 1999). In some cases, free‐living biological nitrogen fixation is detectable in woody material (Jurgensen et al. 1984; Hicks et al. 2003), though the rates are very small compared to annual N flux in forests. The soils that develop beneath fallen logs might be expected to differ substantially from soils away from logs, but this may not be the case. Stutz and Lang (2017) compared soils beneath logs with paired soils 2–3 m away. Thirty‐two comparisons from eight beech forests in Germany showed no differences in mineral soil concentrations of C or N, cation exchange capacity, pH, porosity and water holding capacity. Mineral soils beneath logs did show higher base saturation. People in other disciplines may envision O horizons and wood materials as “sitting on the soil,” and fire specialists may consider these as “duff” or simply fuels, but these components of forest soils are as fundamental to forests as mineral soil horizons. The mass of woody material in (or on) O horizons ranges from less than 20 Mg wood ha−1 in tropical forests and areas subject to frequent fires to more than 500 Mg wood ha−1 for some cool rain forests with large, long‐lived trees. Grassland soils typically have minimal O horizons, and many ecologists simply refer to grassland O horizons as the “litter layer.” Wetland and tundra soils also tend to have large O horizons; indeed, many wetland and tundra soils are classified as organic soils, rather than simply possessing an upper O horizon like in forests.

Organic Matter in Forest Soils

61

25 39 temperate and tropical forests

15 10

30

(Binkley 2002)

20

10

5 0

32 loblolly pine forests in the SE United States

(Yanai et al. 2003)

Percent of sites

Percent of sites

20

40

1

2

3

4

5

6

7

8

9

O horizon mass (kg m–2)

0

1

2

3

4

5

6

7

8

9

O horizon mass (kg m–2)

Figure 4.1  The mass of O horizons commonly varies from 1 to 4 kg m−2 (left). Even within a single forest type, O horizon mass varies several fold (right) as a result of differences in climate, parent material, and soil animal activity.

O HORIZONS HAVE HAD A BUMPY TIME IN THE HISTORY OF SOIL SCIENCE AND ECOLOGY Much of the classification and nomenclature of soils was developed to provide insights into agricultural soils, where plowing prevents the normal development of O horizons. A 2006 Food and Agriculture Organization (FAO) report (UNFAO/ ISRIC 2006) that aimed to foster communication and understanding about soils of the world essentially omitted consideration of the sorts of O horizons found in forests. When Berthrong et al. (2009) set out to examine the effects of afforestation on soil C accumulation, they found that more than 90% of published studies omitted the O horizon! At the other extreme, some forest scientists have focused so heavily on O horizons that classification schemes outstrip our understanding of the horizon’s formation, structure, and function (for example, the “order, group, subgroup” approach of Klinka et  al. (1981) or the nine humus forms defined for French soils (Ponge et al. 2011)). Forest scientists sometimes refer to the O horizon as humus or the humus form. These terms are confusing because humus is generally defined as only a portion of the O horizon and because humus or humic materials exist within the mineral soil

(Stevenson 1994; Piccolo 1996). Similarly, the term “forest floor” has been used, but this has led many non‐soil scientists to think of the O horizon as not being part of the soil. The O horizon designates all organic matter that might pass beneath our feet as we walk through a forest, down to the depth where mineral material dominates (often defined as 45

Mycorrhizae needles mosses

Fragmented litter Humus 1 Humus 2

E horizon Mineral soil

B horizon

Figure 4.3  Molecular and isotopic techniques were used to examine the composition of the fungal communities in a

Scots pine forest in Sweden (Source: Lindahl et al. (2007), used by permission of Wiley).

dominated by fungi, but the types of fungi were not uniform across the horizon. Saprophytic fungi that obtain carbon and energy by decomposing organic matter predominated in the upper O horizon. Mycorrhizal fungi (obtaining C primarily from roots) were the important functional group in the lower O horizon and upper A horizon. Burrowing (fossorial) mammals are among Nature’s other agents operating to alter O horizons, including gophers, moles, and shrews. These animals often pile soil around burrow entrances and move and mix soil when making tunnels (Troedsson and Lyford 1973; see Figure 2.11). Abandoned runways eventually collapse or become filled with soil material from above. Transport of mineral soil particles into the organic horizons may also occur as a result of the activity of ants, termites, earthworms, rodents, and other small animals. Where the mineral soil materials are deposited by these fauna, soil

properties are likely to show sudden changes. The reverse process, transport of organic matter into the mineral soil, also results from animal activity. Forest soil O horizons are often divided into three structural layers, progressing from less processed litter at the top to highly humified material at the bottom of the O (or in the transition from the O to the A horizon). These strata may not occur in all soils (Hesselman 1926; Green et  al. 1993). The upper layer was classically called the litter or L layer, and is currently referred to as the Oi horizon in the USA, or OL horizon in Europe (Zanella et al. 2011) and other parts of the world (UNFAO 1990). The Oi or OL horizon consists of unaltered dead remains of plants and animals whose origin can be readily identified. The middle layer is the “Formultning” (decay, sometimes mistranslated as fermented) layer with partly decomposed organic materials that are sufficiently well preserved to

64

Chapter 4

­ ermit identification of their origin (Prescott et al. p 2000). This horizon is termed the Oe in the USA, and the OF in most other places. The lowermost layer contains material that is not easily identified by original litter structure; highly humified material gives the layer a name of humus (though this sometimes is used to refer to entire O horizons), the H‐layer, the Oa horizon in the USA, and the OH horizon in much of the rest of the world. These structural differences correspond to a wide variety of difference in biotic activity and rates of biological processes that result in different rates of organic matter turnover. Franklin et al. (2003) used isotopic techniques to assess the mean age of O horizon layers in a Scots pine forest in northern Sweden. The OL had an average age of 5.2 years, the OF 13.0 years, and the OH 38.4 years. Summing up these layers and accounting for their masses led to an overall average age of the O horizon of 17.1 years.

MANY O HORIZON CLASSIFICATIONS HAVE BEEN DEVELOPED: ARE THEY USEFUL? As with most natural systems, people have tried to develop classification systems to describe and account for differences in O horizons among forests. The variety of forms and processes in O horizons prevents any simple, unequivocal classification, so many approaches have been developed. It’s not clear that classification efforts have led to clear insights that might, for example, provide an index of soil fertility or guide forest management. For example, Romell and Heiberg (1931), as revised by Heiberg and Chandler (1941), described humus layer types for northeastern United States upland soils. They defined mull as a humus layer consisting of mixed organic and mineral matter in which the transition to a lower horizon is not sharp. They divided mull humus into (i) coarse; (ii) medium; (iii) fine; (iv) firm; and (v) twin types. They described mor humus as a “layer of unincorporated organic material usually matted

or compacted or both, distinctly delimited from the mineral soil unless the latter has been blackened by the washing in of organic matter.” They divided mor humus into (i) matted; (ii) laminated; (iii) granular; (iv) greasy; and (v) fibrous types on the basis of morphological properties. An example of mor humus is shown in Figures  4.2 and 4.3. However, while this classification was adequate for the northeastern United States, some O horizons in other regions were not clearly identified under this system. Klinka et al. (1981) applied a species‐taxonomy approach to classification of O horizons, defining orders of mor, moder, and mull, with 19 groups (such as hemimor and hemihumimor), and 69 subgroups (such as amphihemihumimor and microvelomoder). Green et al. (1993) proposed a comprehensive classification of organic and organicenriched mineral horizons. They defined three taxa: mor, moder, and mull at the order level based on the type of F horizon and the relative prominence of the Ah horizon. Below this order level are 16 taxa at the group level, reflecting differences in the nature and rate of the decomposition process. The latest foray into O horizon classification comes from Europe, with the proposition of six major types: Anmoor, Mull, Moder, Mor, Amphi and Tangel (Zanella et al. 2011). This scheme accounts for variations that depend on soil features outside the O horizon. For example, the same sort of a mull horizon would be called a Terromull if it occurs on a well‐drained soil; an Entimull if lying directly on bedrock; or a Paramull if plant roots are particularly dominant. What practical insights might be supported by characterization of the structure of O horizons? One of the most important is the challenge presented in determining the mass (and element content) of the O horizon. Mor‐type O horizons have a distinct lower boundary, and typical mass estimates would be reliable to about ±20% (Yanai et  al. 2003a). O horizons with less distinct lower boundaries may be difficult to quantify with less than ±50% uncertainty, and this may cloud discussions of rates of change over time (Yanai et al. 2003b).

Organic Matter in Forest Soils 2.5 2.0 Foliar P (mg g–1)

Another practical insight is that the acidity of O horizons tends to relate to the degree of incorporation of mineral material. Organic matter tends to be more strongly acidic than mineral surfaces, so mor‐ type O horizons may have lower pH (= more strongly acidic) than mull‐type O horizons. Tree species and sites with acidic litter (and lower calcium) may have lower activities of soil animals (Reich et al. 2005), leading to circular effects of O horizon chemistry, morphology, and mixing by animals. Perhaps the most interesting aspect of O horizon structure would simply be the indication of the importance of certain types of soil organisms. Mull‐ type structures form as a result of mixing by soil animals (particularly worms), and macrofaunal biomasses may be three to five times greater than found in nearby soils with mor‐type structure (cf. Staaf 1987; Schaefer and Schauermann 1990). Moder‐form O horizons may result from activities of epigeic worms or other soil fauna such as millipedes which do not mix the organic material with mineral soil but rather ingest it (once it has become partially decayed and invaded by fungi). About 90% of the initial mass consumed is subsequently egested and may persist as a layer of fecal pellets, which have variable decay rates (C. Prescott, personal communication). Interest in O horizon morphology was driven implicitly or explicitly by hopes that structure would provide a key to important features such as tree nutrition and site productivity. This expectation has not been borne out; no universal pattern of O horizon morphology or decomposition patterns has proven to be of much use in forest nutrition studies, probably because so many important interactions (of tree species, soil communities, and a suite of soil forming factors) prevent simple patterns. Old expectations of mull O horizons (where mixing by soil animals comminutes the organic matter and buries it) being somehow richer than mor O horizons (where organic matter remains separate from mineral soil) have been supported only in the broadest terms: the richest forest soils have animals that mix litter into the mineral soil,

65

1.5 1.0 0.5 0.0

Mor O horizon

Mull O horizon

Figure 4.4  Sugar maple phosphorus deficiency develops

in mull O horizons because mixing of organic phosphorus pools with mineral soil allows phosphorus to precipitate with iron and aluminum (Source: data from Paré and Bernier 1989a,b).

and the poorest forest soils do not. Between these extremes, many examples have been found in which one O horizon type appears to be more fertile than another, with other cases showing the reverse pattern. For example, the supply of phosphorus to sugar maple trees is lower on mull soils than mor soils in Quebec (Figure 4.4), in contrast to classic expectations of higher fertility of mull soils. We suggest that after almost a century, these hopeful (but unsupported) ideas about litter decomposition rates and O horizon morphology should be set aside and the processes that determine the turnover of soil organic matter should be examined in both O horizons and mineral horizons, especially recognizing potentially dominant influences of belowground sources of plant litter (cf. Pollierer et al. 2007).

DEAD WOOD IS ALSO PART OF THE SOIL C As noted above, forests produce about as much wood each year as they do roots and leaves, and through the lifetime of a forest the amount of C added to the soil may be roughly equal for woody material and fine material like leaves and small roots. Two key questions for the woody material would be how long does it take for the material to

Chapter 4

5 yr

rates of decomposition of woody materials, but difference may be as great based simply on the species of wood (Kahl et al. 2015). Stumps and roots are buried in the soil, and moist conditions might be expected to speed decomposition compared to logs the soil surface. Good ideas need to be challenged by evidence. The rate of decomposition of dead logs on the soil surface at the Calhoun Experimental Forest in South Carolina was much faster than the rate of loss for stumps (Figure 4.5, Mobley et al. 2013). Indeed, the rate of log decomposition was much faster at the Calhoun site than the decomposition rate of needle litter in

20 yr

55 yr

10 yr

Remaining stump and root biomas (kg tree–1)

decompose, and how much is moved into longer term (sequestered) soil C? The rate of mass loss from stumps and coarse roots in loblolly pine stands in North Carolina, United States was about 5% each year (Figure  4.5, Ludovici et  al. 2002). The rate of decay removed about one‐third of the woody mass in the first eight years, and the next one‐third was gone after a total of 20 years. For comparison, stumps and coarse roots of in Eucalyptus plantations in Brazil take only three years after harvesting to lose one‐third of their original mass, and six years to lose two‐thirds (Stape et  al. 2008). Climate typically influences

Remaining mass (% of original)

66

160 140 120 100 80 60 40 20 0

0

20 40 Time since harvest (years)

60

100 80 60 40

Stumps+ coarse roots

20 0

Wood 0

10

Litter

20

30

40

50

60

Time since harvest (years) Figure 4.5  Decomposition over six decades transforms most of the C in stumps and coarse roots into CO2, leaving less

than 5% of the original material (Source: upper graph, and photos from Ludovici et al. (2002), used by permission of NRC Research Press). For comparison, decomposition of logs (wood) on the soil surface may decompose much faster, while decomposition of foliage litter in the O horizon may be intermediate (Source: lower graph; wood and litter data for Calhoun Experimental Forest, from Mobley et al. 2013 and Richter et al. 1999).

Organic Matter in Forest Soils the O horizon (Richter et al. 1999). Other studies that compared wood decomposition on the soil surface and buried in the soil have shown faster decay for surface locations (for example, Mosier et  al. 2017). In these cases, it seems the good idea of moister conditions in the soil favoring wood decomposition does not warrant general confidence.

ORGANIC MATTER IS A VITAL PART OF MINERAL HORIZONS TOO Identifiable pieces of plant material comprise only a portion of soil organic matter. The largest and most important fractions of soil organic matter are the colloidal and soluble fractions (McColl and Gressel 1995). The majority of soil organic matter is insoluble and is bound as macromolecular complexes with calcium, iron, or aluminum or in organomineral complexes. Organic matter in mineral soils “knits” together the small mineral particles into larger soil aggregates (especially in soils with 2:1 clays; Six et al. 2000), with far‐ranging importance for soil porosity, aeration, and water infiltration, flow and retention. The organic portion of these complexes is made up of diverse complex compounds, classically lumped into the term “humus.” The huge variety of chemical compounds within humus has led to intensive efforts to fractionate the bulk pool into subpools that differ in chemical composition, reactivity, and turnover rates. The most common characterization of organic matter in the twentieth century entailed separating fractions that were soluble in strong acids, weak acids, or neither. The defined humic acid fraction is soluble only in strong acid (pH 80% amorphous material) derived from wood and litter from a western redcedar‐western h ­ emlock forest. Differences in the origin of humic materials are evident from the high lignin‐methoxyl peak for the wood‐ derived humus and for the tannin peak for the ­litter‐derived humus (after deMontigny et al. 1993). 13

We know that organic matter in mineral soils is important, diverse in composition and dynamics, and not likely to be easily characterized in ­powerful ways. The laboratory based measurements of soil C or soil organic matter involve intensive, time consuming, and costly methodology that limits applicability for large land areas. Recently, research efforts have focused on measuring soil C in situ using a variety of methods (Gehl and Rice 2007). These methods include Laser Induced Breakdown Spectroscopy (LIBS), Inelastic Neutron Scattering (INS), near‐ infrared spectroscopy (NIRS), and remote sensing (Table 4.1). However, none of these techniques is in common usage.

Organic Matter in Forest Soils

69

Table 4.1  Comparison of field instrumentation for determination of soil C.

Method

Principle

Approximate detection limit

Sampling time

Sampling attributes

Laser induced breakdown spectroscopy

Intense laser pulse is focused on intact sample core, microplasm emission spectrally resolved Inelastic scattering of neutrons from C nuclei emits gamma rays that are detected and analyzed for C peak intensity Diffuse reflectance properties in the 700 to 2500 nm spectrum are correlated to soil C content

300 mg total C kg−1

38 cm in length Stony or Bouldery

Structural classes are determined by comparing a representative group of peds with a set of standard‑ ized diagrams. Grade is determined by the relative stability or durability of the aggregates and by the ease of separating one from another. Grade varies with moisture content and is usually determined on nearly dry soil and designated by the terms weak, moderate, and strong. The complete descrip‑ tion of soil structure, therefore, consists of a combi‑ nation of the three variables, in reverse order of that given above, to form a type, such as weak sub‑ angular blocky. Important drivers of soil aggregation include mineral chemistry, salts, clay skins, oxide coatings, growth and decay of fungal hyphae and roots, freezing and thawing, wetting and drying, and the activity of soil organisms (especially earthworms). Soil texture has considerable influence on the development of aggregates. The absence of aggre‑ gation in sandy soils gives rise to a single‐grain structure, whereas loams and clays exhibit a wide variety of structures. Soil animals, including earth‑ worms and millipedes, help form crumb structure in the surface soil by ingesting mineral matter along with organic materials, and producing casts that provide structure to the soil. The intermediate products of microbial synthesis and decay are effec‑ tive stabilizers, and the cementing action of the more resistant humus components that form com‑ plexes with soil clays gives the highest stability. Soil aggregation develops differently under the influence of different tree species. In an experi‑ ment with 35‐year‐old plots with different tree species, the average size of aggregates ranged from 1.5 mm under white pine to 2.1 mm under Norway spruce (Scott 1998). Across the species, average

Physics in Forest Soils: Soil Structure, Water, and Temperature

BULK DENSITY ACCOUNTS FOR THE COMPOSITION OF MINERALS, ORGANICS, AND PORE SPACE Bulk density is the dry mass (of  1 (dimensionless)

Water content

Aggregated soil

Compacted soil

This model is not perfect and other models have been developed to overcome some of its specific weaknesses, however no model has as yet replaced it in general use.

SEVERAL FACTORS AFFECT INFILTRATION AND LOSSES OF WATER Water tension Water potential Figure 5.15 Effects of soil aggregation on the water

release curve.

(more negative) potentials. At any given potential, peaty soils will usually display much higher moisture contents than clayey soils, which would be expected to hold more water than sandy soils. The water hold‑ ing capacity of any soil is due to the porosity and the nature of the bonding in the soil (Figure 5.14). The degree of soil aggregation, particularly in finer textured soils, can have a pronounced effect on the water release curve (Figure 5.15).

The term infiltration is generally applied to the entry of water into the soil from the surface. Infiltration rates depend mostly on the rate of water input to the soil surface, the initial soil water con‑ tent, and internal characteristics of the soil (such as pore space, degree of swelling of soil colloids, and organic matter content). Overland flow happens only when the rate of rainfall exceeds the infiltra‑ tion capacity of the soil. Because of the sponge‐like action of most O horizons and the high infiltration rate of the mineral soil below, there is little surface runoff of water in mature forests. Overland flow is rarely a problem in undisturbed forests even in steep mountain areas.

Physics in Forest Soils: Soil Structure, Water, and Temperature 0 Former agricultural –10

Soil depth (cm)

When rainfall does exceed infiltration capacity, the excess water accumulates on the surface and then flows overland toward stream channels. This surface flow concentrates in defined stream channels and causes greater peak flows in a shorter time than water that infiltrates and passes through the soil before reap‑ pearing as stream flow. High‐velocity overland flow can erode soils. Soil compaction by harvesting equip‑ ment, disturbances by site preparation, and practices that reduce infiltration capacity and cause water to begin moving over the soil surface are of great concern to forest managers. Special care is warranted on shal‑ low soils and those that have inherently low infiltra‑ tion rates. Soils that are wet because of perched water tables or because of their position along drainage ways are unusually susceptible to reduction in porosity and permeability by compaction. The O horizon is especially important in main‑ taining rapid infiltration rates. This layer not only absorbs several times its own mass of water, but also breaks the impact of raindrops, prevents agita‑ tion of the mineral soil particles, and discourages the formation of surface crusts. The incorporation of organic matter into mineral soils, artificially or by natural means, increases their permeability to water as a result of increased poros‑ ity. Forest soils have a high percentage of macropo‑ res through which large quantities of water can move‐sometimes without appreciable wetting of the soil mass. Most macropores develop from old root channels or from burrows and tunnels made by insects, worms, and other animals. Some macropores result from structural pores and cracks in the soil. As a consequence of the better structure, higher organic matter content, and presence of channels, infiltration rates in forest soils are consid‑ erably greater than in similar soils subjected to con‑ tinued cultivation (Figure 5.16). The presence of stones increases the rate of water infiltration in soils because the differences in expansion and contraction between stones and the soil produce channels and macropores. However, higher infiltration rates in in the presence of stones are matched by lower retention capacity of soils for 40 Oregon sites (Figure 5.17).

101

soil

–20

Forest soil

–30 –40 –50 –60 0.01

0.1

1

10

Percolation rate (cm

100

1000

hr –1)

Figure 5.16  Comparative percolation rates by soil depth

for a forest soil and an adjacent old‐field soil in the South Carolina Piedmont (Hoover 1949).

Burning of watersheds supporting certain types of vegetation may result in a well‐defined hydro‑ phobic (water‐repellent) layer at the soil surface, which may increase erosion rates (see Chapter 12). This heat‐induced water repellency results from the vaporization of organic hydrophobic substances at the soil surface during a fire, followed by con‑ densation in the cooler soil below. Evaporation from surface soils increases with increases in the percentage of fine‐textured materials, moisture content, and soil tempera‑ ture and decreases with as atmospheric humidity increases. Evaporation losses are also influenced by wind velocity. Heyward and Barnette (1934) found that the upper soil layers of an unburned longleaf pine forest contained significantly more moisture than did similar soils in a burned forest because of the mulching effect of the O horizon in the unburned areas. Fine‐textured soils have higher retention capac‑ ity for water than sands, and can store larger amounts of water following storm events. The effective depth of the soil (rooting depth) and the initial moisture content are factors that also influ‑ ence the amount of rainwater that can be retained in soils for later use by plants. The condition of a forest has a strong influence on soil water as a result of differences in water use and water input. Clear cutting a mixed conifer

Chapter 5

Soil water retention capacity (cm)

102

50 40 30 20 10 0

0

20 40 60 80 Soil stone content (% by volume)

100

Soil moisture (% dry mass)

Figure 5.17  The water retention capacity in the surface 120 cm of 40 soil profiles in Oregon, USA, declined linearly with increasing rock content (Dyrness 1969).

30

Clearcut

20

Uncut

10

0 February

April

June

August

October

Figure 5.18  Seasonal course of soil moisture in a clearcut and an uncut forest in the mountains of Montana (Source: data

from Newman and Schmidt 1980).

f­orest in Montana led to substantially wetter soils throughout the year (Figure  5.18). The difference was related to lower water use in the absence of trees but also to greater snow input (Figure 5.19). Snow that accumulates in conifer canopies may sublimate without reaching the soil surface, lower‑ ing the input of water to the soil relative to clearcut areas, with less sublimation (or interception loss). Differences in interception loss can also be impor‑ tant among tree species that support substantially different leaf areas. For example, Helvey and Patric

(1988) contrasted the interception loss for a water‑ shed dominated by a mixed‐age, mixed‐species hardwood forest with a watershed planted with a single‐age stand of white pine. Average annual interception loss from the mixed‐hardwood stand (250 mm year−1) was less than half of the loss from the white pine stand (530 mm year−1). Soil lea‑ chates were more concentrated under the pine stand, due in part to the reduced quantity of water leaching through the soils (data in Johnson and Lindberg 1992).

Physics in Forest Soils: Soil Structure, Water, and Temperature 25 Snowpack water content (cm)

Clearcut 20

Uncut

15

10

5

0

Year 1

Year 2

Year 3

Figure 5.19 Water content of the snowpack across

three years was substantial higher in clearcut areas than in adjacent mature conifer forests in the mountains of Montana (Source: data from Newman and Schmidt 1980).

TREES MAY GET WATER FROM THE CAPILLARY FRINGE The upper surface of the zone of saturation in a soil is called the groundwater table. Extending upward from the water table is a zone of moist soil known as the capillary fringe. In fine‐textured soils, this zone of moisture may approach a height of 1 m or more, but in sandy soils it seldom exceeds 25–30 cm. In some soils, the height of the water table fluctu‑ ates considerably between wet and dry periods. In forested soils in New England water tables are high‑ est during late autumn, winter, and early spring; gradually lower in late spring and early summer; and rise again in the autumn. In the flatwoods of the southeastern coastal plain of the United States, the water table may be above the soil surface dur‑ ing wet seasons and fall to a depth of 1 m or more during dry seasons. Soil depth often reflects the volume of growing space for tree roots above some restricting layer, and in wet areas this restricting layer may be a high water table. The water table may be a temporary (perched) zone of saturation above an impervious layer following periods of high rainfall.

103

Large seasonal fluctuations in the depth to the water table are harmful to root development for many species. Roots that develop during dry p ­ eriods in the portion of the profile that is later flooded may be killed under the induced anaerobic condi‑ tions. A reasonably high water table is not ­detrimental to tree growth as long as there is little fluctuation in its level. In an experiment in the southeastern coastal plain of the United States, slash pines grown for five years in plots where the water table was maintained at 45 cm by a system of subsurface irrigation and tile drains were 11% taller than trees in plots where the water table was maintained at 90 cm, but they were 60% taller than trees in plots where the water table was allowed to fluctuate normally (White and Pritchett 1970). The primary benefits derived from drainage of wet soils may result more from an increase in the nutrient supply than from an increase in the soil oxygen supply. Nutrition may be improved by an increase in the volume of soil available for root exploitation, and by faster mineralization of organic compounds. For example, fertilizer experi‑ ments in nutrient‐poor wet soils of the coastal plain indicated that tree growth was increased as much by draining as by fertilization (Pritchett and Smith 1974).

TREES REQUIRE GREAT QUANTITIES OF WATER Tree roots absorb vast quantities of water to replace that lost by transpiration and that used in metabolic activities. Under favorable conditions, this loss can be as much as 6 mm day−1 (60,000 L  ha−1) during summer. Evapotranspiration by forests is generally higher than grasslands, owing to the combined effects of greater interception loss (evaporation from external water on the canopy surfaces) and greater rooting depth in the soil and subsoil (Figure  5.20). Differences among vegetation type on ecosystem hydrology can be complex. Afforestation alters soil structure, often increasing the ability of soils to absorb water during rain

Chapter 5 Evapotransipration (mm yr –1)

104

1,500

Forest

1,200 900 Grassland

600 300 0

0

500

1,000

1,500

2,000

2,500

3,000

Precipitation (mm yr –1) Figure 5.20  Watershed studies from around the world have documented higher water use by forests than grasslands

across broad gradients of annual precipitation (Source: from data of Zhang et al. 2001).

events, leading to greater water supply for vegeta‑ tion as well as recharging ground water (Figure 5.21; Ilstedt et al. 2007; Ellison et al. 2017). Greater infil‑ tration can also alter the hydrograph (the timing and volume of flows) of streams (Brown et  al. 2004). Tree cover is also important to the physics of ecosystem temperatures; evapotranspiration con‑ verts intense solar radiation energy into the latent energy of water vapor, substantially cooling land‑ scapes (Figure 5.22).

PHYSICS IN A LANDSCAPE CONTEXT CAN DETERMINE SOIL CHEMISTRY The physical properties described above need to be placed in a landscape context. Most forest soils are affected by their context on landscapes. Ridge top soils may be excessively well drained. Mid‐slope sites receive water and nutrients from upslope sites but lose some water to down‐slope sites. Lower‐ slope positions may receive more water and nutri‑ ents than leach away, and in some cases may become saturated and experience periods of low oxidation–reduction potential. These landscape issues are sometimes lumped together under the term topography, and many generalizations may be offered. Landscape position (or topography) is very important in forest soils but that broad generaliza‑ tions may not be helpful. For example, the sequence

of soil moisture supply, nutrient supply, and overall fertility commonly increases from ridges, to mid slopes, to toe‐slopes in relatively young soils on the sides of hills, mountains, and valley sides. Other situations, such as the Eucalyptus plantation on an Oxisol in Argentina, show exactly the reverse pat‑ tern (Dalla‐Tea and Marco 1996). Rather than try to establish generalizations about the role of landscapes in soil physics, we present some of the biogeochemistry of a two landscape case studies. The first case study developed from concerns that forest regeneration was strikingly variable after logging of some old‐growth forests on Vancouver Island, Canada (Prescott et  al. 2013; Sajedi et  al. 2012). The old forests were comprised of western hemlock, western redcedar, and amabilis fir. After logging, some areas regenerated well, and others poorly, even though the soils seemed very similar (Figure  5.23). Concerns about forest regeneration led to an intensive research collaboration among sci‑ entists and foresters: the Salal Cedar Hemlock Integrated Research Program (SCHIRP). Possible explanations included competition with various spe‑ cies of understory plants (especially woody salal, an ericaceous shrub), nutrient limitation, historical leg‑ acies of soil mixing by windthrown trees, and prob‑ lems with excessive amounts of organic compounds (tannins) that could interfere with nitrogen cycling. A variety of intensive measurements and experi‑ mental manipulations showed that very small

Physics in Forest Soils: Soil Structure, Water, and Temperature

transpiration + interception

groundwater recharge

groundwater recharge

surface runoff

soil evaporation

revi

sed

domin

105

infiltration

para

digm

ant pa

radigm

ground surface water table

canopy cover Figure 5.21  Trees increase the infiltration capacity of soils, increasing groundwater recharge in dry areas. In wetter areas, water use by dense forests can reduce groundwater recharge (Source: from Ellison et al. (2017), based on Ilstedt et al. (2016), used by permission of Elsevier).

­ifferences in topography led to differences in d water flow and retention, and with an annual rain‑ fall of 2 m in a cool climate even small differences in soil water retention can change soil redox, chemis‑ try, and oxygen supply to roots and the soil biota. The experiments also showed strong nutrient limi‑ tation of the regenerating forests. The original expectation was that a poorly performing site would have a strong response to fertilization than a site that performs better. The experiments showed very strong nutrient limitations on both types of sites, with about fivefold greater growth of both planted hemlock and redcedar. However, a fivefold response for a productive site is a far greater abso‑ lute response (in m3 ha−1 of wood) than the same relative response on a poorer site (Figure 5.24, and see Chapter  13). An investment in fertilization

would have a much greater payback on the type of site that was already the more productive type. The second case study of small difference in topography generating large differences in soils, biogeochemistry, and forests comes from northern Sweden (see also Figure 8.21). The Betsele transect, in the valley of the Umeå River, was 100 m long and descends a gentle (2%) slope of soils formed from sandy till deposits. At the upper end it was dominated by a 125‐year‐old stand of Scots pine with an understory of dwarf shrubs (mostly Vaccinium). The lower end was dominated by a Norway spruce stand of the same age, with a vari‑ ety of tall herbs in the understory. The upper end was much less productive, site index of 17 m at 100 years, than the lower end, site index 28 m at 100  years (Figure  5.25). The key driver of the

Germany

Poland

Czech Republic

a

i vak Slo

Austria Hungary

48°C

20°C 0

250

500 m

Old hemlock Old redcedar Clearcut hemlock Clearcut redcedar

B horizon redox potential (Eh in mV at pH 7)

Figure 5.22  Evapotranspiration from tree canopies in the lower part of the image cools the landscape, with temperatures far cooler than less vegetated areas (Source: from Ellison et al. (2017), based on Hesslerová et al. (2013); used by permis‑ sion of Elsevier).

350 Hemlock 325

Redcedar

300 275 250

Forest

Clearcut

Figure 5.23  Sites that appear to be uniform may still differ enough in soil water regime to drive major differences in forests. Old‐growth forests characterized by hemlock and by redcedar form a mosaic on Coastal Vancouver Island, Canada, on soils that appeared very similar. Post‐logging regeneration was much better on former‐hemlock sites than former‐redcedar sites. After many ideas were tested, the key difference appeared to be slight differences in topographic position and water regime, leading to differences in soil aeration (as indexed by redox potential; Source: from data of Sajedi et al. (2012); photo courtesy of Cindy Prescott).

Physics in Forest Soils: Soil Structure, Water, and Temperature 400 Redcedar site

Stem volume at 22 years (m3 ha–1)

Stem volume at 22 years (m3 ha–1)

120 100 80 60

107

Hemlock Redcedar

40 20 0

Hemlock site 300 200

Hemlock Redcedar

100 0

Control

Fertilized

Control

Fertilized

Figure 5.24  The volume of wood accumulated at age 22 years was greater for hemlock planted on the site formerly

characterized by than when planted on site formerly characterized as redcedar. The reverse was true for redcedar trees. Both species responded very strongly to fertilization with nitrogen + phosphorus, but fertilization gave much greater responses (in m3 ha−1) for the hemlock sites (note the threefold difference in the scales of the Y‐axes; Source: data from Prescott et al. (2013)).

Feature

Upper End

Lower End

Dominant tree site index (m at 100 years) Basal area (m2/ha) O horizon morphology

Scots pine 17 22 Thin Oa (L), indicating little soil faunal mixing Thick Oe (F) Thin Oi (H)

Norway spruce 28 32 Thick Oi (H) only, indicating major soil fauna activity

3.8 10 300 2

6.4 160 8 35

O horizon solution pH NO3 (μmol/L) PO4 (μmol/L) O horizon Extractable Fe + Al (mg/g)

Figure 5.25  Stand and soil characteristics along a 90‐m transect of soils developed in sandy till in Sweden (Source: data

from Giesler et al. (1998), photos courtesy of Peter Högberg).

108

Chapter 5

­ ifferences in soil chemistry and overall soil fertility d was soil water, and its influence on biogeochemis‑ try. Soils at the upper end the Betsele transect were excessively drained; water leached downward out of the soil. The lower end was in a position to receive the “discharge water” from the upslope ecosystems (both laterally and from up the valley), providing abundant water supply and changing nutrient cycles. The major biogeochemical implica‑ tions were that the upper ecosystems acidified and are nitrogen limited, and the lower ecosystems received abundant alkalinity (acid‐neutralizing capacity) from the discharge water and had high nitrogen availability. Another notable feature was the low (and limiting) concentrations of phospho‑

rus at the lower slope position. Mixing of the organic litter with mineral soil brings phosphorus into contact with iron and aluminum, allowing sorption and removal of phosphorus from the soil solution. Alternatively, Giesler et  al. (1998) sug‑ gested that iron and aluminum enrichment of the O horizon at the toe‐slope may have resulted from dissolved inputs, particularly iron, which may have high solubility during anaerobic periods under saturated conditions. Overall, the authors conclude that the three‐unit gradient in soil pH was driven primarily by the differences in base saturation, caused by input of alkalinity into dis‑ charge water, rather than by differences in acid quantity or strength.

C ha pte r   6

Life in Forest Soils

OVERVIEW Soils are living systems, where tree roots connect with the microbial and animal communities in the soil. These living systems are very flexible and indetermi­ nate, as feedbacks between processes and structures create malleable “physiological” networks. Root sys­ tems penetrate deep into soils, typically several meters or more. Roots anchor trees to the soil, resisting the forces of wind on canopies, and roots also help anchor soils on hillslopes and mountainsides. The surface area of roots is augmented with very small diameter filaments in the form of mycorrhizal fungal hyphae; much of the uptake of water and nutrients come through these hypha rather than directly into roots. Trees fuel the soil food webs through the input of aboveground detritus, and especially through the direct supply of carbohydrates to roots (while living, and after death) and the associated myco‐rhizosphere community. Soil foodwebs are the locus of the great majority of genetic biodiversity in forests, exceeding the diversity of plants by orders of magnitude.

THE SINGLE MOST IMPORTANT THING TO COMMUNICATE ABOUT SOILS Perhaps the most important insight to share with people who know little about soils is that soils are  vibrant, living systems. Hans Jenny (1984)

expressed this idea, “Many ecologists glibly desig­ nate soil as the abiotic environment of plants, a phrase that gives me the creeps.” A non‐living view of soils would expect that geochemical equilibria would explain soil chemistry, that change would be minimal, and that risks of damaging or enhancing soils would be small. A living view of soil recog­ nizes the fundamental role of organic molecules in shaping soil chemistry, that soil processes are far removed from equilibrium as a result of free‐energy generated by living organisms, and that soils are very dynamic and subject to both harm and improvement. Soils have often been an integral part of ecosystem‐focused studies, and just as often have been ignored or placed in a “black box” where only the inputs and outputs warrant attention (reviewed by Binkley 2006). Charles Darwin may have been the first scientist to begin to appreciate the biological nature of soil and to make systematic observations. A friend of Darwin’s called his attention to the “sinking” of rocks placed on the surface of meadow soils, leading Darwin to ponder “On the Formation of Mould” (Darwin 1883):

I was thus led to conclude that all the vegetable mould over the whole country has passed many times through, and will again pass many times through, the intestinal canals of worms. Hence the term “animal mould” would be in some respects more appropriate than that commonly used of “vegetable mould.”

Ecology and Management of Forest Soils, Fifth Edition. Dan Binkley and Richard F. Fisher. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

109

110

Chapter 6

In modern terminology, we would use the “humus” or “soil organic matter” rather than “mould.” Darwin continued to develop a very integrated, bio‐centric focus on soil development: Beneath large trees few castings can be found dur­ ing certain seasons of the year, and this is appar­ ently due to the moisture having been sucked out of the ground by the innumerable roots of the trees; for such places may be seen covered with castings after the heavy autumnal rains. Although most coppices and woods support many worms, yet in a forest of tall and ancient beech‐trees in Knole Park, where the ground beneath was bare of all vegetation, not a single casting could be found over wide spaces, even during the autumn.

The presence, activity and legacies of organisms largely determine structure and dynamics of soils. Without life we would have only regolith at the surface of the earth. Life manages to enhance the complexity of soils, taking high‐quality solar energy and converting it to thousands of high‐energy car­ bon compounds that support the most concen­ trated biodiversity on Earth. Soils are very dynamic systems, but the massive suites of interactions and feedbacks sustain long‐term persistence (Paul 2015). Sustained inputs of carbon compounds by plants allows for expansion and collapse of populations of soil organisms, from scales of micrometers to meters. Expanding populations hasten the consumption of substrates, and the opportunity for predator populations to rise. Symbioses may be as ubiquitous and important in soil dynamics as are competition and predation (for example, Haruna et al. 2017; for fascinating insights on symbioses, see Yong 2016). Changing concentrations of substrates (as well as soil oxygen and water) over space and time ensure that soils are never homogenized, and never remain in exactly the same state for long. Out of the seemingly chaotic interactions arise the soils that sustain our forests and civilizations. The Global Soil Biodiversity Initiative (www. globalsoilbiodiversity.org) produced a wonderful atlas that provides the best summary available on

the huge and fascinating topic of soils and life (Orgiazzi et al. 2016). The atlas can be downloaded at no charge: http://atlas.globalsoilbiodiversity.org.

LIVING SOIL HAS THREE FUNDAMENTAL FEATURES An inanimate system may be understood based on its physical structure and its passive interactions of energy and matter. Living systems have additional complications because feedbacks lead to elaboration of structure, alteration of processes, and the emergence of novel features. Small changes in inanimate systems, such as automobile engines, can lead to complete failure of the system. Living systems routinely change in small and large ways, and catastrophic failures are quite rare. This living view of the soil has three fundamental features: • Soils are “physiological” networks with intimate relationships at microscopic scales • Feedbacks between processes and structure lead to changes in future processes and structures, and • The state of living soils is flexible and indeterminate. The processing of matter and energy in soils entails intricate interactions between the supplies of high‐energy molecules (such as carbohydrates from plants), the generation and excretion of exoen­ zymes by microbes that digest large molecules into small, soluble molecules, the transport of these mol­ ecules through soils and across membranes of cells, and all the biochemistry that occurs within cells. The rates of these processes depend on the processes that influence the growth and death of plants (and plant tissues), the microenvironmental conditions that influence chemical and biological processes (such as soil moisture and oxygen supply), and the genetic potential of the suite of soil organisms. This physiological nature of living soil leads to a sort of “uncertainty principal”: the more an individual detail is isolated for study from the time and space

Life in Forest Soils context of the living soil, the less certain we can be about this detail actually functions in real, dynamic soils. Soil science depends largely on samples of soils that are cut from the living soil, placed in vials or on microscope slides, treated with chemicals and meas­ ured, but these acts of measurement in part deter­ mine what will be measured. Högberg and Read (2006) emphasized the need to take an integrated, physiological approach to soils: The inherent complexity of the connected plant– soil system has long served to deter soil scientists from carrying out appropriately integrated studies of naturally intact systems. Although fragmenta­ tion of such systems is sometimes justified, we should recognize that soils are connected to plants almost as integrally as arteries are to veins in ani­ mals, and that our investigations need to embrace these connections rather than sever them…

The current condition of any soil is a legacy how previous processes altered the structure of the soil, with these changes in structure altering current rates of processes. The addition of organic matter alters the physical structure of soils, changing the three‐dimensional movement of water, oxygen and other chemicals over time. The processing of organic matter, and the nutrients and energy contained in organic matter, depends on organisms that are influenced by the supplies of water, oxygen and other chemicals, leading to further changes in soil structure. A comparison of the properties of unaltered parent material and a well‐developed soil reveals the cumulative history of feedbacks between chemical and biological processes and soil structure. When a factory begins the assembly of a new automobile, there is no uncertainty in the outcome of the assembly process. The development of living soils is almost entirely unlike this process. The transformation of parent material into soil depends on the interactions of so many important factors, each of which is variable with some degree of uncertainty, that soil development is always somewhat indeterminate and only broadly predictable. The influence of very dry or very wet environments can generally be expected to lead to

111

indurate, calcified soils or saturated peatland soils. But across smaller gradients in moisture regimes the contingent influence of features (such as plant and animal species that may or may not be present) can largely determine soil development. The indeterminate nature of living soils means that soils change over time, and that management activities can substantially alter soils (for better or worse).

LIFE IN THE FOREST SOIL BEGINS WITH ROOTS Hans Jenny also remarked that “trees and flowers excite poets and painter, but no one serenades the humble root, the hidden half of plants.” Roots can be excavated from soils, but the act of removing them for study breaks the intimate living connec­ tions with mycorrhizal fungi, the rhizosphere microbe community, and the rapid chemical reactions that are in part driven by organic molecules from the roots (Figure 6.1). Roots supply the connecting link between the plant and the soil, and studies of tree root systems are especially pertinent to forest soil science. Roots provide anchorage for the tree (Figure  6.2) and serve the vital functions of absorption and translo­ cation of water and nutrients. They exert a signifi­ cant influence on soil profile development by fueling most of the biogeochemistry within the soil, via the input of organic matter while roots are alive, when fine roots die over a period of months or years, and when whole root systems die (McClaugherty et  al. 1982; Fahey et  al. 1988). It should not be surprising that the growth and distri­ bution of roots are influenced by essentially the same environmental factors that affect growth of the aboveground portion of the tree. Not only do varia­ tions in the chemical, physical, and biological prop­ erties of the soil have profound effects on tree roots, but the influence of such factors as light intensity, air temperature, and wind may be reflected as much in root growth as in shoot growth. Around the world, trees typically have about 20% of their biomass hidden belowground (Cairns et al. 1997).

112

Chapter 6

Figure 6.1 Roots connect the soil system with the sky and sunlight, bringing the organic materials that shape soil

­structure and fuel biogeochemistry (Source: photo by Cristian Montes).

ROOT SYSTEMS HAVE CHARACTERISTIC FORMS AND ENORMOUS EXTENTS When grown under favorable soil conditions, tree species tend to develop distinctive root systems. Some send taproots deep into the soil; others tend to spread roots laterally in the upper soil; and some tree species do both (Figure  6.3). Large roots provide anchorage for the tree, and serve as core conducting members of the large number of feeder roots (Figure 6.4) that support the mycorrhizal net­ work where absorption of water and nutrients occurs. The actual root system developed by indi­ vidual trees is shaped in part by the microsite condi­ tions of soil texture, strength, available moisture, impeding layers, and nutrition (Sutton 1991).

Factors such as stand density, competition among individuals, soil micro‐relief, and soil rock content have substantial effects on the extension of lateral roots. Under the most favorable conditions, it is not uncommon for trees to possess some lateral roots that extend two to three times the radius of the crown (Zimmerman and Brown 1971; Stone and Kalisz 1991).

THE FORM OF ROOT SYSTEMS MAY BE RIGID OR FLEXIBLE Root systems can be characterized on the basis of rooting habit (relating to the form, direction, and distribution of the larger framework roots) and root intensity (the form, distribution, and number of

Life in Forest Soils

113

Figure 6.2  The immensity of root systems and their influences on soils are easy to overlook under typical conditions. One of the most severe windstorms in four centuries toppled this ancient western hemlock in British Columbia, Canada, revealing the root system that fueled the biogeochemistry of the soil and shaped soil structure and biotic communities for centuries. The pile of soil raised by the falling tree, as well as the pit, will persist for centuries or longer (see review by Šamonil et al. 2010).

114

Chapter 6

Tap root

Combination tap roots/lateral roots

Lateral root form

Figure 6.3  Diagrams of typical root structures, and photo of the root system of a two‐year‐old Eucalyptus tree in Brazil.

small roots). The habit, or form, of a root system is influenced by local site conditions, but also tends to be under some degree of genetic control. Taprooted trees are characterized by a strong downward‐ growing main root, which may branch to some degree. Such root systems occur in species of Carya, Juglans, Quercus, Pinus, and Abies. Strong laterals from which vertical sinkers grow downward, as in Populus, Fraxinus, and some species of Picea, characterize the flat, lateral root habit. Root systems with strong roots radiating diagonally from the base of the tree without a strong taproot are characteristic of the combination form (sometimes called “heart root” form). Such systems are found in species of Larix, Betula, Carpinus, and Tilia. The three‐dimensional “architecture” of root systems have been measured and modeled, revealing differences among species and responses to environmental factors (see Danjon and Reubens (2008) for a review). The rooting habit of a tree has considerable influ­ ence on the type of habitat in which it will thrive. Root form may determine whether or not a species is capable of fully exploiting a site and competing successfully with neighboring species. Root systems

All roots > 0.5 cm

All roots > 4 cm

Figure 6.4  Root system of a mature maritime pine tree, showing the major coarse roots (1600 roots >4 cm diameter) and all roots (6700 roots >0.5 cm) (Source: reprinted with permission from Danjon and Reubens (2008), figure 4).

Life in Forest Soils of species are under strong genetic control, and spe­ cies such as longleaf pine tend to retain their root­ ing habit regardless of the soil conditions to which they are subjected; consequently, they grow well only in a limited range of site conditions. On the other hand, a few species, such as red maple, can adapt their juvenile root systems to a variety of environments, and the species can become estab­ lished and grow on both wet and dry sites.

ROOTS PENETRATE MORE DEEPLY INTO SOILS THAN MOST INVESTIGATIONS SAMPLE Schenk and Jackson (2002) summarized the pat­ terns of rooting system depths around the world, and found that most investigations stop digging at less than half of the depth exploited by roots (Figure  6.5). In forests from the high‐latitudes to the tropics, the typical depth of the deepest tree roots is between two and three meters; much greater depths are not uncommon. Trees that develop strong taproots are capable of penetrating the soil to great depths for support and

0.0

Boreal

Temperate

Tropical

Rooting depth (m)

–0.5 –1.0 –1.5 –2.0 –2.5 –3.0

Commonly sampled Actual median

Figure 6.5  Common sampling for rooting depth stops at one or two meters for studies from boreal to tropical forests, but the typical depth of rooting is two to three meters (after Schenk and Jackson 2002).

115

moisture. For example, many pines, such as lon­ gleaf pine, have well‐developed taproots and are capable of surviving on deep, relatively dry, sandy sites. This well‐developed taproot form is also shared by a number of deciduous trees, exemplified by burr oak, black walnut, and many eucalyptus species (Figure  6.6). Many taprooted trees also have extensive laterals and sinkers that permit them to survive on shallow soils and soils with fluc­ tuating water tables. Although roots may not per­ sist in a zone of permanent water saturation, taproots of pitch and slash pine do develop below the depth of perched water tables (McQuilkin 1935; Van Rees and Comerford 1990). Cypress (Taxodium) and some other species common to swamps have special root adaptations that permit them to grow in saturated zones. If young taproots are injured, several descending woody roots may occur at the base of the tree in place of a single taproot. Taproots of slash pine are often fan‐shaped from the capillary zone to below the mean water table. The fan‐ shaped roots were reported by Schultz (1972) to be two‐ranked, profuse, and often dichotomously branched. Boggie (1972) reported that where the water table was maintained at the surface of deep peat, root development of lodgepole pine was con­ fined to a multitude of short roots from the base of the stems, giving a brush‐like appearance. Laterals developed only on trees where Sphagnum hum­ mocks had raised the soil surface above the mean water level. Tree species with inherently shallow or flat root systems and those with root systems under weak genetic control have particular advantages on shal­ low soils. Black spruce is an example of a species that can be found growing over a wide range of soil conditions, from peats with high or fluctuating water tables to deep sands. However, trees with shallow root systems may have a real disadvantage compared to other species on deep, easily pene­ trated soils, for they are poorly equipped to exploit these conditions and are more subject to wind throw than taprooted trees. Species possessing a combination (heart root) form with a number of lateral and oblique roots

116

Chapter 6

Height (m)

30 20 10

Depth (m)

0 –10 –20

0

1

2

3

4

5

6

Age (years) Root length (cm L–1) 0

0.1

1

10

100

1,000 10,000

–2 Soil depth (m)

–4

1.5 years

–6 –8

6 years

–10 –12 –14 –16 –18

Figure 6.6  The penetration of roots into the soil equaled about 85% of the reach of the canopy into the sky for this eucalyptus stand. The proliferation of fine roots declined exponentially with depth, dropping from more than 1,000 cm L−1 soil volume in the upper soil to 100 cm L−1 at 3 m. The plantations developed between 3 and 6 km of fine roots in each m2 (based on Christina et al. 2011, 2017; see also Figure 15.1).

arising from the root collar grow best on deep, per­ meable soils. However, they are also capable of exploiting fissures in fractured bedrock to a greater extent than other types of root systems.

SOIL CONDITIONS ALTER ROOT GROWTH The physical, chemical, and biological properties of soils have profound effects on the rate of root growth and development, root habit, and intensity

of root systems of established trees. Consequently, variations in root systems among individuals of a pliable species grown in different soils may be as great as those among different species grown on the same soil. Soil bulk density strongly influences root growth because of limitations on the ability of roots to pen­ etrate soils (soil strength), and because of influ­ ences on soil water supply and aeration. The bulk density that limits root penetration varies with tree species, soil moisture content, and soil texture (Sutton 1991). The range of bulk densities that has

Life in Forest Soils been reported to limit root penetration is rather wide, 1.1 kg L−1 in silty clay to about 2 kg L−1 in clay loam. However, in general, roots grow well in soils with bulk densities of up to 1.4 kg L−1, and ­significant root penetration begins to cease at bulk densities of around 1.7 kg  L−1. Basal tills and fragipans are examples of soil layers with high bulk density that often limit root growth. Soil strength is now often used to determine whether or not roots might penetrate a given soil layer. Soil strength is measured with a penetrometer. The values obtained with a penetrometer are determined not only by the bulk density and moisture content of the soil but also by the pore pattern of the soil and the shape and size of the penetrometer probe. Some roots are able to penetrate soil that has soil strength in excess of 3 MPa; however, the force that roots can actually exert is in the range of 0.50–1.5 MPa. Obviously, roots penetrate soil using mechanisms other than that of the penetrometer; nonetheless, soil strength is a convenient guide to root growth limitation. In medium‐ and fine‐textured soils, soil strength of 2.5 MPa, as measured by a penetrometer, curtails root extension. In coarse‐textured soils, the pores permit the entry of the root tip into crevices in which wedging action permits the root to elongate. Thus, roots can extend into coarse‐textured sand that has a soil strength of 5.0 MPa, as measured by a penetrometer. Texture and structure may influence rooting through physical impedance and by their influence on soil aeration. A soil does not need to become anaerobic before root growth and function are harmed. Oxygen concentration needs to be approximately 20% to avoid interfering with respiration at the root tip; however, oxygen concentrations as low as 10% may be sufficient to sustain respiration of older root tissues (Waisel et  al. 1996). Trees differ in tolerance to reduced aeration. Loblolly pine apparently has a greater tolerance of poor aeration than does shortleaf pine, and red and ponderosa pines are among the most sensitive of all North American conifers to low‐ oxygen conditions in the root zone.

117

Moisture has a greater influence on root devel­ opment and distribution than most other soil fac­ tors. Slow tree growth rates on shallow soils are not due directly to lack of space for root develop­ ment but rather to the limitation of the water (and nutrient) supply associated with shallow rooting. Lyford and Wilson (1966) suggested that red maple roots grow rather well over a broad range of soil texture and fertility conditions if the soil is maintained at near‐optimum moisture levels. Rooting habits change appreciably when maples are grown under very moist, poorly aerated condi­ tions. Kozlowski (1968) reported that large root systems develop in tree seedlings grown in soils maintained close to field capacity, in contrast to the sparse root systems found in soils allowed to dry to near the wilting point before rewetting. Many investigators have noted that small roots die almost immediately in local dry areas (Sutton 1991). Lorio et  al. (1972) found that mature loblolly pine trees on wet, flat sites were nearly devoid of fine roots and mycorrhizal roots com­ pared to neighboring trees on mounds. It is doubt­ ful that any significant amount of root growth takes place when the soil moisture drops to near the wilting point (Waisel et al. 1996), and growth may be exceedingly slow during the dry summer months that would otherwise be favorable for rapid root growth. Optimum soil moisture content for root develop­ ment depends on soil texture and temperature, among other factors, but soil moisture near field capacity is optimum for most species. Roots of most trees grow best in moist, well‐aerated soils, and they generally proliferate in layers providing the greatest moisture supply only if well aerated. Water saturation of the soil results in a deficiency of oxy­ gen and an accumulation of carbon dioxide. Such conditions usually result in reduced root growth and, eventually, in root mortality. Some species, such as bald cypress and water tupelo and certain species of willow and alder, are capable of obtaining oxygen and growing in saturated soils; however, roots of most tree species will not survive long under such conditions.

ROOT DEVELOPMENT DEPENDS ON SOIL CHEMISTRY Acidity and nutrient deficiencies can restrict plant root growth and development. Acid soil may inhibit roots because of the toxicity of aluminum, the solu­ bility of which increases as acidity increases. However, the influences of organic compounds, in the soil solution at large and in the rhizosphere, may limit problems with dissolved aluminum. Root toler­ ance to acidity differs widely among plant species, and many tree species are very tolerant to acid con­ ditions (Fisher and Juo 1995; Tilki and Fisher 1998). Nutrient availability in the soil affects both the growth rate and distribution of roots. Roots (and their mycorrhizal hyphae) proliferate in zones of high nutrient availability. These increases in rooting density are due to the increased biological activity that takes place in these zones rather than to the tree’s activity. Increases in site fertility, both along geographic gradients and in response to fertilization, typically lead to increases in aboveground tree growth but constant or declining belowground growth (Litton et al. 2007; Högberg et al. 2010).

ROOTS CAN SHAPE SOIL PHYSICAL PROPERTIES Tree roots grow in a variety of forms and to varying degrees of intensity, governed to a large extent by genetic forces but modified by local conditions of soil and site. The influence of these latter factors on the nature and abundance of roots is very diverse and difficult to class into average conditions because of the challenges of studying roots under realistic field conditions. The fact that woody plant roots are per­ ennial, penetrate the soil to great depth, and have the ability to concentrate in soil layers most favora­ ble to growth makes deeper horizons much more important in forest soils than in agronomic soils. Other functions of roots that are often overlooked, but that are of tremendous importance, are those related to stability and development of the soil. Tree roots provide a significant stabilizing force in mountainous areas where soil is subject to erosion

Root tensile strength (MPa)

Chapter 6 35 30 25

Nothafagus

20 15

Pinus

10 5 0

0

1

2

3

4

Years since harvest 20 Sheer stress (kPa)

118

16 Uncut forests

12 8 4 0

Clearcuts 0

20

40 60 Displacement (mm)

80

100

Figure 6.7  Tensile strength curves for decaying roots of

Nothofagus and pine show the loss of strength over several years (upper; O’Loughlin and Watson 1981), and the shear stress displacement curves from field tests of the upper 5 cm of soil in a Nothofagus forest (Source: O’Loughlin et al. 1982, used with permission).

or mass movement. Fine roots and fungal mycelia serve as binding agents for surface soil, and where larger roots penetrate below the surface horizons, they can anchor the soil mantle to the substrate (Swanston and Dyrness 1973). The number of land­ slides after harvesting typically increases for three to five years because this is the time frame for roots to decompose enough to lose tensile strength (Figure 6.7). The harvesting of Japanese cedar for­ ests in Japan led to a strong increase in landslide frequency shortly after harvesting as the strength of stump roots declined; recovery of root strength with regrowing trees dramatically lowered the frequency of landslides (Figure 6.8, Imaizumi et al. 2008). The value of living tree roots in anchoring shal­ low soils to the underlying subsoil is particularly important in small drainage areas, where winter storms can cause a sharp rise in groundwater level. Clear‐felling further contributes to a decline in soil

119

12

5 4

10

Total root strength

8

3

6 2 1

4

Landslide frequency

New root strength

Root strength (kPa)

Frequency (landslides km–2 yr–1)

Life in Forest Soils

2

Decaying root strength 0

0 0

10

20 30 40 Time since harvest (years)

50

Figure 6.8  The frequency of landslides following harvesting of Japanese cedar (Cryptomeria japonica) forests is very high shortly after harvesting, as decay of stump roots weakens slopes. Slope failures decline as new roots increase the stability of the slope (Source: modified from Imaizumi et al. 2008).

retention by roots through greater exposure of the soil surface and increased soil moisture levels fol­ lowing removal of the intercepting and transpiring tree canopy. The form and extent of the tree root system also influence the amount of soil disturbance resulting from windthrow. Soil mixing and the development of a pit and mound microrelief are  brought about by the uprooting of indivi­ dual trees during storm periods (Figure  6.2). Windthrows also influence the opportunities for soil animals. Kooch et  al. (2014) found that a beech forest in northern Iran had a density of windthrow mounds and pits of about one per ha, and twice as many earthworms were present per m2 of mound than in the general soil (worms were absent from pits).

MYCORRHIZAL FUNGI CONNECT PLANT ROOTS WITH THE LIVING SOIL In reality, the fine roots of trees live in a milieu determined by symbiosis with fungi rather than directly in the solid and liquid phases of the soil (Brundrett et  al. 1996; Smith and Read 2008). In 1885, Albert Bernhard Frank, a German forest pathologist, coined the term mycorrhiza (fungus root), to denote these particular associations of

roots and fungi. The morphology of fine roots and mycorrhizae varies among plant and fungal species; individual tree species may form these associations with hundreds of species of fungi. Mycorrhizal fungi also mediate the connections between trees (the primary providers of carbon for the soil community) and the free‐living soil com­ munities. The complexity of possible interactions is hard to imagine, and even some simple overall expectations may not work (see for example Moore et al. 2015; Kaspari et al. 2017).

TWO CLASSES OF MYCORRHIZAE ARE PARTICULARLY IMPORTANT FOR TREES Mycorrhizae are divided into two broad classes based on the spatial interrelation of thread‐like fungal hyphae and root cells. Ectomycorrhizae are characterized by fungal hyphae that penetrate the spaces between cortical cells and usually form a compact mantle around the short roots (Figure 6.9). Arbuscular mycorrhizae are characterized by hyphae that penetrate directly into the cells of the root cortex and do not form a fungal mantle (Figure 6.10). Two other types of mycorrhizae (eri­ coid, and those that associate with orchids) also occur in forests.

Fungal mantle

Hartig net

Root vascular stele

Root endodermis

Root cortex

Lateral root

Figure 6.9  Graphic presentation (not to scale) of an ectomycorrhiza (after Ruehle and Marx 1979).

Large spores

External vesicles Root hair

Cortex

Stele

Arbuscules

Vesicles

Figure 6.10  Graphic presentation (not to scale) of an arbuscular mycorrhiza (Source: from Nicolson (1967), used with

permission).

Life in Forest Soils Ectomycorrhizae are restricted almost entirely to trees, and they occur naturally on many forest spe­ cies. These fungi have been present for at least 50 million years based on fossil evidence, and probably more than 100 million years based on patterns of plant taxonomy and distribution (Smith and Read 2008; van der Heijden et al. 2015). More than 5,000 species of ectomycorrhizal fungi are estimated to exist, and most of these are Basidiomycetes, but some also belong to the Ascomycetes. These fungi show a low level of specificity for host plants, and a single tree may host dozens of species of ectomycorrhizae. The fruiting bodies of ectomycorrhizae produce spores that are readily and widely disseminated by wind, water, and animals. Ectomycorrhizae are characteristic of the families Pinaceae, Fagaceae, and Betulaceae. Eucalyptus and some tropical hard­ wood species are largely ectomycorrhizal, while angiosperm families such as Salicaceae, Juglandaceae, Tiliaceae, and Myrtaceae may be either ectomycor­ rhizal or arbuscular mycorrhizal. Ectomycorrhizal infection is initiated from spores or hyphae of the fungal symbiont in the rhizos­

121

phere of feeder roots. Contact between hyphae of a mycorrhizal fungus and a compatible short root may originate from spores germinated in the vicin­ ity of the roots, by extension through the soil of hyphae from either residual mycelia or established mycorrhizae, or by progression of hyphae through adjacent internal root tissue. Fungal mycelia usu­ ally grow over the feeder root surfaces and form an external mantle or sheath (Figure 6.11). Following mantle development, hyphae grow between cells without actually penetrating cell walls, forming a network of hyphae around root cortical cells. This “Hartig net” of hyphae may completely replace the middle lamellae between cortical cells, and is the major distinguishing feature of ectomycorrhizae (Kormanik et  al. 1977). Ectomycorrhizae may appear as simple, unforked feeder roots or as bifur­ cated, muitiforked coralloids, nodule‐shaped modi­ fications of feeder roots. Hyphae on short roots radiate from the fungal mantle into the soil, greatly increasing the absorbing potential of the roots. Roots that are infected with ectomycorrhizae are comprised of about 60–90% of plant mass, and

Figure 6.11  Ectomycorrhizal fungus forming a mantel on short roots, with hyphae bundled into mycelium that permeate

the soil, obtaining water and nutrients for the tree in exchange for carbohydrates (Source: photo courtesy of Kjell Olofsson).

122

Chapter 6

10–40% of fungal mass (Courty et  al. 2010). The ­ iomass of ectomycorrhizal fungus might comprise b about one‐third of the total soil microbial biomass (Högberg and Högberg 2002). The sheaths of fungal tissues surrounding roots are joined to vast net­ works of “extrametrical” hyphae that permeate soils. Many coniferous soils experience about 0.5– 1.0 Mg ha−1 year−1 of growth of these extrametrical hyphae, and individual hypha may persist for sev­ eral years (Ekblad et al. 2013). About one‐quarter of the CO2 efflux from soils comes from respiration by ectomycorrhizae, accounting for 3–15% of total forest photosynthesis (Courty et al. 2010). Ectomycorrhizae sometimes form fascinating “mats” in conifer forests. In a classic study, Hintikka and Näykki (1968) explored mats formed by an ecto­ mycorrhizal fungus in pine forests in Finland. The mats were described as being 1–5 cm thick, creating a felt‐like layer within the O horizon that sometimes extended into the mineral soil. The hyphae knit together the mats so well that they can be lifted up intact from the soil. The mats were 1–5 m2 in size (expanding at rates of 2–3 cm year−1), covering 2–4% of the ground area. What is special about these mats? They tend to be hydrophobic, with water ponding on top of the soil rather than infiltrating (especially visi­ ble after snowmelt). The rate of CO2 efflux from mats was two to four times the rate from non‐mat O hori­ zons, indicating the mats were hotspots of biological activity. Mineral soils beneath the mats had E hori­ zons (with iron and aluminum removed) that were much more pronounced than under the non‐mat O horizons. The O horizon within the mats was thinner, having lost about 1/3 of the mass found in the non‐ mat O horizon. Finally, the O horizon mats appeared to have lower N and P, as a result of (presumed) removal from the organic pools by the mycorrhizae. Arbuscular mycorrhizae are the most wide­ spread and important root symbionts, occurring throughout the world in both agricultural and for­ est soils. The fossil record shows these fungi can be traced back more than 400 million years; all spe­ cies are obligate symbionts (none have the sort of saprophytic capacity that would allow them to subsist independent of plant connections). The

number of species of arbuscular mycorrhizae is quite small, with perhaps 150 identified species for hundreds of thousands of species of plants (Smith and Read 2008). Among the forest tree genera with these types of mycorrhizae are Acer, Alnus, Fraxinus, Juglans, Liquidambar, Liriodendron, Platanus, Populus. Robinia, Salix, and Ulmus (Gerdemann and Trappe 1975). Arbuscular mycorrhizae of trees are produced most frequently by fungi of the family Endogonaceae. They generally form an extensive network of hyphae on feeder roots and extend from the root into the soil, but they do not develop the dense fungal sheath typical of ectomycorrhizae. Arbuscular mycorrhizae form large, conspicuous, thick‐walled spores in the rhizosphere, on the root surface, and sometimes in feeder root cortical tissue (Kormanik et al. 1977). The fungal hyphae pene­ trate epidermal cell walls of the root and then grow into cortical cells, where they develop specialized absorbing structures called arbuscules in the cyto­ plasmic matrix. In some instances, the fungus com­ pletely colonizes the cortical region of the root, but it does not invade the endodermis, stele, or meris­ tem (Gray 1971). Thin‐walled, spherical vesicles may also be produced in cortical cells. Arbuscular mycorrhizal infection does not result in major mor­ phological changes in roots; therefore, the unaided eye cannot detect it. The fungi that form arbuscular mycorrhizae do not produce aboveground fruiting bodies or wind‐disseminated spores, as do most ectomyc­ orrhizal fungi. Arbuscular mycorrhizae produce sporocarps belowground, and the spread of arbuscular mycorrhizae is not as rapid as that of ectomycorrhizae.

TYPES OF MYCORRHIZA MAY DETERMINE THE EFFECTS OF TREE SPECIES ON SOILS The influence of tree species on soils may depend on whether a species forms associations with ecomycorrhizae or arbuscular mycorrhizae (see ­

Life in Forest Soils Figure 9.13, Phillips et al. 2013; Lin et al. 2017). The general pattern seems to be that ectomycor­ rhizal species build larger O horizons, and arbus­ cular species lead to organic‐rich A horizons (in part by fostering higher animal activity). Ectomycorrhizal species may rely more heavily on organic‐N compounds, and arbuscular species may show higher rates of N mineralization and production of inorganic ammonium and nitrate. More work is needed to see how strong these gen­ eralizations are, and how often local cases prove to be exceptions to the generalization. For exam­ ple, a common garden experiment showed that ectomycorrhizal Norway spruce did indeed show lower net N mineralization than arbuscular myc­ orrhizal green ash, but the species fostering the highest net N mineralization was ectomycorrhizal white pine (Valentine and Binkley 1992). Was this example of greater differences between ectomyc­ orrhizal species due to the particular soil of this common garden, or to these particular tree species as representative of each class of mycorrhizal part­ ners? The answer for this case isn’t known, but questions about generalizations and case‐specific details are always important for forest soils.

SOIL FACTORS AFFECT MYCORRHIZAL DEVELOPMENT Environmental factors influence mycorrhizal development by affecting both tree roots and fun­ gal symbionts. After the formation of a receptive tree root, the main factors influencing susceptibility of the root to mycorrhizal infection appear to be photosynthetic potential and soil fertility (Marx 1977). High light intensity and low to moderate soil fertility enhance mycorrhizal development, while the opposite conditions may reduce or even pre­ vent mycorrhizal development. The effects of soil fertility and fertilizer addi­ tions on the development of ectomycorrhizae appear to vary with the original fertility of the soil and the nutrient content of the host plant. Mikola (1973) reported that ectomycorrhizal formation

123

on white pine seedlings is stimulated by applica­ tions of phosphate fertilizers to soils containing a low population of relatively inactive mycorrhizal fungi. It appears that growth suppression of pine following nitrogen applications to phosphorus‐ deficient soils results from a reduction in mycor­ rhizal development (and phosphorus absorption) caused by the high nitrogen: phosphorus ratio in host plant tissue. Apparently all mycorrhizal fungi are obligate aer­ obes, and mycelial growth likely decreases as oxy­ gen availability decreases. The requirements of mycorrhizal fungi for nutrient elements are proba­ bly not very different from those of the host plants, although little research has been conducted on this point. The formation of ectomycorrhizae on tree roots is often greatest under acidic conditions, though this may be due to low nutrient availability rather than to soil acidity. Richards (1961) con­ cluded that poor mycorrhizal formation in alkaline soils was due to nitrate inhibition of mycorrhizal infection rather than alkalinity per se. However, Theodorou and Bowen (1969) reported that alka­ line conditions in the rhizosphere severely inhib­ ited the growth of some types of mycorrhizal fungi, apart from the possible effect of nitrate inhibition of infection. They further stated that nitrate inhibition of mycorrhiza formation under acid conditions is mainly due to a paucity of infection, not to poor fungal growth. The presence of antagonistic soil microorgan­ isms can influence survival of the mycorrhizal symbiont as well as root growth of the host plant. Fungicides used in plant disease control can inhibit mycorrhizal fungi under some conditions, or they may stimulate mycorrhizal development by reduc­ ing microbial competition. Eradicating ectomycor­ rhizal fungi from nursery soils by fumigation is generally not a problem because these fungi pro­ duce wind‐disseminated spores that soon colonize the soil. The soil‐borne spores of most arbuscular fungi are eliminated by fumigation of nursery soils, and artificial inoculation is essential if spe­ cies requiring arbuscular mycorrhizae are to be grown successfully.

124

Chapter 6

MYCORRHIZAE BENEFIT TREES LARGELY BY INCREASING SURFACE AREA FOR ABSORBING WATER AND NUTRIENTS The dependence of most species of forest trees on mycorrhizae to initiate and support healthy growth has been most strikingly illustrated by the problem encountered in introducing trees in areas devoid of symbionts. Kessell (1927) failed to establish Pinus radiata in western Australian nurseries that lacked mycorrhizal fungi. The seedlings grew normally only after soil from a healthy pine stand was added to the beds and ectomycorrhizae formed. Similar results have been obtained where exotic species have been used in the Philippines, Rhodesia, New Zealand, South America, and Puerto Rico (Vosso 1971). Other areas where ectomycorrhizal trees and their symbiotic fungi do not occur naturally include former agricultural soils of Poland, oak shelterbelts on the steppes of Russia, and the formerly treeless plains of North and South America (Marx 1977). The observed benefit of mycorrhizae in the growth and development of trees largely results from the proliferation of absorbing surface area that provides trees with greater access to soil water and nutrients. Hyphae have much smaller diameters than roots, providing more than an order of magnitude increase in absorbing surface area per g of mass; small diam­ eter hyphae are also more effective at absorbing ions from soil solutions (see Figure 7.11). Ectomycorrhizae may also provide access for trees to organic forms of N dissolved in soil solu­ tions. For example, Turnbull et al. (1995) showed that eucalyptus seedlings without mycorrhizae could use ammonium, nitrate, and organic glu­ tamine, but no other form of organic N. Inoculation with fungus allowed the seedlings to utilize a vari­ ety of amino acids and even soluble protein. Recent work has shown that organic forms of N may com­ prise the majority of N taken up by trees (Inselsbacher and Näsholm 2012; see Figure 7.17). Can decomposition be enhanced by mycorrhizae, gaining access to nutrients for trees? This question has two potential mechanisms. The provision of

readily available C does seem to enhance the a­ ctivity of decomposers (saprotrophs), and mycor­ rhizae‐infested roots may support a host of microbes in the root “mycorhizosphere.” Some mycorrhizae play a direct role by increasing the proteolytic activ­ ity in decaying organic matter; indeed, ectomyc­ orrhizae appear to span the entire gamut from obligate symbionts (receiving all their carbon from trees) to largely saprophytic (obtaining carbon from dead organic matter; Courty et al. 2010). In a recent review, Lindahl and Tunlid 2015 concluded the evidence is strong for mycorrhizal fungi being able to degrade some organic compounds in soils, but they remained skeptical that these fungi could actually subsist on detritus as their sole C source. Saprophytic fungi may be more effective in rela­ tive fresh litter (with broad C:N compounds), whereas ectomycorrhizae may be more important in processing of older, narrower C:N materials. For example, Lindahl et al. (2007) used DNA and iso­ tope techniques to examine the spatial pattern of fungi in a Scots pine soil. Saprophytic fungi were found only in relatively fresh litter in the O hori­ zon, where very few mycorrhizal hyphae devel­ oped. The lower O horizon and the mineral soil horizons were dominated solely by mycorrhizal fungi (see Figure 4.3). The overall outcome of the influence of ectomy­ corrhizae on decomposition may be mixed. Sterkenburg et  al. (2018) trenched around small plots in a forest of pine and spruce in Sweden, effectively removing the influence of ectomycor­ rhizae within the plots. They found that decompo­ sition in the upper O horizon increased in the plots, perhaps as a result of higher N supply to saprobes when competition with mycorrhizal hyphae was reduced. However, the lower O horizon showed a reduction in activity of oxidizing enzymes, indicat­ ing a reduction in the active role of ectomycorrhi­ zae in decomposition. Mycorrhizae appear to be able to access both organic and inorganic pools of P in soils. The high absorbing area of the fungal hyphae may be most  important, but other mechanisms favoring P acquisition include the presence of phosphatase

Life in Forest Soils enzymes (to break down organic‐P), localized acidification, and excretion of anions such as ­oxalate that may bind with cations and release P from inorganic salts (Smith and Read 2008). The apparent enhancement of mineral weathering and dissolution of insoluble salts may depend on the community of bacteria found in the myco‐rhizosphere (Courty et al. 2010; see Figure 4.3).

CONNECTIONS MAY LINK MULTIPLE TREES THROUGH A COMMON MYCORRHIZAL NETWORK A single genetic individual (genet) of a mycorrhizal fungus may extend for several meters through a forest soil, linking with more than one tree. Individual trees may also be infected with more than a single fungus (or species of fungus), creating a common mycorrhizal/tree network at the scale of tens to hundreds of square meters (Courty et  al. 2010). Isotopic tracer studies have demonstrated flows of carbohydrate from one tree, through the common mycorrhizal network into another tree, and this sort of community may foster establish­ ment of seedlings that can tap into the rich resource network. The actual magnitude of flows of water and nutrients between plants is difficult to deter­ mine, and issues of competition versus facilitation between plants need to be worked out for these networks. At this point, these interconnections provide a fascinating explanation for how mycor­ rhizal fungi may impede nitrogen nutrition of trees on poor sites.

SYMBIOSES DO NOT ALWAYS ENRICH BOTH PARTNERS: MYCORRHIZAE CAN IMPEDE TREE GROWTH Systems with good designs and engineering can maximize performance, such as the fuel efficiency of a car. If car mileage is an important criterion for consumer choices among car models, car models

125

with lower efficiency would be at a disadvantage and might even disappear from the market. The  complex components of cars must all work together smoothly to achieve overall fuel efficiency; there is no place for one component of the car “competing” for energy at the expense of another component. Should the symbiosis of tree roots and mycor­ rhizal fungi be thought of in the same way? It might seem to make no sense for the tree to benefit more than the fungus, or the fungus to benefit more than the tree. The scientific method consists of developing logically plausible ideas (such as this expectation about symbiotic balance) and chal­ lenging the ideas with evidence. Fascinatingly, evi­ dence from forests seems to disprove this expectation of balance, as the benefit does not always appear to be strongly mutual. Näsholm et  al. (2013) examined how 15N added to soil moved from mycorrhizal fungi into trees. Two types of plots were examined, one very low in N supply and the other higher as a result of previous fertilization. The mycorrhizal fungi in the low‐N plot showed a high enrichment of 15N, indicating the fungi retained much of the added N for its own use rather than pass it to the tree (Figure 6.12). In the high‐N plot, the 15N was not retained in the fungi. Needles from the pine trees showed the opposite pattern, as mycorrhizal fungi passed on large amounts of 15N in the high‐N plot, but trees received little of the 15N in the low‐N plot. Why would trees provide carbohydrates to ­support mycorrhizal fungi on N‐poor sites where the mycorrhizae hoard a large share of N for their own use? Why don’t trees stop supporting the fungi, or form associations with other species of mycorrhizal fungus that are more generous in shar­ ing N? In complex systems such as these, questions of “why?” may not have simple or clear answers. Franklin et  al. (2014) came up with a plausible explanation that lines up well with the limited evidence currently available from experiments. ­ Trees can be considered to have “options” for investing carbohydrates to obtain nutrients: invest in fine roots, or in a varied distribution of species of

Chapter 6 70 excess in neeles (mg kg–1 org matter)

Not fertilized

4 3 2

Fertilized

1 0

0

5

15N

15N excess in fungi (mg kg–1 org matter)

126

10

15 Days

20

25

30

Fertilized

60 50 40 30

Not fertilized

20 10 0

0

10

20 Days

30

Figure 6.12  15N was used to trace how N was retained by mycorrhizal fungi or passed on to Scots pine trees in plots that were low in N (not fertilized) or high in N (previously fertilized). Mycorrhizal fungi held onto much more N in the low‐N plot, passing less on to the trees (Source: from data in Näsholm et al. 2013).

mycorrhizal fungi. When soils have low concentra­ tions of nutrients, trees would generally obtain more nutrients by investing in mycorrhizal fungi, but in higher‐N soils it may be better to simply invest in growing their own fine roots. A key to this overall model is that each tree typically has a “choice” of multiple species of mycorrhizal fungi, so a tree can “choose” a species that provides the best return of nutrients for each unit of carbohydrate passed to the fungus. The fungi also have a “choice” of passing nutrients on to trees, or retaining nutri­ ents for their own growth and reproduction. The network of fungal hyphae in soils includes connec­ tions with more than one tree for each fungus, so the fungus also gets to apportion the nutrients among competing trees. The overall story is similar to a market economy, with features of efficiencies, trade‐offs, and exchange rates (see also Bender et al. 2014; van der Heijden et al. 2015). The overall conclusion from this story is that trees depend more heavily on mycorrhizal fungi (and less on their own roots) as soil nutrient supplies decline, and trees remain vested in supporting mycorrhizal fungi even when the symbiosis shifts so strongly in favor of the fungus that tree growth is reduced. This is clearly an arena with potential for new ideas, and many more experiments to challenge these ideas. Indeed, the topic of unequal benefits and costs for organisms engaged in symbioses is a hot topic across many fields of biology (see Yong 2016;

Jiang et  al. 2017). The frontiers for research are even broader when c­ onsidering the interactions of symbioses with conjunction with the full suite of organisms in soils.

ALMOST ALL THE DIVERSITY IN A FOREST IS FOUND IN THE SOIL Soils are fundamental to forest ecosystems, as they form a reservoir for available water and nutrients, and anchorage for supporting massive canopies. Less apparent is the contribution of soils to the genetic biodiversity of forests; the diversity of plant species is barely a footnote in the complete community of organisms inhabiting the forest! A diverse tropical rain forest might contain more than 200 tree species in a hectare, plus a few 1,000 species of understory plants, epiphytes, arboreal arthropods, reptiles, amphibians, birds, and m ­ ammals. The biodiversity in the soil is almost inconceivably richer than this. The concept of “species” applies well to some soil organisms, such as soil animals that pass genes through the generations by mating between males and females. Many small animals may be difficult to assign to traditional species, owing to very limited information about their biology and ecology. The species concept is even more difficult to apply to the microbial world, as genes can be transferred between unrelated types of cells. Modern DNA techniques

Life in Forest Soils can use a “barcode” approach, where operational taxonomic units (OTU) are determined based on the level of shared versus unique segments of DNA between organisms (Bardgett and van der Putten 2014; Orgiazzi et al. 2016). Fierer et al. (2007) used this DNA barcode/OTU approach to evaluate the biodiversity of a rainforest soil. They roughly esti­ mated the soil community contained 20,000 OTUs of bacteria, 1,000 OTUs of archaea, 2,000 OTUs of fungi, and more than 10 million OTUs of viruses. Other chemical methods are commonly used to assess the diversity and function of soil microbes. Phospholipid fatty acid analysis (PFLA) has been used to measure bacteria and fungi in the soil (Frostegård et al. 1993; Frostegård and Bååth 1996; Bardgett and McAlister 1999; Kaur et  al. 2005; Strickland and Rousk 2010). The incredible diversity of the soil community forms an intricate food web through which nutrients and energy pass (Coleman and Crossley 1996; Moore and de Ruiter 2012; Paul 2015; Moore et al. 2017). The importance of the processes within this web to the maintenance of the soil’s productive potential cannot be overestimated (Fisher 1995b; Baldrian 2017; Lladó et al. 2017), though the malleability of food webs may ensure sustained processing of matter and energy despite major impacts of disturbances or management. Food webs have been constructed for few soils, but it is clear that cultural practices applied to the soil can dramatically alter their structure. The mass of the microbial community changes through seasons, influenced by temperature, mois­ ture, and the supply of substrates and nutrients. A fourfold or greater range may be found in temper­ ate forests, along with shifts in the mass of N con­ tained in the microbes, and several‐fold shifts in the C:N (Kaiser et al. 2011). These shifts are driven in large part by current (and recent) photosynthesis by trees, and the rapid allocation of varying amounts of carbohydrates to the belowground por­ tion of the ecosystem (see Högberg et  al. 2010). Huge changes in the biomass of the soil microbial community lead to remarkable variations in the structure of soil food webs, and the rates of processing of energy and matter.

127

Perhaps the biggest challenge, and opportunity, is to take the “big data” opportunities of recent and emerging analytical techniques and extract mean­ ing relevant to forest soil processes (Jansson and Hofmockel 2018). Wilhelm et al. (2017) analyzed the microbial communities from long‐term forest productivity sites across North America, identify­ ing patterns in community composition related to both geography and harvesting treatments. They concluded: Biogeographical differences in the distribution of taxa as well as local edaphic and environmental conditions produced substantial variation in the effects of harvesting. This extensive molecular‐ based investigation of forest soil advances our understanding of forest disturbance and lays the foundation for monitoring long‐term impacts of timber harvesting.

The first sentence is straightforward, but the asser­ tion of the second sentence remains to be established.

BACTERIA AND ARCHAEA ARE SMALL, BUT WITH HUGE POPULATIONS Although bacteria and archaea are small, typically between 0.1 and 10 μm in size, they are especially prominent in soils because of their large total bio­ mass (1–10 Mg ha−1, Brady and Weil 2002) and numbers (on the order of 1018 ha−1). The diversity of bacteria might be expected to show patterns in relation to latitude, temperature, or moisture regime; however, the best association seems to be with soil acidity (Figure  6.13). At a smaller scale and across the full suite of the soil microbial com­ munity, other approaches for characterizing micro­ bial communities (phospholipid fatty acid profiles and enzymes) have shown notable differences among forests that relate to water regime and tem­ perature (Brockett et al. 2012). Bacteria, archaea and fungi dominate in well‐aer­ ated soils, but bacteria and archaea alone account for

Chapter 6

128

Phylotype richness

50 40 30

r2 = 0.6, P 2 mm) are large enough to disrupt soil structure through their activity and include small mammals, earthworms, termites, ants, and other arthropods (Dindal 1990; Stork and Eggleton 1992; Berry 1994). These organisms create macropores that influence infiltration rate and gas exchange. They are responsible for a great deal of soil mixing and

131

can profoundly influence horizonation and rooting pattern. They often consume very high carbon: nitrogen ratio organic material and produce lower carbon:nitrogen detritus. mm) are large enough to Mesofauna (0.1–2  escape the surface tension of water and move freely in the soil but small enough not to disrupt soil structure through their movement. They include some mollusks, springtails, mites, wood lice, and other arthropods. These organisms are extremely important in decomposition and mineralization processes (Crossley et al. 1992). Microfauna (5 Proportion of value for species with lowest value

25 20

A horizon pH

15 10 5 0 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10

30

N content of litterfall

25

Percent of comparisons (n = 44)

Litterfall mass

40

30

Percent of comparisons (n = 47)

60 50

Percent of comparisons (n = 18)

Percent of comparisons (n = 34)

Tree Species Influences

pH units higher than species with lowest pH

Figure 9.9  Lowest values reported within common gardens were set at 1.0, and values for other species calculated as proportions of this lowest value. For pH, species effects were tabulated in pH units above the lowest value found for any species within the common garden (Source: modified from Binkley and Giardina (1998); with additional comparisons from Menyailo et al. (2002); Hagen‐Thorn et al. (2004); Reich et al. (2005); Hansen et al. (2009); Hansson et al. (2011)).

216

Chapter 9

Net N mineralization (mg kg–1 day–1)

25

r2 = 0.10 P > 0.7

20 15 10 5 0 3.5

4.5

5.5

6.5

Soil pH Figure 9.10  Soil net N mineralization from 11 species in a common‐garden species trail at Clonsast Bog, Ireland, showed no relationship with the soil pH that developed under the influence of each species (Source: modified from Prescott et al. 1995).

(acid quantity) under alder (Figure  9.11). The dissociation of the acid complex was 41% in Douglas‐fir soil but only 29% in red alder soil. If the alder soil had the same total quantify of acid as the Douglas‐fir soil, its pH would have been higher (5.2). If the acid strength of the alder soil were adjusted to match that of the Douglas‐fir soil, the pH would also have risen to about 5.2. If the acid quantity in the Douglas‐fir soil were

increased to match that of the alder soil, the pH would have dropped from 5.2 to about 4.8. If the acid quantity in the Douglas‐fir remained the same but the strength matched that of the red alder soil, the pH would have dropped to 5.0. These calculations indicate that the most important difference between these soils was the quantity of acid in the soil (which altered the degree of dissociation), with the greater acid strength playing an additional role. Determining the pattern of tree species effects on  soils is a first step that can be followed by examining the biogeochemical mechanisms by ­ which species alter soils. A common garden experiment in Wisconsin, USA showed net N mineralization was more than twice as high under larch as under other species (Figure  9.12). The difference corresponded to a very low ratio of lignin:N in aboveground litterfall, which might explain rapid decomposition and N turnover. However, the ­pattern was just as strong (or stronger) for fungal biomass as a driver of net N mineralization. A strong (inverse) covariance between fungal and bacterial biomass indicated that bacterial biomass would be equally related to the pattern of species influence on N mineralization. The multiple covarying factors still prevent clear insights about direct causes for any observed effects.

Table 9.3  Explanations for how tree species differed in soil pH.

Comparison

Difference in pH

Relative ranking of factors

Reference

50 year Norway spruce vs. white pine 50 year Norway spruce vs. green ash 50 year white pine vs. green ash 8 year Eucalyptus vs. Falcataria 55 year old conifer/alder vs. conifer, rich site 55 year old Douglas‐fir/ alder vs. Douglas‐fir 50 year old Douglas‐fir vs. red alder

Spruce pH 3.8 Pine pH 4.2 Spruce pH 3.8 Ash pH 4.6 Pine pH 4.2 Ash pH 4.6 Eucalyptus pH 4.5 Falcataria pH 4.3 Conifer pH 3.9 Alder/conifer pH 3.6 Douglas‐fir pH 4.0 Alder/Douglas‐fir pH 4.0 Douglas‐fir pH 5.2 Red alder pH 5.0

Strength>Dissociation>Quantity

Binkley and Valentine 1991

Strength>Dissociation>Quantity Dissociation>Strength>Quantity Dissociation>Strength>Quantity

Rhoades and Binkley 1995

Dissociation>Strength>Quantity

Binkley and Sollins 1990

Increased dissociation under alder offset by increased acid strength Quantity>Strength>Dissociation Calculated from Van Miegroet et al. 1989

Tree Species Influences 6.5 6.0

Red alder

Soil pH

5.5 5.0

Douglas-fir

4.5 4.0 3.5 3.0 –50

–20

10 OH– added

40 (mmolc

70

100

kg–1)

Figure 9.11 Titration curves for soils from adjacent stands of Douglas‐fir and red alder (Source: data from Van Miegroet et  al. 1989). Both species had similar ANC to pH 4, representing the same level of exchangeable base cations. Alder soil had twice the BNC to pH 5.5, indicating much larger pools of soil acids.

DO CONIFERS DIFFER CONSISTENTLY FROM BROAD‐LEAVED TREE SPECIES IN SOILS EFFECTS? Science advances substantially when generalizations allow us to apply knowledge from past experiments to predict the outcome in situations not yet measured.

As mentioned in several places above, there is a long history of ideas about how categories of ­species differ in soil effects. Recent meta‐analyses have tallied the evidence supporting some of these possible generalizations. Boča et  al. (2014) examined 77 studies of differences in soils beneath conifers and broadleaved species, spanning 28 countries. Conifer forest had 38% larger stores of C in O horizons, but similar C stocks in mineral soils. The sum of O horizons and mineral soils resulted in an overall 14% greater C storage in soils under conifers. Some of the studies in the meta‐analysis were not from common gardens, so factors that influenced the presence of tree species on each site might ­confound any apparent influence of species. Augusto et  al. (2015) did a similar analysis but restricted the studies to common garden designs. They also omitted larch from the conifer group, which is winter‐deciduous and often differs substantially in soil effects relative to other conifers. Soils were typically drier under conifers, as a result of higher interception loss from the larger leaf area of conifer forests. O horizons were more massive under conifers, about double the mass of O horizons under broad‐leaved species. Mineralization of N may have averaged about 140

r2 = 0.76 P = 0.05

120

European larch

100 White pine

80 60

Red oak

40

Red pine

Norway spruce

20

In situ N mineralization (kg ha–1 yr–1)

In situ N mineralization (kg ha–1 yr–1)

140

0

217

120

10

20 Leaf litter lignin:N

30

40

European larch

100 White pine

80 60

Red oak Red pine

40

Norway spruce

20 0

0

r2 = 0.89 P = 0.02

0

500

1,000

1,500

Total fungal biomass (mg/kg)

Figure 9.12  Annual net N mineralization related strongly to the lignin:N of aboveground litterfall, but also with fungal

biomass. These sorts of correlations can form an interesting basis for new hypotheses, but they cannot establish cause and effect (Source: from data of Son and Gower 1991; Scott 1998).

218

Chapter 9

1/3 lower under conifers. The omission of larch from the meta‐analysis may have strong influence on these apparent differences, as larch may show much higher N mineralization than other conifers in common garden experiments (see Figure 9.12). Generalizations that apply to classes such as conifers and broad‐leaved species need to be considered from a statistical perspective, just like other soil properties. A general difference in the mean value between two groups is only one piece of the information that would be needed for overall insight. The distributions around the means might be very tight, in which case knowing that the means differ on average would be powerful. Where the distributions are broad and overlapping, the value of any difference on average may have limited value in understanding specific cases. As noted in Table  9.1, differences between the conifers Norway spruce and white pine may be as large as differences between a conifer and broad‐leaved species.

MYCORRHIZAE MAY BE A KEY TO THE EFFECTS OF TREE SPECIES ON SOILS All trees depend on associations with mycorrhizal fungi for the uptake of water and nutrients. The extent of this dependence varies with soil fertility, and the types of mycorrhizal fungi differ among tree species (Chapter 6). Tree species supporting ectomycorrhizal fungi might have very different influences on soils than those supporting arbuscular mycorrhizae. Phillips et al. (2013) examined soils in a mixed‐ species forest in Indiana, USA, and plotted soil traits as a function of the proportion of plot basal area belonging to each class of tree species. The forest was about 80 years old, with some species forming associations with ectomycorrhiza (American beech, red oak, black oak, white oak, and pignut hickory), and others with arbuscular mycorrhizae (sugar maple, tulip poplar, sassafras). The plots were chosen in similar landscape positions, and had at least one species of each type. The distribution of species across the landscape was not controlled, and the success of

tree colonization and growth could relate in part to pre‐existing soil conditions that also influenced properties measured in the study. As the proportion of tree basal area of ectomycorrhizal trees increased from near 0 to near 100%, the extractable ­ammonium + nitrate as a percentage of total soil N (0–15 cm) declined by about half (Figure 9.13). The rate of net nitrification showed a much steeper response to increasing basal area of ectomycorrhizal species (rate of net nitrification can be negative if an initial soil sample has more nitrate than is present at the end of an incubation period). The differences between tree groups and biogeochemistry may be even greater than processes shown in these graphs. Brzostek et al. (2015) girdled trees in some plots to reduce or eliminate the carbohydrate supply to mycorrhizal fungi. Girdling of ectomycorrhizal trees led to a decline of 40% in the soil enzymes responsible for decomposing organic matter, indicating that the C supply to the soil played a major role in priming decomposition. In contrast, girdling of arbuscular mycorrhizal trees increased decomposition by 45%, perhaps indicating that intact arbuscular mycorrhizae compete strongly with the decomposer community for nutrients (with the nutrient supply to decomposers increasing after girdling reduced competition with the mycorrhizae).

TREE SPECIES INFLUENCE SOIL ANIMAL COMMUNITIES Soil animal communities are complex webs that respond very strongly to tree species and the organic matter the trees add to the soil. An example of the sorts of patterns that might be found comes from an afforestation experiment in Siberia, where several tree species were planted in a former agricultural field (pictured in Figure 9.3). After 10 years, the density of large invertebrates (worms, beetles, ants, fly larvae and others) ranged from about 60 to 130 m−2 (Figure 9.14, Bezkorovaynaya 2005) compared with 40 m−2 in the adjacent agricultural field. Twenty years later, the population density more than doubled for  some species, and dropped by half in another.

Tree Species Influences

r2 = 0.37, p < 0.01

r2 = 0.82, p < 0.0001

90

1.5 Net nitrification (μg N g–1 soil hour–1)

Inorganic N as % of total soil N

1.8

219

1.2 0.9 0.6 0.3 0

0

20

40

60

80

70 50 30 10 –10

100

Basal area of ECM species (%)

0

20

40

60

80

100

Basal area of ECM species (%)

Figure 9.13  Soil properties in a mixed species forest related to the percentage of plot basal area comprised of tree species

that were ecotomycorrhizal (Source: reworked from Phillips et al. (2013), by permission of Wiley).

sampled multiple times within a year before we can be confident that apparent differences between years are much larger than any potential bias from sampling during different seasons (or recent weather conditions). Do the apparent changes in large invertebrate populations in the Siberian experiment indicate true changes at a decade scale, or might the season/environment conditions at the time of sampling mistakenly confound seasonal and decadal trends? The Polish study, for example, found that

300 10 years

250

30 years

200

Thousands m–2

Large invertebrates (animals m–2)

A common‐garden experiment in Poland also found huge differences in the density of small invertebrates (enchytreids and collembola) among tree species. Some soil properties (such as clay content and %C) are very stable though a growing season, but biological properties such as animal populations differ substantially as a result of changes in temperature, moisture, and overall biological activity of the food web. Properties that are sensitive to environmental changes through seasons may need to be

150 100 50 0 Arolla pine

Larch

Scots pine

Spruce

Birch

Aspen

50 45 40 35 30 25 20 15 10 5 0

Enchytreids Collembola

Betula Carpinus Larix

Pinus Quercus Spruce

Tilia

Figure 9.14  Density of large invertebrates in the Siberian afforestation experiment (left) differed by more than twofold

across the species, and also over stand development (Source: data from Bezkorovaynaya 2005). The density of small enchytreids and collembola in a common garden experiment in Poland showed several‐fold difference among tree species (Source: from data of Rożen et al. 2010).

Chapter 9

populations of enchytreids and collembola fluctuated by several fold between samplings in spring and autumn. Another important source of variation is the depth of sampling. The Siberian study found that 80% of the large invertebrates were found in the 0–20 cm mineral soil, with the O horizon and the deeper mineral soil accounting for the remaining 20%. Soil animals move up and down in soil profiles in response to temperature and moisture, so this distribution may shift seasonally. Differences in animal populations can be measured in common gardens, but the biogeochemical significance can remain murky. For example, an experimental plantation in Hawaii examined the effects of eucalyptus and an N‐fixing leguminous tree (see later in this chapter). The N‐fixing species enriched the total N capital of the soil, and dramatically increased N availability while decreasing P availability (Garcia‐ Montiel and Binkley 1998; Kaye et al. 2000). At one sampling period, the density of small earthworms was about 90 m−2 under eucalyptus, compared with 470 m−2 under the N‐fixing trees (Zou 1993). As the worms feed on litter, we may reasonably expect that the N‐fixer’s litter was higher quality worm food. What might the effect of such a large population of worms be on the turnover of N and P? Would the apparent effects of the tree species be different if the exotic worm species had not been present? The difficulty of interpreting the importance of animal populations is evident in the Polish common

75

45 30 15

THE INFLUENCE OF TREE SPECIES MIGHT BE EXAMINED AS A FUNCTION OF LITTER CHEMISTRY RATHER THAN SPECIES IDENTITY A major step in developing generalizable insights about species effects on soils may come from studies that intentionally vary the chemistry of litter inputs by selecting and arranging species in pure and mixed plantings. Litterfall inputs (as well as root inputs) in the middle of a monoculture plot

90

r2 = 0.53, p < 0.01

6.0

60

0

6.5

r2 = 0.40, p = 0.02 O horizon pH

O horizon mass (g m–2)

90

garden study mentioned above. The biomass of the O horizon declined (across tree species) as earthworm densities increased (Figure 9.15, Reich et al. 2005). Given that earthworms are major mixing agents of soils, this pattern would be expected if trees differed as suitable providers of earthworm food. Curiously, the pH of the O horizon increased with increasing earthworm density. Does this reflect a change in acidity for material passing through earthworm guts, an effect of mixing mineral particles with organic material, or perhaps the original chemistry of tree litter that also happens to influence both decomposition and pH in the O horizon? To complete the circular story, the pH of the O horizon also correlated with O horizon mass. Unavoidable feedbacks in ecosystems lead to circles of cause, effect, and covariation.

O horizon mass (g m–2)

220

5.5 5.0 4.5 4.0

0

5 Worm biomass (g m–2)

10

3.5

0

5 Worm biomass (g m–2)

10

75

r2 = 0.40, p = 0.02

60 45 30 15 0

3.5

4.5

5.5

6.5

O horizon pH

Figure 9.15  Covarying patterns of earthworm (dry matter) mass, O horizon mass, and O horizon pH in the Polish common garden experiment illustrate why circular feedbacks make it difficult to assign simple cause and effect in the study of soil ecology (Source: from data of Reich et al. 2005).

Tree Species Influences would be influenced by a single tree species, but at the boundaries between plots the influence of neighboring species blends in. Mixed‐species plots have the same sort of variation in species influence, but at a smaller (single‐tree) scale. A combination of spatially explicit sampling with a litter quality characterization (perhaps including the balance among nutrient elements, or “stoichiometry”) may be promising (see Vivanco and Austin 2008; Hättenschwiler et  al. 2008; Marichal et  al. 2011). A  particular challenge to this approach might be the divergence of chemistry and decomposition between leaf litter and fine roots within the same species. Hobbie et al. (2010) examined these aspects in the Polish common garden experiment, and found that leaves and roots may decay at the same rate for some species, or leaves may decompose at twice (or half) the rate of roots of the same species.

CHANGES IN SOILS DO NOT LAST FOREVER A major remaining question about the effects of different species on soils is how long these effects will last after the original species is replaced by another. Will these changes have major legacies, or will the species effects evaporate because they’re ephemeral features that disappear quickly as soils adjust to new species? Would a species that doubles the rate of soil N supply leave a legacy of high N for the next generation of trees, or would the legacy be very short lived? Laboratory incubations of soils from the Wisconsin common garden trial showed a fivefold difference in net N mineralization among species after 30 days (Scott 1996). After 300 days, the differences among species had shrunk to less than twofold. The laboratory incubations suggest that the apparent effect of species may be most pronounced on the relative small, rapidly cycling pools of N, with less influence on the larger, more stable pools. Field tests are clearly needed to determine the robustness and longevity of tree species effects.

221

Perhaps some “field trial” insights are possible from descriptive experiments of species influences on soils. For example, Giardina et al. (2001) found that the C accumulated under aspen was more resistant to decomposition than C under adjacent conifer stands. Laganière et  al. (2017) reviewed two dozen studies from across North America and concluded that although aspen may accumulate somewhat smaller pools of soil C than conifers, the C accumulated under aspen was more securely stabilized and resistant to decomposition. This consistent pattern for aspen underscores that knowing the total size of a pool is only part of the important story in forest soils: dynamics (and resistance to change) may be particularly important for key aspects. One second‐rotation example is available of the longevity of tree‐species influences, from the plantations in Hawaii of Eucalyptus and N‐fixing Falcataria (Figure  9.16, see also Figures  9.20 and 9.22). Twenty‐year‐old plots of pure Eucalyptus and pure Falcataria were harvested, and the plots replanted with the same or the opposite species. The first‐­ generation species had no influence on the next generation of Falcataria. Plots with soils enriched in N by Falcataria in the first generation led to almost double the growth of Eucalyptus ­compared with soils that had Eucalyptus for both generations. The greater growth of Eucalyptus was  associated with twice as  much available N in former‐Falcataria soils as compared to former‐­Eucalyptus soils. This single case study does not provide a foundation for extrapolation, and the influences of N‐fixing trees might be stronger (with greater inertia) than other species. Future experiments of this sort should provide ­fascinating outcomes.

NITROGEN FIXING SPECIES REDUCE N FROM THE ATMOSPHERE AND ENRICH SOILS The N fixing enzyme, nitrogenase, reduces atmospheric N to ammonia through addition of electrons and hydrogen ions (see Chapter 8):

Chapter 9

222

Net Increment (Mg ha–1yr–1)

20 16 12 8 4 0 Euc. after Euc. after Falc. after Falc. after Euc. Falc. Falc. Euc.

Eucalyptus

Falcataria

Figure 9.16  The growth rates of second‐generation species were examined in Hawaii in relation to the first‐generation species. Eucalyptus planted after a previous generation of Eucalyptus grew at about half the rate (from age four to five years) of Eucalyptus planted after N‐fixing Falcataria. Growth of second‐generation Falcataria did not depend on the first‐ generation species (Source: from unpublished data of Senock, Powers, and Binkley).



N 2 8e

8H

2NH3

H 2

Because electrons are added to N2, this is a reduction process, requiring energy input. The production of hydrogen gas represents an unproductive energy drain, and some N‐fixing organisms have a hydrogenase system that oxidizes hydrogen to form water and release energy. The theoretical energy requirement is about 35 kJ mol−1 of N fixed, but it is difficult to assess the actual cost under field conditions. A good approximation of the actual cost might be about 15–30 g of carbohydrates (6–12 g C) for every gram of ammonia N produced (Schubert 1982; Marschner 1995). About half the carbohydrate costs is for the N‐fixation process, and the other half is for the growth and maintenance of nodules and for assimilating and exporting the fixed ammonium. The nitrogenase enzyme is also very sensitive to oxygen, and the high energy requirements coupled with the need for oxygen protection set limits on N fixing systems. Interestingly, a substantial amount (25–30%) of the CO2 released during the N‐fixation process can be “refixed” inside the root nodules by phosphoenolpyruvate (PEP) carboxylase.

A wide range of tree species have evolved an association with N‐fixing bacteria, including many legume family genera (such as Acacia, Robinia, Falcataria, Albizia), and “actinorhizal” species of Alnus, Eleagnus, Casuarina and others. The bacteria infect roots, causing formation of nodules where bacteria proliferate, consume carbohydrates (while protected by low oxygen partial pressures), and fix N. The fixed N is exported to the trees as amino acids, and used in all growing tissues. When tree tissues die, the N enters the decomposition process, and may be taken up by the roots and mycorrhizae of all plants in the soil. Rates of N fixation vary among species, sites, and over time, and the local details of these factors need to be considered for each case. Overall, the major N‐fixing species that have been intensively studied tend to fix about 75 kg N ha−1  year−1, though rates half this large, or twice this large, are not uncommon (Figure  9.17). To put these rates in context, 75 kg N ha−1  year−1 would equal 10–25 times the annual rate of input from atmospheric deposition, and about the same rate at which N circulates from soil into trees through the course of one year.

Tree Species Influences

223

35

Percent of reports

30

Red alder (n = 27) Tropical species (n = 32)

25 20 15 10 5 0 0

25

50

75

100

125

Nitrogen fixation (kg N

150

175

200 >200

ha–1 yr–1)

Figure 9.17  Reported rates of N fixation for forests dominated by red alder in the Pacific Northwest (Source: data from

Binkley et al. (1994)), and for a wide variety of tropical species (Source: data from Binkley and Giardina (1997); Garcia‐ Montiel and Binkley (1998); Pearson and Vitousek (2001); Forrester et al. (2007); and Bouillet et al. (2008) (including original data and summarized tropical forest cases)).

NITROGEN FIXATION ACCELERATES NITROGEN CYCLING What happens to the N once fixed? The ammonia is rapidly aminated to become part of organic amino acids and proteins. In free‐living systems, the bacteria and cyanobacteria simply use the N as any other microbe would – to grow and synthesize new cells. In symbiotic systems, the fixed N is shipped from the nodules into the plant roots where it performs the usual functions. Soil enrichment from N fixation comes only after microbe and plant tissues have died and decomposed to release the extra N. Increases in N availability may have a positive feedback that stimulates N mineralization. This is one of the major benefits of mixing N‐fixing species with crop trees, and can be illustrated in mixed stands of alders and conifers. Portions of a Douglas‐ fir plantation on Mt. Benson in British Columbia, Canada, contained shrubby Sitka alder, red alder, or no alder (Binkley 1983, 1984). The site index in the absence of alder was low, about 24 m at 50 years. Sitka alder fixed only about 20 kg N ha−1 year−1, but after 23  years the N content of litterfall was

110 kg N ha−1  year−1 with Sitka alder, compared to only 16 kg N ha−1 year−1 without alder. The N fixation rate of red alder was about twice that of Sitka alder, and the N content of litterfall was about 130 kg N ha−1  year−1. What accounted for this increase in N cycling? The N content of alder leaves (and alder litter) was very high, about 3% of the dry weight, as compared with less than 1% for Douglas‐fir needles. Most litter immobilizes N from the surrounding soil in the initial stages of decomposition, but N‐rich litter begins mineralizing N immediately. Ecosystems with N‐fixing species enjoy both greater inputs of N and produce litter rich in N; these two factors combine to greatly accelerate N cycling and availability.

NITROGEN FIXATION RATES ARE DIFFICULT TO MEASURE Nitrogen fixation rates can be estimated by a variety of methods, with four approaches used commonly in forest research: 15N methods, N accretion, chronosequences, and acetylene reduction assays.

224

Chapter 9

The addition of highly enriched 15N to soils may provide the best estimates of N fixation, but the cost of sufficient 15N to label soils in the field has limited the use of this method for trees. Some processes in N cycling alter the ratio between naturally occurring 15N and 14N, and if certain key assumptions are met, this “natural abundance” approach can provide inexpensive estimates of N fixation. Unfortunately, the key assumptions are probably not met in many (if any) forests. Nitrogen accretion is also a simple approach, involving measuring the total N content of an ecosystem at two points in time. The difference between samplings represents the addition of N from fixation, plus any precipitation inputs. This accretion method assumes outputs are negligible, and underestimates true N fixation rates on sites that experience high nitrate leaching or denitrification. The chronosequence approach is really a variation of the accretion approach but with more assumptions (see also Chapter  15). Several sites of different ages are used to represent trends to be expected within one site over time. If all site factors are constant except for stand age, then spatial patterns can be interpreted to represent temporal patterns. Finally, the acetylene reduction technique takes advantage of the nitrogenase (N‐ fixing) enzyme’s preference for reducing acetylene to ethylene (both of which are easily measured) rather than dinitrogen to ammonia (not measurable without 15N methods).

15 N LABELING SHOWS HIGH RATES OF N FIXATION FOR LEUCAENA AND CASUARINA

Parrotta et al. (1996) used the 15N addition method for estimating N fixation to estimate the rate of N fixation in planted pure and mixed replicated plots of Eucalyptus, Leucaena, and Casuarina. Small subplots were labeled with 15N, so that N derived from the soil would be substantially enriched in 15N relative to N derived from the atmosphere. The ratio of 15 N:14N in Eucalyptus trees represented the isotope ratio of N derived from the soil, and the lower

enrichment in the N‐fixing trees allowed separation of the N‐fixer’s reliance on the atmosphere and soil. This approach showed that Leucaena fixed about 90% of its N supply through age two, and about 60% of its N supply by age four; Casuarina derived about 50–60% of its N from the atmosphere for both periods. This decline in N derived from the atmosphere for Leucaena was only a relative decline; the quantity of N fixed remained steady at about 70–75 kg N ha−1  year−1, but the uptake of N from the soil increased as the tree demand for N increased. The same experiment showed that more of the added N was stabilized in microbially unavailable pools under the N‐fixers than under Eucalyptus (Kaye et  al. 2002). Year‐ long incubations showed that microbes managed to release almost 40% of the 15N (which had been added seven years before) from the Eucalyptus soil, but only 25% of the 15N from the soil from the N‐ fixing Albizia and Casuarina.

15 N NATURAL ABUNDANCE IS AFFECTED BY N FIXATION

In the atmosphere, 15N comprises about 0.3664% of the N2, and 14N comprises 99.6336%. The isotope ratio is soils is typically enriched in 15N because the minor difference in mass of the N atoms favors the reaction, mobility, and loss of lighter 14N. If a soil is sufficiently enriched in 15N relative to the atmosphere, it might be possible to use the natural abundance of N isotopes to estimate the input of N from fixation. Natural abundance levels of 15N are often gauged as delta 15N (or σ15N), which equals the parts per 1,000 deviation of a sample from the percentage 15 N in the atmosphere. A soil sample that contained 0.3710%15N would have a σ15N of 4.6 ((0.3710– 0.3664) *1000). As a first approach, the contribution of the atmosphere to the N economy of an N‐fixing tree could be estimated by comparing the σ15N of foliage with the σ15N of total soil N. Reality is actually messier: the σ15N within trees may differ by several per mil from the roots to the leaves, and the

Tree Species Influences σ15N of soil total N often increases by several per mil with depth. In addition, the σ15N of soil ammonium and nitrate commonly do not match that of the total soil N, or each other. One approach to circumventing these problems is to use a non‐N‐fixing reference tree to gauge the σ15N of available soil N. For example, total soil N in a pure Douglas‐fir plantation had a σ15N of 5.8, significantly higher than the 2.3 σ15N for a mixed stand of red alder and Douglas‐ fir (Binkley et al. 1985). Accounting for the higher %N in the mixed stand soil, these data suggest that the alder had contributed about 60% of the N present in the 0–15 cm soil depth. The confidence in this pattern had to be tempered, however, with data from two other sites where the σ15N of total soil N was too close to the atmospheric level to be useful, and by a fourth site where the σ15N in the mixed alder/Douglas‐fir stand was more enriched in 15 N. Högberg (1997) suggested that differences in σ15N might need to be greater than five per mil before interpretations about N fixation could be made with confidence. A meta‐analysis of N fixation studies that used 15 N natural abundance indicated that leguminous trees show fairly high variation in the proportion of N obtained from the air (via fixation) and that obtained from the soil (Andrews et  al. 2011). Actinorhizal species were notably more consistent in obtaining the major portion of their N from the air. Curiously, legumes were more likely to shift to reliance on soil N on more fertile sites, whereas actinorhizal trees continued to rely heavily on N fixation regardless of soil fertility. Bouillet et  al. (2008) investigated rates of N fixation by Acacia mangium in pure and mixed‐ species plots, using several approaches. Based on dilution of labeled 15N added to the soil early in the plantation, they estimated the acacia fixed about 30 kg N ha−1  year−1 when mixed 1:2 with eucalyptus trees. The natural abundance patterns suggested that N fixation was double this rate, indicating again that natural abundance approaches are too imprecise for reliable quantitative estimates.

225

NITROGEN ACCRETION AND CHRONOSEQUENCES CAN DETECT MODERATE AND LARGE N FIXATION RATES These two approaches can be illustrated by several research projects on an N‐fixing shrub, snowbrush (Ceanothus velutinus). Youngberg and Wollum (1976) and Youngberg et  al. (1979) intensively characterized the N pools of a recently logged site, and resampled for N accretion several times over the next 20 years. Zavitkovski and Newton (1968) chose a chronosequence approach in hopes that differences among sites of varying ages would track the N accretion occurring within each site, providing a rapid estimate of N fixation. Youngberg and Wollum (1976) and Youngberg et  al. (1979) found that N accumulated in vegetation + soil (to 23 cm) at a rate of about 110 kg N ha−1 year−1 for the first 10 years, and 40 kg N ha−1  year−1 for the next 5  years. Zavitkovski and Newton (1968) found a very similar pattern (in N content of vegetation + soil to 60‐cm depth), despite high variability among sites (Figure 9.18). A regression equation of ecosystem N content with age shows a significant slope in the chronosequence series of sites of about 80 N ha−1  year−1. A 95% confidence interval around this relationship ranges from 20 to 140 kg N ha−1  year−1. However, Zavitkovski and Newton (1968) found few nodules on any plants younger than five years, and concluded that the three youngest stands were probably unrepresentative of the rest of the sequence. Deleting these stands, they concluded that N fixation was negligible in the chronosequence. The chronosequence method can be converted into the accretion method if sites are resampled at a later date. Hopefully, each ecosystem would progress along the trajectory established for the chronosequence. Using this approach, 6 of the original 13 sites examined by Zavitkovski and Newton (1968) were resampled 15 years later. Five sites showed no significant change in soil (0‐ to 15‐cm depth) N content, but one showed a very significant increase of 280 kg N ha−1 (17 kg N ha−1  year−1; Figure  9.18).

226

Chapter 9 5 Biomass + 0–60 cm soil

Nitrogen (kg ha–1)

4 3 2

Biomass + 0–23 cm soil

1 0

0

5

10

15

Plantation age (years) 1,400

Nitrogen (kg ha–1)

1,200 1,000 800 600 400 200 0 1962

1972 Year

1982

Figure 9.18  N‐fixation estimates for snowbrush based on a chronosequence (blue circles, based on Zavitkovski and Newton 1968), and at a nearby site where accretion was followed within a single site over time (red squares, based on Youngberg et al. 1979). The rate of change in the chronosequence seemed to match that of the within‐site resampling, but a later within‐site resampling of the chronosequence sites (graph on right) found only one of six sites showed an increase in N. The open‐grown snowbrush is a very large specimen from British Columbia, Canada.

Some N accretion may have occurred in other ecosystem pools, but it is unlikely that high rates of N fixation took place without evidence of increased N in the upper soil. This resampling showed that the apparent trend in the original chronosequence was not in fact a reliable estimate of N fixation for other sites, and that the sites in the chronosequence probably differed fundamentally from the single site sampled by Youngberg. As with most features of forest soils, site specific details were important in addressing the rates of N fixation by snowbrush.

ACETYLENE REDUCTION ASSAYS ESTIMATE CURRENT RATES OF N FIXATION The fourth approach to estimating rates of N fixation also involves large sources of variation. If the average rate of N fixation per gram of nodule (activity rate) were known, an estimate of the annual N fixation rate could be obtained by multiplying the rate times the nodule biomass per hectare. Nodule activity usually is assessed with an acetylene reduction assay,

Tree Species Influences

Nodule activity also varies on a diurnal cycle; rates usually decline at night (Figure 9.19). Seasonal cycles are also important, as are patterns in water availability. It is very difficult to account for all these sources variability; most studies simply try to perform assays on many nodules at many points in the growing season in hopes of hitting a reasonable approximation of average rates. In most cases, it would be fortunate to achieve a confidence interval of ±50% on a seasonal estimate of average nodule activity. Nodule biomass per plant is easy to determine on small plants by excavating entire root systems, but large plants present mammoth problems. Ecosystems with large plants usually are sampled for nodule biomass by digging pits or taking soil cores at random across the site. The soil then is sieved to collect all nodules. The variability is again quite high, and a standard error of the mean is likely to be 25–50% of the mean. Because high variability is unavoidable in the acetylene reduction method, these estimates should be viewed as very rough approximations. Acetylene reduction assays can be useful in gauging the general magnitude of N fixation (such as less than

where the nitrogenase enzyme reduces acetylene to ethylene rather than reducing N gas to ammonia: C2H 2



2H

2e

C2H 4 .

Acetylene and ethylene are easier to measure than the N compounds. Recalling that N fixation requires the addition of six e− for every mole of N2 consumed, it would appear that the reduction of 3 mol of acetylene would be equivalent to 1 mol of N. However, H2 gas is an unavoidable drain on the N‐fixing system when N2 is reduced, but no generation occurs when acetylene is reduced. Another potential artifact in the acetylene reduction method is called nitrogenase derepression. Normally, the product of N fixation (ammonia) has a negative feedback on further synthesis of the nitrogenase enzyme. This feedback balances the rate of N fixation with the plant’s ability to process ammonia into proteins. With acetylene reduction, no ammonia is produced, no feedback develops, and the synthesis of extra nitrogenase is uninhibited. For these reasons, the conversion factor for acetylene to N is not always constant, but usually averages between 3 and 10 mol C2H2 per mole of N2 under carefully controlled incubation conditions. 20

16 Inland site

14 Acetylene reduction (μmol g nodule–1 hr–1)

227

15

12 Coastal site

10 10

8 6

5

4 2

0

4

6

8

10

12

14

Time of day (hour)

16

18

20

0 Apr

May

June

July

Aug

Sept

Oct

Nov

Month

Figure 9.19  The rate of N fixation by nodules of Himalayan alder begins to rise after canopy photosynthesis begins in the

mornings, and tapers off as evening approaches (Source: left, from data of Sharma 1988). Nitrogen fixation by nodules on red alder trees occurred for a longer period at a coastal site in Oregon with a mild, wet climate compared to red alder trees at a higher elevation, inland site in Washington (Source: right, from data of Binkley et al. 1992a). The inset picture shows N‐fixing nodules on the upper part of a root system of a four‐year‐old red alder tree.

228

Chapter 9

20 kg ha−1, or more than 100  kg  ha−1). They are probably most valuable for assessing physiological aspects of N fixation, such as response to light and water regimes. How well do rates of N fixation from acetylene reduction approaches compare with N accretion estimates? An adjacent 55‐year‐old pair of stands of Douglas‐fir and Douglas‐fir mixed with red alder were compared by the accretion method, and the current rates of N‐fixation were estimated by the acetylene reduction method (Wind River site from Binkley et al. 1992a). At the less fertile site, the biomass and soil (to 0.9 m) in the mixed stand contained about 2,800 kg ha−1 more N than the pure Douglas‐fir stand, for an average rate of accretion of about 54 kg N ha−1  year−1. Leaching losses were about 20 kg N ha−1 year−1 higher in from the mixed stand, raising the estimated rate of N fixation to 74 kg N ha−1  year−1. The seasonal average rate of acetylene reduction was 8.3 μmol g−1  hour−1, and the nodule biomass was 325 kg ha−1. The rate of acetylene reduction at a hectare scale would be: 8.3 μmol g−1  hour−1  *  325,000 g nodules ha−1 * 2,880 hour season−1, divided by 3 mols C2H2 per mol N2, times 28 g N mol−1, which totals to 73 kg N ha−1 for the season. This almost perfect match between methods is coincidental in part. A similar comparison at another site yielded an accretion‐based estimate of 102 kg N ha−1  year−1, and a current acetylene‐reduction based rate of 85 kg N ha−1 year−1 (Binkley et al. 1992a).

DOES N FIXATION DECLINE AS SOIL N INCREASES? Although N fixation by agricultural legumes does tend to decrease as availability of soil N increases, no feedback inhibition has been shown for actinorhizal species under field conditions. Several greenhouse studies showed that high concentrations of ammonium or nitrate can decrease N fixation, but the concentrations used in these studies

far exceeded soil solution concentrations under field conditions. Abnormally high concentrations of ammonium or nitrate around the roots may have misleading effects on N fixation. For example, Ingestad (1981) found that a high supply rate of N to alder roots at a low concentration actually stimulated N fixation rates. Under field conditions, alders have been found to fix substantial quantities of N on forest soils with some of the highest N contents in the world (Franklin et al. 1968; Sharma et al. 1985; Binkley et al. 1994). The high rate of nitrate leaching from many alder forests (sometimes in excess of 50 kg N ha−1  year−1) provides a final piece of evidence for lack of strong feedback regulation by N availability on N fixation. Nitrogen accretion under N‐fixing Falcataria trees in Hawaii was greater on soils that had less N, but even the richest site experienced substantial increases (Garcia‐Montiel and Binkley 1998). In some cases, N fixation does decline as trees age, such as an example with Acacia koa in Hawaii (Pearson and Vitousek 2001). However, the koa decline in N fixation was not driven by increasing soil N supply, as soil N supply actually declined along with N fixation. Why would plants expend energy to fix N on sites where soil N was abundantly available? No conclusive evidence is available, but some speculations can be advanced. Nitrogen fixation requires more energy than assimilation of ammonium, so ammonium uptake might appear to be more efficient than N fixation. Nitrate assimilation costs may be similar to those of N fixation. The picture is actually more complex, as metabolic costs are not the only energy costs involved. From a whole‐plant perspective the cost of developing and maintaining root systems to obtain ammonium and nitrate must also be included. Ammonium is not very mobile in soils, so a larger root system may be needed to obtain soil ammonium than for nitrate. Nitrogen gas is far more mobile than either ammonium or nitrate. Indeed, nodulated (N‐fixing) alders have been shown to develop smaller root systems than

Tree Species Influences non‐nodulated (non‐N‐fixing) plants do. From a whole‐plant energy perspective, N fixation may not cost more than assimilation of soil N, but this speculation needs further testing.

N‐FIXING PLANTS ALSO AFFECT CYCLES OF OTHER NUTRIENTS A mixture of N‐fixing plants and crop trees may show changes in the availability of nutrients other than N. Increased N availability may directly affect the cycling of other nutrients, and increased ecosystem production may rapidly tie up available nutrients in tree biomass. Returning to the Mt. Benson study, the P concentration of Douglas‐fir foliage was 0.22% without alder and only 0.12% with Sitka alder. Sitka alder leaves contained a healthy level of 0.30% P. Red alder trees in the same plantation further reduced Douglas‐fir foliage P concentrations to 0.09%, and red alder leaves contained only 0.14% P. Why did the alder impair Douglas‐fir P nutrition? The rate of biomass accumulation increased by 75% with Sitka alder, and by 260% with red alder. The P contained in the ecosystem’s biomass totaled 45 kg ha−1 without alder, 105 with Sitka alder, and 150 kg ha−1 with red alder. Thus, the decrease in soil P availability may simply be due to accumulation of P in rapidly growing biomass. As the stand without alder develops, the accumulation of P in biomass may also reach a point where P availability in the soil declines. Acceleration of stand production increases demands for all nutrients, and the benefits of N‐fixing species may be limited by the availability of other nutrients on very poor soils. This pattern would also be expected if N fertilizer were used to greatly accelerate stand growth. Limitations of nutrients other than N will affect N fixation as well as the performance of crop species. Other studies with red alder have found both substantial declines in P availability (Compton and Cole 1998, assuming the large site difference in

229

total P did not confound species effect on available P) and substantial increases (Giardina et al. 1995). A strong pattern of reduced P supply was also apparent in 16‐year‐old plantations of Eucalyptus and N‐fixing Falcataria. Nitrogen availability in pure Falcataria plots was four times higher than in pure Eucalyptus plots, but pure Eucalyptus plots had five times more available P than pure Falcataria plots (Figure  9.20). Phosphorus availability under N‐fixing species appears to be increased in some cases and depressed in others; this is another situation where simple generalizations are not helpful.

NITROGEN‐FIXING SPECIES ALTER SOIL CARBON IN TWO WAYS Soil organic matter increases under N‐fixing species, and rates of turnover may change as well. Across a range of N‐fixing case studies, the accumulation of soil N was accompanied by large increases in soil organic matter (soil C). Typical rates of C accretion range from 0.5 to 1.25 Mg ha−1 year−1 (Figure 9.21). The size of a pool in a soil could increase because of increased inputs to the pool, or reduced losses from the pool (or both). N‐fixing species often increase the total productivity of a forest, so it may seem reasonable that C accretion results from larger inputs of organic matter in aboveground litter and belowground tissues. Nitrogen‐rich tissues may decay faster than low‐N materials, and faster decomposition might be expected to lead to less accumulation of soil organic matter. However, as noted in Chapter 4, litter decay rates should not be assumed to relate well to rates of organic matter accumulation, because these processes are not very directly related, and are subject to many other influences. The Eucalyptus/Falcataria study described above was planted on a site following several decades of sugarcane agriculture on a high‐organic matter Andisol soil. Sugarcane is a C4 plant, the different photosynthetic pathway of C4 plants leads to difference in the ratio of the stable isotopes 13C and 12C.

230

Chapter 9

N or P availability, % of eucalypt (P) or Falcataria (N) monoculture

100

Phosphorus

80 60 40 20

Nitrogen 0

0

20

40

60

80

100

Falcataria (% of trees planted) Figure 9.20  A study with 18‐year‐old monocultures and mixtures of Eucalyptus saligna (straight trees, with dark bark)

and Falcataria moluccana (mottled bark) showed exponential increases in soil N availability (with resin bags) with increasing% of N‐fixing Falcataria in the plot (r2 = 0.95). Conversely, the availability of P increased exponentially with increasing% age of Eucalyptus (r2 = 0.89; Source: from data of Kaye et al. 2000).

After 17 years of tree growth, this means the soil had three pools of C: very old C3‐derived material from millennia of forests prior to clearing for agriculture; moderately old C4‐derived material from sugarcane agriculture; and recent C3‐derived material from the current crop of trees. The stable‐isotope approach cannot separate the pool of very old C3 material

from the recently added C3 material, but it may be reasonable to expect very little change in an old pool that lasted through decades of agricultural plowing. Any difference in the C3‐derived soil organic matter between Eucalyptus and Falcataria should represent the accumulation of new material during the current rotation. Given that the input of C4 material

Tree Species Influences 35

35

30

N

30

25

Percent of studies

Percent of studies

231

20 15 10 5

C

25 20 15 10 5

0 0

2

4

6

8

N accumulation in soil (g

10 m–2

12

14

yr–1)

0 0.0

0.3

0.5

0.8

1.0

C accretion (Mg

1.3

ha–1

1.5

1.8

yr–1)

Figure 9.21  Stands with N‐fixing trees increase both soil N and soil C. Carbon accretion averages about 8–9 Mg ha−1 year−1,

or about 17 g C for each g of N accretion in the soil (Source: from data of Binkley 2005).

stopped with the last crop of sugarcane, any difference in the remaining C4‐derived material would result from different rates of loss of soil C between the species. As with other N‐fixation experiments, soil C accumulated more under N‐fixing Falcataria relative to Eucalyptus. The divergence between the species was a linear function of the percentage of Falcataria trees in a plot, with the monocultures diverging by about 1.2 Mg C ha−1 year−1. Intriguingly, the stable isotope evaluation revealed that about 2/3 of the difference resulted from the accumulation of extra tree‐derived C under the N‐fixing trees, and from a 1/3 lower rate of loss of older C derived from sugarcane (Figure 9.22). This pattern of greater addition of C and greater retention of C was clear for this single experiment, but would the same pattern apply to other cases with N‐fixing trees? Resh et al. (2002 performed a similar analysis for three other experiments, including another Andisol in Hawaii, and a Vertisol and an Inceptisol in Puerto Rico; the patterns all supported the generalization that higher N‐fixing trees raise soil C by adding more C, and reducing rates of loss of old C. What mechanism might account for the better stabilization of soil C under N‐fixing species slow the

loss rate of older soil C? The first possibility would be a mechanism where higher N supply helps stabilize humified C pools. This idea was tested on an Andisol in Hawaii (five km away from the Eucalyptus/ Falcataria experiment) by adding over 1 Mg N ha−1 as inorganic fertilizer in a Eucalyptus plantation over a six‐year period (Binkley et al. 2004c). High u­niformity in soil properties gave a very precise estimate of the effect of added inorganic N on the accumulation of new soil C and loss of old soil C: exactly zero (Figure  14.6). The soil changes engendered by the N‐fixing Falcataria did not directly relate to increasing soil N supply. Other possibilities include the enhanced populations (and activities) of earthworms, a lower ratio of fungi/bacterial b­iomass, and other biological factors rather than simple chemistry.

WHEN DOES FIXED NITROGEN SHOW UP IN OTHER VEGETATION? Trees growing near N‐fixing trees may benefit from increased soil N supplies, and secondary benefits from this enhanced N nutrition (for a review, see Richards et  al. 2010). Only a few studies have examined how long it takes for a N‐fixing plant to

232

Chapter 9

135

17

C3-C from trees

130 13 125 C4-C from sugarcane

120

115

0

20

40

60

80

9

Soil C from C4 sugarcane (Mg ha–1)

Soil C from C3 trees (Mg ha–1)

140

5 100

Falcataria (% of trees planted) Figure 9.22  The accretion of C (0–50 cm depth) under N‐fixing Falcataria resulted from a 2/3 contribution of addition

new C derived from trees, and 1/3 by the greater retention of older C that was added under the previous land use of sugarcane agriculture. (Source: from data of Kaye et al. (2000); see also Figure 9.20).

provide significant amounts of N for associated s­ pecies. The major way that fixed N becomes available to associated trees is through decomposition; the annual litter produced by the N‐fixer decomposes, and the released N is taken up by trees. Minor amounts of N may pass directly from the root system of the N fixer to associated trees through mycorrhizal connections (Ekblad and Huss‐Danell 1995), but the rates of exchange among roots are probably too low to be important. About one to two percent of the fixed N may leak from roots into the soil (Uselman et al. 1999). The Puerto Rico 15N labeling study described above found that the Eucalyptus trees began taking up N fixed by Casuarina or Leucaena sometime between 2 and 3.5 years of age (Parrotta et  al. 1996). A study using natural abundance levels of 15 N concluded that a Leucaena plantation derived about 70% of its N from the atmosphere and that by age six years, the understory plants were as reliant on N fixed recently by Leucaena as was the Leucaena itself (van Kessel et al. 1994). In temperate forests, three studies have examined mixtures of poplars and alders for ­

short‐rotation biomass production (DeBell and Radwan 1979; Hanson and Dawson 1982; Cote and Camire 1984). All found that poplars neighboring on alders grew better in the first two to four year than poplars neighboring on poplars. This increased growth could have resulted from soil N enrichment by alder (DeBell and Radwan 1979) or from the size differences between the species. Because the poplars were larger than the alders, poplars with alders probably experienced less competition than poplars with poplars (Cote and Camire 1984).

N‐FIXING PLANTS HAVE A FULL RANGE OF POSSIBLE INFLUENCES ON STAND GROWTH Trees that fix N appear to always increase soil N pools and supply rates, and all plants in the forest likely experience higher N availability. All plants would also experience competition with the N fixers for light, water, and nutrients other than N. Boyden et  al. (2005) looked at the growth of

Tree Species Influences

233

Eucalyptus affected by Falcataria neighbors

10

Change in Eucalyptus growth (%)

10

5

5

0

0

–5

–5

–10

–10

–15

–15 100

0.20 80

0.15

60

40 F igh alca 20 t bo rho aria od ind ex

0.10

ne

0.05 0

0.00

il

So

l

pp

u Ps

)

ag

g/b

m y(

Change in Falcataria growth (%)

Falcataria affected by Eucalyptus neighbors

100 80 60 40 20 0 –20 –40 0.0

So

il N

0.2

su

0.4

pp

ly

40

(m

0.6

g/b

)

80

s x ptu de aly od in c Eu orho hb i eg

20 0.8

ag

60

120 100 80 60 40 20 0 –20 –40 100

0

n

Figure 9.23  The growth of individual Eucalyptus trees (left) in an 18‐year‐old set of mixed‐species plantations showed that Eucalyptus were somewhat influenced by neighboring Falcataria trees, and the balance between competition and facilitation depended on the soil supply of P. Growth of individual Falcataria trees was much more sensitive to the size of neighboring Eucalyptus trees, and the balance between competition and facilitation depended on soil N supply (and the ability of Eucalyptus to use N to grow and compete) (Source: from data of Boyden et al. 2005).

234

Chapter 9 Brazil, 6 years

Congo, 7 years

Australia, 11 years

15

15

10

10

5

5

0

0

–5

–5

–10

–10

–10

–15

–15

–15

15 Production (Mg ha–1 yr–1)

Foliage

10 5

Wood

0 –5

–20

TBCA

Eucalyptus Mixed Acacia

–20

Eucalyptus Mixed Acacia

–20

Eucalyptus Mixed Acacia

Figure 9.24  Three experiments with mixed‐species and monocultures of Eucalyptus or Acacia showed a wide range of

responses of stem growth, and of belowground production (TBCA, see Figure 14.3) (Source: from data of Forrester et al. (2004, 2005, 2006) and Epron et al. 2013).

i­ndividual trees in plots that contained varying ratios of Eucalyptus and N‐fixing Falcataria (Figure  9.20). A neighborhood index was calculated as a function of the sizes and distances of neighbors, and this index was used to examine the influence of neighbors on individual “focal” trees. When the focal tree was an average size Eucalyptus, the influence of neighboring Falcataria trees depended not only on the neighborhood index, but also on the supply of P in the soil. The low supply of P at this site limited the growth of both species, and restricted the rate of N fixation by Falcataria (Binkley et al. 2003). The net effect of Falcataria on growth of Eucalyptus was p­ ositive (up to a 15% increase) if soil P was high (Figure 9.23), likely as a result of both reduced competition for P and greater N supply resulting from higher N fixation. The growth of average‐sized focal trees of Falcataria was much more sensitive to the influence of neighboring Eucalyptus trees, and the response differed dramatically with soil N supply. At higher soil N supplies, Falcataria suffer from greater competition with Eucalyptus (for both light and P), because high N leads to very vigorous growth of Eucalyptus. At lower N supplies, Falcataria benefited substantially from a neighborhood of Eucalyptus trees. What gen-

eral insights can be taken from this study? The details are quite specific to the research site and probably the age of the plantation. The strongest general insights might be framed as negative statements: competition between the N‐fixers and non‐N‐fixers would not be consistent across gradients in soil nutrient supplies, and would not be independent of the sizes and distances of trees in neighborhoods. These insights from experiments with N‐fixing trees probably yield stronger effects than would be apparent from mixed‐species stands that did not contain N‐fixers. However, the strength of the effect might just be that it takes longer for more subtle effects to develop, not that the long‐term effects would necessarily be small. For insights about the silviculture and ecology of mixed‐species stands, see Forrester and Pretzsch (2015) and Pretzsch et al. 2017. Insights about forest growth come from examining both the growth rates of individual trees (above), and growth of stands of trees. Stand‐level responses of mixing N‐fixers with other trees have shown a wide range of responses (Figure  9.24). A  mixture experiment of Eucalyptus and N‐fixing Acacia in Brazil showed greatest wood growth as

Tree Species Influences well overall growth for the Eucalyptus monoculture. A similar study in Congo found greatest wood growth in the mixture, largely as a result of reduced belowground production. A third study in Australia again found greater wood growth in the mixture,

235

which also had the highest belowground growth. Collectively these studies underscore the importance of soil and other local environmental factors in determining the outcome of mixed‐species situations (for a review, see Ammer 2018).

C ha pte r   10

Characterizing Soils Across Space and Time OVERVIEW We know that soils vary over the dimensions of space and time, and it can be challenging to characterize these dynamics in ways that help us understand and manage soils and forests. Because soils are not tightly constrained systems, each soil trait (such as oxygen content, organic matter content, or number of nematodes) can vary with substantial independence of many other soil traits. Useful characterizations of soils require carefully designed efforts to capture the key traits of interest at the appropriate scales of time and space. Field sampling is followed by laboratory analyses, and both steps entail choices and limitations for interpreting soils. Forest and soil management always proceed with only partial understanding of soil dynamics, so attention to the possible important of unknown aspects is important for avoiding unintended damage. We measure soils, or more particularly their properties, for a variety of reasons. We may wish to better understand the genesis of a particular soil, how it has developed and how it may develop in future. We often wish to see how soils relate to vegetation type, or how they relate to vegetative ­productivity. We may wish to determine whether or not the addition of one or more chemical elements will enhance productivity. We may wish to create a “benchmark” so that we can determine how soils change over time, or with a change in vegetative cover. Measuring soils is a difficult and challenging task, but accurate information about

soils is crucial to reliable soils research and sound ecosystem management. In this chapter we discuss how we measure soils, and how the knowledge gained from those measurements leads to a better understanding of how ecosystems function, and how we might better manage ecosystems either for preservation or production. Recent advances in geographic information systems, global positioning systems, our understanding of the temporal and spatial scales of soil variation, and the development of more sophisticated taxonomies for the classification of ecosystem attributes have all contributed to our ability to better manage and preserve forests. By cleverly using these advances, we can better understand why a particular forest exists as it does. We now can also begin to understand why a change in ecosystem attributes, whether natural or human induced, affects the forest the way it does. Robust experimental designs and appropriate and adequate sampling are essential if we are to gain reliable knowledge. We have the tools – all we need to do is to use them wisely.

VARIATION OVER TIME AND SPACE IS FUNDAMENTAL IN FOREST SOILS Soils on the landscape are complex biological, physical, and chemical systems that involve processes and interactions occurring within and between the ­atmosphere, hydrosphere, biosphere, and geosphere.

Ecology and Management of Forest Soils, Fifth Edition. Dan Binkley and Richard F. Fisher. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

237

Chapter 10

238

In short, some soil properties vary from centimeter to centimeter, others vary from meter to meter, and others (such as soil parent material) may be constant for kilometers. It is important to consider both the temporal and spatial scales over which the property of interest varies if we are to sample it effectively.

These systems are dynamic in space and time and involve a constant interaction with environmental drivers. Scale issues are at the heart of many environmental problems because different processes may be dominant at different spatial and/or temporal scales. Few processes in nature operate in a linear fashion, and most interact with multiple other processes to provide complex patterns in space and over time. Sometimes the interactions generate patterns of gradual change, but thresholds for rapid changes are not unusual. As an example, rapid and slow changes in concentrations of inorganic N are illustrated in Figure 10.1.

ce

(m )

2

7 8 9 10 11 12 13 14 15 Di st an

0 1 2 3 4 5 6

(m )

2

ce

3

an

4

st

6

5

10 5 0 1

Di

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

15

April 15 - April 13

5

3

2

(m )

4

10 5 0 1

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Di st an

15 10 5 0

ce

NO3–

Distance (m)

May 13 - June 10

6

2

(m )

3

ce

4

Di

Distance (m)

5

an

3 2 1 0

3 2 1 0 1

st

NO3– (μg N/cm2)

2

(m )

3

2 1 0 1

3

15

6

ce

4

Di

6

5

Distance (m)

(h) 3

4

April 1st - April 15

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

2

(μg N/cm2)

3

(m )

4

ce

5

2 1 0 1

Di st an

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

6

3 2 1 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

(μg N/cm2)

(m )

ce

an

(g) 3

May 13 - June 10

Distance (m)

(μg N/cm2)

(m )

ce

an

st

2

April 15 - May 13

Distance (m)

NH4+

3

6

15 10 5 0

st

6

5

4

5

2 1 0 1

Distance (m) NO3–

2 1 0 1

3

March 13 - April 1st

3 2 1 0

(f) 3

3 2 1 0

(d)

2

Di

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Distance (m)

NH4+

3

April 1st - April 15

3 2 1 0

(c)

4

an

NH4+

6

5

2 1 0 1

Di

Distance (m)

NO3–

3

3 2 1 0

(b)

(μg N/cm2)

March 13 - April 1st

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

(μg N/cm2)

NH4

st

+

(μg N/cm2)

A soil is an entity (sometimes called a pedon) that exists at a particular geographic location, and areas with similar soils are often defined as soil types. (e)

(a)

(μg N/cm2)

SOILS ARE GEOGRAPHIC ENTITIES

Figure 10.1  Spatial maps of NH4+ and NO3− availability over four time spans: (a and e) March 13–April 1; (b and f) April

1–15; (c and g) April 15–May 13; (d and h) May 13–June 10 (After Cain et al. 1999).

Characterizing Soils Across Space and Time They tend to change as parent material and topography change. In some regions, very small changes in topography can lead to a change in soil type. The sampling design needs to take into account all of the soil types in the area of interest. When we sample the soil of an area, we also must remember that we sample soil volumes. Each volume from which a sample is drawn is a population composed of many individuals that vary among themselves, both horizontally and vertically (Cline 1944). Soil properties often vary greatly with depth from the surface. Of course, soil organic matter is a prime example. The variation in soil properties with depth can have profound influences on vegetation. For example, available phosphorus might be low in the A horizon of a particular soil, but high in the B horizon. In such cases, plants that root only in the surface soil would experience P deficiency, but trees that root more deeply would be able to acquire sufficient P.

MOST OF WHAT MATTERS ABOUT SOILS IS HIDDEN FROM VIEW One major difficulty in measuring soils is that they are seldom visible. They lie beneath our feet, with organic horizons covering mineral horizons. To understand soils we must expose then, and that means digging or boring into them. The exposed vertical section of a soil is called its profile, and there are conventions for describing it. Soils develop in parent material from the time of its deposition under the influence of local biota and climate in a particular geographic (topographic) setting. Over time, a number of environmental forces act to create distinct layers or horizons that may be parallel to the soil surface. This occurs through the differential downward movement of materials, such as Ca, Fe, organic matter, or clay particles. The movement and accumulation of materials at depth affects soil texture, structure, and/or color. These are the three principal properties that are used for distinguishing soil horizons (Soil Survey Staff 1993; Buol et al. 2003; Schoeneberger et al. 2011). There

239

Table 10.1  Master, transitional and combination horizon nomenclature.

Horizon

Criteria

O A

Organic soil materials Mineral; Organic matter accumulation, loss of Fe, Al, clay AB or AE Dominantly A horizon characteristics but also contains some characteristics of B or E A/B, A/E, Dominantly A horizon characteristics with A/C discrete, intermingled bodies of B or E or C AC Dominantly A horizon characteristics but contains some characteristics of C horizon E Mineral; loss of Fe, Al, Organic matter, or clay EA or EB Dominantly E horizon characteristics but also contains some characteristics of A or B E/A, E/B Dominantly E horizon characteristics with discrete, intermingled bodies of A or B E and Bt Thin lamellae (Bt) within a dominantly E horizon B Subsurface accumulation of clay, Fe, Al, Si, humus, CaCO3, CaSO4; or loss of CaCO3; or accumulation of sesquioxides; or subsurface soil structure BA or BE Dominantly B horizon characteristics but also or BC contains some characteristics of A or E or C B/A or Dominantly B horizon characteristics with B/E or discrete, intermingled bodies of A or E or C B/C B and E Thin lamellae (E) within a dominantly B horizon C Little or no pedogenic alteration; unconsolidated earthy material; soft bedrock CB or CA Dominantly C horizon characteristics but also contains some characteristics of A or B C/B or Dominantly C horizon characteristics with C/A discrete, intermingled bodies of A or B L Limnic material R Bedrock W A layer of liquid water or permanently frozen water (Wf) within the soil

are many soil classification systems in use around the world, but the nomenclature for describing soils is similar among them. We present the nomenclature followed by the United States Department of Agriculture’s Natural Resources Conservation Service and the soil horizon designation nomenclature (Tables 10.1–10.3, from Soil Survey Staff 1993).

240

Chapter 10

Table 10.2  Horizon designation suffixes.

Table 10.4  An example of a profile description.

Horizon Criteria suffixes

Horizon Description

a b c co d di e f ff g h i j jj k m ma n

Highly decomposed organic matter Buried genetic horizon Concretions or nodules Coprogenous earth (used only with L) Densic layer (physically root restrictive) Diatomaceous earth (used only with L) Moderately decomposed organic matter Permanently frozen soil or ice (permafrost); continuous, subsurface ice; not seasonal ice Permanently frozen soil; no continuous ice; not seasonal ice Strong gley Illuvial organic matter accumulation Slightly decomposed organic matter Jarosite accumulation Evidence of cryoturbation Pedogenic carbonate accumulation Strong cementation Marl (used only with L) Pedogenic, exchangeable sodium accumulation

Oi Oe Oa A

Bh

Eg

Btg Table 10.3  Other horizon modifiers.

Numerical prefixes

Used to denote lithologic discontinuities. By convention 1 is understood and is not written; e.g., A, Bt, 2Bt, 2C, 3C

Numerical suffixes The prime

Used to denote subdivisions within a horizon; e.g., A1, A2, Bt1, Bt2, Bs1, Bs2 Used to indicate the second or subsequent occurrence of an identical horizon descriptor in a profile or pedon; e.g., A, E, Bt, E’, Btx, C. Double and triple primes are used to indicate subsequent identical horizon descriptors in a profile or pedon; e.g., A, E, Bt, E’, Btx, E”, C

To describe a soil profile we begin by locating the preliminary horizon boundaries based upon differences in texture, structure, and color, and assigning each horizon a master horizon designation. If an O horizon is present, we measure its depth and determine its degree of decomposition. Remember that there may be more than one type of O horizon present (see Table  10.4 for an example soil profile description).

Cg

10 to 7 cm; intact needles and leaves; abrupt smooth boundary. 7 to 4 cm; moderately decomposed needles and leaves; indistinct wavy boundary. 4 to 0 cm; highly decomposed organic matter; indistinct wavy boundary. 0 to 20 cm; very dark gray (10YR 3/1) fine sand; weak fine granular structure; clear irregular boundary. (10 to 25 cm thick) 20 to 36 cm; dark brown (7.5YR 3/2) fine sand; weak fine granular structure; friable; gradual irregular boundary. (1 to 30 cm thick) 36 to 94 cm; light gray (10YR 7/2) fine sand; many coarse distinct grayish brown (10YR 5/2) mottles; few medium prominent (7.5YR 3/2) Bh bodies; few fine prominent strong brown (7.5YR 5/8) streaks along root channels; single grained loose; clear wavy boundary. 94 to 145 cm; light brownish gray (10YR 6/2) fine sandy loam; many coarse prominent yellowish brown (10YR 5/8) and few fine prominent red (2.5YR 4/6) soft masses of iron accumulation; weak coarse subangular blocky structure; wavy boundary. (10 to 48 inches thick) 145 to 203 cm; reticulately mottled gray (10YR 6/1), strong brown (7.5YR 5/8), and red (2.5YR 4/6, 10R 4/6) fine sandy loam; massive.

For each preliminary mineral horizon we determine its texture (Figure 5.1). Texture classes can be accurately determined only by particle size analysis; however in the field they are determined by feel. The ability to do this accurately comes only with practice, and field determination should always be followed up with particle size analysis. It is important to also note the presence of coarse fragment, e.g. gravel, channers, flags, or stones (see Table 5.2), that might be present either on the surface or within a horizon. We also need to estimate the percentage of the soil volume that the coarse fragments occupy, and to apply the appropriate

Characterizing Soils Across Space and Time modifier to the texture designation (e.g. gravelly sandy loam). Next we break out a handful of structural aggregates and determine the horizons structure. Structure is described in terms of shape, size, and grade. Shape: Granular: The units are approximately spherical or polyhedral and are bounded by curved or irregular faces that are not casts of adjoin peds. Blocky: The units are block like or polyhedral and are bounded by flat or slightly rounded surfaces that are casts of adjoining peds. The structure is described as angular blocky if the faces intersect at sharp angles and as subangular blocky if the faces are a mixture of rounded and planer faces and the corners are mostly rounded. Platy: The units are flat and plate like. They are generally oriented horizontally. Prismatic: The units are bounded by flat to rounded vertical faces and are distinctly longer vertically. The faces are typically casts of adjoin units. Columnar: The units are similar to prisms and are bounded by flat or only slightly rounded vertical faces. The tops of columns, in contrast to those of prisms, are distinct and rounded. Size: Five classes are employed: very fine, fine, medium, coarse, and very coarse. The size limits of the classes differ according to the shape of the units (Table 10.5). Grade: Grade describes the distinctness of units. Criteria are the ease of separation into discrete units and the proportion of units that hold together when the soil is handled. Weak – The units are barely observable in place. When gently disturbed, the soil material parts into a mixture of whole and broken units and much material that exhibits no planes of weakness. Moderate – The units are well formed and evident in undisturbed soil. When disturbed, the soil material parts into a mixture of mostly whole units, some broken units, and material that is not in units.

241

Table 10.5  Size class limits for describing structural

shapes. Size class

Platya  Prismatic and mm columnar mm

Blocky  Granular  mm mm

Very fine Fine Medium Coarse Very coarse

10

50`

100

10

 Plates are described as thin instead of fine and thick instead of coarse.

a

Strong – The units are distinct in undisturbed soil. They separate cleanly when the soil is disturbed. Then we determine the color or colors present in each preliminary horizon using the Munsell Soil Color Charts (Munsell 1990; Soil Survey Staff 1993). Like texture by feel, the ability to accurately determine soil color is only gained through practice. It is important to note whether or not there is any mixing of different colors (mottling) in the horizon, since this is an indication of the soil’s drainage class, that is how often and for how long the soil is saturated with water. Soil colors are predominantly associated with the presence or absence of organic matter and iron. Weathering causes the release of mineral constituents to the environment. Soluble weathering‐products are removed from the soil profile, while more stable compounds will precipitate. Iron migrating through the soil coats soil particles with thin oxide‐coatings. Well‐ aerated soils in areas of moderate weathering are typically are brown to yellowish‐brown in color from these ferric (Fe3−) oxide coatings. In areas of intense weathering, red and yellow colors develop. Prolonged soil‐saturation results in anaerobic conditions that lead to the formation of mobile ferrous (Fe2−) iron. The migrating groundwater redistributes the iron throughout the soil profile. Subsequent drainage restores aerobic conditions, but some iron coatings on the minerals may have been entirely removed, leaving the grayish surface of the mineral grains

242

Chapter 10

exposed. During drainage, some areas around pores, cracks, and root channels become dry and aerated more quickly than the rest of the soil. Ferric iron precipitates in these places, forming reddish‐ brown spots. During periods of alternating wetting and drying cycles, such as seasonal high groundwater, ferrous iron does not transfer out of the soil profile entirely but moves over short distances only and precipitates during the drying phase. Such conditions are characterized by blotches of gray and reddish‐brown soil colors occurring at the same depth. The longer the saturation period, the more pronounced the reduction process, and the grayer the soil becomes. This pattern of spots or blotches of different color or a shade of color, interspersed with the dominant color is called soil mottling. A quick look at the sample profile description will reveal that it is a somewhat poorly drained soil. Now, if we are satisfied that the preliminary horizons delineations are correct, we can apply the master and subordinate horizon designations based upon our soil description. We then need to measure and record each horizon’s thickness, which may vary from point to point along the exposed face, and describe the boundary between horizons (e.g. is it abrupt or gradual and is it smooth, wavy, or irregular). If we are interested in the soil only at the location of the soil description, we are finished; however, if we are interested in the soil in a particular stand, we need to determine whether or not the soil we have described is the only one present in the stand. We can estimate this soil’s extent in the landscape by traversing the stand, taking soil cores, and comparing horizons in those cores to our profile description. In general, but not always, changes in the soil profile occur at changes in topography, and this serves a key as to where must check for potential profile changes.

INFERENCES ARE DEVELOPED FROM PROFILE DESCRIPTIONS Having a profile description allows us to infer a great deal about a site. Texture is an important variable in determining site quality for tree

growth. Fine particles provide exchange sites for nutrients that are in cationic form, and they produce a large proportion of small soil pore sizes that retain water in the soil, while coarse particles produce large pore sizes and increase soil aeration. Coarse fragments provide neither nutrient exchange sites nor do they aide in water retention, but they do produce many macro‐pores that increase infiltration rate and reduce erosion risk. Therefore we can infer something about soil nutrient and moisture status, trafficability, and erosion hazard from soil texture. Structure tends to modify texture properties. Sandy soils with well‐developed structure retain more water than those with little or no structure. However, clayey soils with well‐developed structure are better aerated and have higher infiltration rates than clayey soils with little or no structure. The clay skins that develop on the peds in soils with well‐developed structure also increase the number of cationic nutrient exchange sites. Soils with well‐ developed structure also have greater soil strength than those with weak or no structure. If dense soil horizons are present within the profile, they may impede root growth and reduce site quality for trees. Therefore we gain knowledge about soil nutrient and moisture status, and trafficability from soil structure. The color that develops in soil horizons is another guide to soil nutrient status and serves as a guide to soil drainage class, the amount of time that the soil is saturated with water. The dark colors of the A horizon are related to soil organic matter content and nutrient status, particularly nitrogen content, and drainage class is an important factor in determining a site’s suitability for a particular tree species, its trafficability and resistance to compaction, and site productivity. By coupling this knowledge gained from soil texture, structure, and color with knowledge of horizon depth and thickness we can gain a great deal of understanding about a site either in the field or by studying the profile description in the office. Descriptions of soil profiles may give the impression that soil horizons are distinct and uniform

Characterizing Soils Across Space and Time with depth, and from one location to another nearby location. This is true sometimes, but in many cases the thickness of horizons varies across space, and the boundaries between horizons may be wavy or even more irregular (Figure  10.2). Sampling of soils for chemical or physical analysis entails clear consideration of these sources of variation (as described later).

243

POWERFUL SAMPLING DESIGNS ARE FUNDAMENTAL TO USEFUL SOIL MEASUREMENTS Sampling soil involves the selection from the total population of a subset of individuals (or samples) upon which measurements will be made. These measurements are then used to estimate the

Figure 10.2  This Dystric Brunisol developed in glacial outwash deposits (similar to Dystrochrepts in the US taxonomy,

and Dystric Cambisol in the UNFAO taxonomy) near Chapleau, Ontario, Canada. The profile in the upper photo has relatively clear and regular boundaries between horizons. A few meters away, horizonation is much more varied, presenting a major challenge for sampling (Source: photos by Mark Primavera).

244

Chapter 10

­roperties of the total population. Sampling is p essential if we wish to characterize the population, because measurement of the entire population would be impossible (Carter and Gregorich 2008). Sampling design involves the selection of the most efficient and effective method for choosing the samples that will be used to characterize the total population. The proper definition of the population is critical. It is important to clearly define the population of interest, and how the study findings will be extrapolated to that population (Binkley 2008). If one wishes to study how a particular plantation of loblolly pine growing on a Cecil soil (a fine, kaolinitic, thermic Typic Kanhapludult) in Georgia responds to the addition of fertilizer, then that particular plantation is the population that must be sampled. If one wishes to study how loblolly pine plantations growing on Cecil soils in Georgia respond to the addition of fertilizer, then all loblolly pine plantations growing on Cecil soils in Georgia are the population that must be considered for sampling. The correct selection of the population to be sampled is crucial if inferences made from the study results are to be valid. The study of how a particular plantation of loblolly pine growing on a Cecil in Georgia responds to the addition of fertilizer might use four replicated blocks within the stand, with four levels of nitrogen fertilizer added to randomly assigned plots in each block. The attention of the study might focus on choosing which levels of fertilizer to add (100 or 200 kg ha−1?), and the measurements to use for characterizing the response to fertilization (diameter, height, volume, mortality?). In many cases, the post‐treatment response might be more precise if pre‐treatment differences among plots were considered (such as basal area, or growth rates within plots). All of these questions need to be preceded by a clear definition of the population the researchers plan to address with the experimental results. The design of this experiment would be strong for inferring the response of trees in this single stand to fertilization during the period in time spanned by the

study. In fact, this design would be suitable for determining the response of all loblolly pine plantations to fertilization, if the response of all plantations to fertilization was not affected by factors such as soils, microclimate, and management. The design would be very weak, however, in providing insights about growth responses of any other plantation because these other factors dramatically confound the response to fertilization. Even if the researchers were interested only in the fertilizer response of loblolly pine growing on Cecil soils in Georgia, this design would provide 0 degrees of freedom for understanding the variability among sites within this narrowly defined population (with one site, n − 1 = 0). A similarly poor design might aim to determine whether the response to fertilization would differ between well‐drained soils and poorly drained soils by placing a well‐replicated trial in a single stand on each soil type. Perhaps the response would be significantly higher in the stand on poorly drained soil, but this design would have no statistical power to test whether the response would differ between the population of well‐drained soils and the population of poorly drained soils.

PSEUDOREPLICATION IS NO REPLICATION AT ALL FOR THE POPULATION OF INFERENCE Pseudoreplication is a major problem with many field studies and often occurs when the target population has not been defined correctly (Hurlbert 1984). If the population for the loblolly pine study referenced above were defined as “this plantation,” then the block design would indeed have four true replicates, and a researcher interested in just that plantation might conduct such a study. Frequently, the mistake is made of extrapolating the results of such a study to an entirely new population. If the population of interest was “loblolly pine plantations in Georgia,” the design would have a single replicate (one plantation) with estimated responses based on four subsamples (actually, split subsamples) of the single replicate. A much better design for this latter

Characterizing Soils Across Space and Time objective would omit any replication of treatments within a single plantation, and apportion the experimental work across four independent plantations within the population of interest. The dispersed approach would have 3 degrees of freedom for interpreting the likely responsiveness of the whole population, despite having 0 degrees of freedom for interpreting the response within any single plantation. Replicate subsamples within a single plantation are only useful if they improve the representativeness of the response for that plantation, which might be important if, for example, basal area or soil type is variable within the plantation. Efforts invested in more true replicates are almost always more profitable than the same effort applied to obtaining more subsamples that contribute no degrees of freedom for extrapolating to the population (Stape et al. 2006). Once we are certain that we have defined the population of interest correctly, the question becomes how to sample that population most accurately and efficiently. After the population of inference is clearly established, an effective sampling design is needed. A key issue of all sampling designs is the number of samples that would be needed to provide an estimate of the mean that warrants confidence, or to determine if an important change over time or space could be detected with confidence. Most soils have such high variability across space (but not so often time) that the number of desirable samples is not small. For thorough consideration of this issue, see the Quantifying Uncertainty in Ecosystem Studies website: http://quantifyinguncertainty.org.

PEDONS ARE SAMPLED FOR INFERENCES ABOUT LANDSCAPES Sampling is often confined to a single pedon for the purposes of characterization; however, sampling of an area of land is more commonly required. Sampling of an area can either be haphazard or systematic. Sampling is haphazard when accessibility or convenience determines the location of samples

245

rather than a conscious effort to ensure that the samples truly represent the range of variability in the population. Haphazard sampling yields data but no real information about the population, and consequently is of little real value. Systematic sampling can either be based on judgment, termed purposeful sampling, or probability. Purposeful sampling involves using a sampling grid approach to spread sampling points over the area in question. Sometimes purposeful sampling is based on knowledge held by the sampler. The reliability of data gained by sampler‐driven purposeful sampling is only as good as the judgment of the sampler. Both of these sampling methods can produce reasonable estimates of population means, variances, and totals, but they provide only a weak measure of the adequacy of these estimates to truly characterize the area. Probability sampling selects sampling points randomly using one of a variety of specific sampling layouts. The probability of sample point selection can be calculated for any specific design, allowing an estimate to be made of the accuracy of parameter estimates, and a range of statistical analyses can be carried out based on the estimates of variability about the mean (Pennock 2004; Bélanger and Van Rees 2008; Pennock et al. 2008). In addition to the issues of choosing locations for sampling soils, decisions need to be made about how a pedon will be sampled vertically. Metal frames are commonly used for quantitative sampling of the O horizon (Figure  10.3). Corers can then be used to sample the mineral soil beneath the same location; using the same location ensures a consistent representation of the whole profile. Plastic sleeves can be used within corers to provide intact samples (see Figure  10.5). Soils with high content of rocks may not be sampleable by corers, and pits need to be dug to expose horizons for sampling with a trowel. Mineral soils may be sampled by defined depth intervals (such as 0–15 cm and 15–30 cm), or by identifiable horizons. Soil chemistry typically differs markedly across the boundary of two soil horizons, so it may seem logical to base the depth of sampling so the intervals coincide with horizons. However,

246

Chapter 10

Figure 10.3  O horizons are commonly sampled using a metal frame. A knife is used around the outside of the frame to severe roots and all material is then removed from within the frame. The boundary between the O and A is sometimes distinct, but other times gradual enough that the precision and repeatability of sampling is challenging.

the reduction of variation that is obtained by keeping horizons distinct might be countered by increasing variance associated with variable thicknesses of horizons across a site. Horizon sampling and extrapolation become even more problematic when the boundaries between horizons are very uneven, such as in the lower photo in Figure  10.2. Sampling by horizon may also be more difficult to repeat in later resampling campaigns. Wang et al. (2017) compared estimates of soil pools based on sampling by fixed depths and by horizon in larch plantations in China. They concluded that sampling by horizons gave higher estimates for soil C and N contents, with 20–40% higher values than those from fixed‐depth samples. Which values were correct, the higher values from horizon sampling or lower values from fixed‐depth sampling? Answering this question would require an independent, reliable measure of the “true” values, which was not available. The best quality assurance information that was available focused on the variance for the estimates of the means. Horizon‐based sampling required 10–40% more samples to obtain a given level of confidence in mean values than fixed‐depth sampling. Analysis of soil samples in a laboratory provides information on chemical concentrations, such as mg C kg−1 of dry soil. Extrapolating a concentration to a mass per unit area of a forest such as kg C m−2

requires an estimate of the mass of soil. The mass value is determined as bulk density for a soil horizon or a given depth. Bulk density is the mass of the fine fraction (usually 500°C), kaolinite can be decomposed (into alu­ minum oxides and silica). Ulrey et al. (1996) exam­ ined four burned forests in California, and found that heavy fuel loadings covered about 1–2% of the areas, and these locations burned intensely enough to decompose kaolinite and halloysite. Soils often have a reddish hue beneath areas where large amounts of fuels led to intense heat­ ing. At high temperatures (and limited oxygen), Fe‐OOH compounds can be altered to Fe2O3; sig­ nificant generation of Fe2O3 may cover 1–15% of a burned area (Goforth et  al. 2005). Ulrey et  al. (2017) noted that burning of heavy fuel loads reduced cation exchange capacity of underlying soils by more than 50%, as a result of organic mat­ ter consumption and alteration of soil minerals. Goethite was transformed to maghemite or hema­ tite, kaolin was destroyed, and loss of water led to permanent collapse of vermiculite and chlorite.

FIRES MAY – OR MAY NOT – DECREASE WATER INFILTRATION INTO SOILS Water infiltration rates are often diminished follow­ ing fire owing to the plugging of surface pores and to increased fire‐induced water repellency (Krammes and DeBano 1965). Soils can also show hydrophobic tendencies in response to simple drying, so fires are not the only factor in determining soil hydrophobil­ ity. Indeed, a severe fire in Spain removed hydro­ phobicity that had developed under normal (no‐fire) forest development (Cerdà and Doerr 2005). Doerr et al. (2006) reported a similar pattern for eucalyptus forests near Sydney, Australia; pre‐fire hydrophobic­ ity was substantially reduced by wildfire. Increased sediment loss as a consequence of the formation of hydrophobic soil layers has been docu­ mented in ponderosa pine forests (Campbell et  al. 1977; White and Wells 1981). Hydrophobic proper­ ties appear to develop when organic molecules vola­ tilize (evaporate) as a result of heating; as the vapor comes in contact with cooler soil surfaces at depth, they condense and form non‐wettable surfaces.

The nature of such hydrophobicity has been worked out in laboratory experiments (DeBano et al. 1998). For example, heating soils to less than 175°C has little effect on rates of water infiltration into soils. Temperatures between 175 and 200°C may substantially reduce infiltration rates, while temperatures over 290°C lead to combustion of the organic molecules that would create hydrophobic conditions at lower temperatures. The spatial pattern of hydrophobicity may be very important to any influence on post‐fire ero­ sion rates. Patches of hydrophobic soils surrounded by high‐infiltration soils may have little chance for increased erosion, as water does not travel far (or in large quantities) before infiltrating soil. Woods et al. (2007) studied the pattern of hydrophobicity at small (1 m2) and moderate (225 m2) scales after fires in Colorado and Montana, USA, and found high variability at both scales (Figures  12.12 and 12.13). The high erosion losses that followed the Hayman fire in Colorado (similar to Figure 12.11) probably related more to the loss of soil cover (can­ opy and O horizon) than to hydrophobicity (MacDonald and Larsen 2009). Fire characteristics that promote hydrophobicity include high (but not extreme) fire intensity, coarse texture, and low soil water content. Despite these insights from controlled experiments, the contribu­ tion of hydrophobicity to overall post‐fire erosion of forest soils remains largely unknown. The gen­ eration of hydrophobic patches in burned land­ scapes is relatively straightforward, and water infiltration rates can be measured in hydrophobic patches and compared with those for unburned or burned‐but‐not‐hydrophobic soils. Water infiltra­ tion rates may be lower in hydrophobic patches for a period of months to several years under field con­ ditions (Dyrness 1976; McNabb et  al. 1989). Saturated conductivity in hydrophobic areas may be reduced by 10–40%, but the effects last for only a few storm events (Robichaud 2000). How much erosion would result from a burned watershed if hydrophobic patches did not develop? The present state of knowledge has too much uncertainty to quantify the contribution of each

Fire Influences

307

Percent of area

100 90

Strong

80

Moderate

70

Slight

60

None

50 40 30 20 10 0 Control 1

m2

Burned plots

Control 225

Burned m2

plots

Figure 12.12  The proportion of the burned areas within the Hayman Fire, Colorado, USA (near Figure  12.11) that showed hydrophobicity depended on the scale of interest. Small plots showed about half of the sampling points had some degree of hydrophobicity (with none evident in unburned plots), compared with about 25% of larger plots (with some evidence of hydrophobicity even in unburned plots; after data from Woods et al. 2007).

hydrologic factor in generating post‐fire erosion. Larsen et  al. (2009) concluded that the extent of cover on the mineral soil was the major determi­ nant of erosion, with any contribution of hydro­ phobicity being difficult to verify. This underscores the importance of re‐establishing vegetation to diminish both erosion and nutrient loss (Knight et al. 1985).

ASH MAY CONTAIN LARGE QUANTITIES OF NUTRIENTS Fires leave behind large amounts of ash that can be subject to wind erosion (see Ewel et  al. 1981; Bodí et al. 2014); typical masses of post‐fire range from 2 to 15 Mg ha−1 (Raison et al. 1985). The concentrations of nutrients is unusually high in ash, ­giving nutrient contents of post‐fire ash layers of 20–100 kg N ha−1, 3–50 kg P ha−1, and 40–1,600 kg Ca ha−1. Most of the nutrients in ash are water soluble or readily released by microbial activity, and this should provide a large supply of available nutrients for recovery of vegetation.

DIRECT NUTRIENT RELEASE FROM SOIL HEATING MAY BE IMPORTANT The obvious layers of ash following fire have led many researchers to focus on the nutrient content of ash as the major pool of nutrients released by fire. However, the direct effects of heating may release comparable quantities of some nutrients that were formerly bound in soil organic matter. Heat may release ammonium and P from organic matter, even if the organic matter is not consumed in the fire.

DECOMPOSITION AND MICROBIAL ACTIVITY CHANGE AFTER FIRE The environmental conditions of the residual O horizon and soil may be very different after the fire. Charred and darkened organic materials may absorb radiation better than do unburned materials, result­ ing in warmer conditions. For example, Neal et al. (1965) reported soil temperatures (at 5‐cm depth) in

308

Chapter 12

1 m2 plot

225 m2 plot

7

14 12

6.75

Meters

Meters

10 6.5

8 6

6.25

4 2

6

2

2.25

2.5

2.75

3

0

0

Meters

2

4

6

8 Meters

10

12

14

Very strongly hydrophobic Strongly hydrophobic Moderately hydrophobic Slightly hydrophobic Hydrophilic Figure 12.13  The spatial distribution of hydrophobic soils after the Hayman fire was very patch at the scale of 1 m2 (left) and 225 m2 (right). Patchiness can reduce the influence of hydrophobicity on erosion by allowing water to infiltrate before moving far enough to gather larger volumes of runoff (Source: reprinted from Geomorphology volume 86, Woods et al. copyright 2007, with permission from Elsevier.

a burned portion of a Douglas‐fir clearcut averaged 6°C higher than in unburned portions. Fires may also increase soil moisture by decreasing water use by vegetation. A pulse of increased nutrient availability typi­ cally follows fires as a result of reduced competition among plants, the release of elements from organic matter, and altered activity of soil biota. Changes in microbial properties of soils after fire have been documented, but few generalizations seem sup­ portable. Decomposition is generally expected to increase after fire because of increased tempera­ tures at the soil surface and increased soil moisture. Such an increase was demonstrated by Van Cleve

and Dyrness (1985), who placed cellulose strips in mesh bags in the upper soil horizons of unburned and adjacent burned white spruce stands. After two months, most of the cellulose had decomposed in the burned plots, whereas the cellulose in the forest remained practically undecomposed. In a more detailed study, Bissett and Parkinson (1980) showed that higher rates of cellulose decomposi­ tion in burned sites occurred only in the field; labora­ tory incubations actually had slower decomposition rates for burned soils. They concluded that the microenvironmental effects of fire were more important than the chemical effects. Decomposition studies using leaf materials generally have not

Fire Influences shown increased rates of decomposition in burned areas. For example, Grigal and McColl (1977) found that aspen leaves decayed more slowly in burned than unburned plots in Minnesota, and Weber (1987) found no differences in the rate of decomposition of jack pine needles on burned and unburned plots in Ontario. Changes in the O horizon during fire are sub­ stantial, and changes continue after the fire has

309

passed (Figure  12.14). Covington and Sackett (1984) reported that eight months of decomposi­ tion after a fire in a ponderosa pine stand released 108 kg N ha−1. Schoch and Binkley (1986) reported a similar increase in N release of 60 kg N ha−1 dur­ ing six months after a fire in a stand of loblolly pine. Indeed, fires can lead to higher soil N sup­ plies for decades. DeLuca and Sala (2006) exam­ ined soil N availability in seven wilderness areas in

Arizona, USA, ponderosa pine 500

30

4 Mg/ha loss

25 20 15 10

108 kg/ha loss

400 O horlzon N (kg/ha)

O horlzon mass (Mg/ha)

35

5

300 200 100 0

0 Before

After

7 months later

Before

After

7 months later

North Carolina, USA, loblolly pine 300

7 Mg/ha loss

20 15 10 5

63 kg/ha loss

250 O horlzon N (kg/ha)

O horlzon mass (Mg/ha)

25

200 150 100 50 0

0 Before

After

6 months later

Before

After

6 months later

Figure 12.14  Prescribed fires consumed some of the O horizon in ponderosa pine (upper) and loblolly pine (lower) stands, but the loss of O horizon mass continued after the fire, releasing substantial amounts of N that might enhance tree growth (Source: data from Covington and Sackett (1984), Schoch and Binkley (1986)).

Chapter 12

310

the northwestern US, comparing stands that remained unburned for 70–130 years with others that burned two or more times. Stands that had experienced fires had about double the soil N sup­ ply (measured by a variety of methods, which increases confidence in the pattern) of those that had no fires. The impact of high intensity fires on soil micro­ bial activity has received little attention. Bissett and Parkinson (1980) found no difference in microbial biomass between burned and unburned plots in a spruce/fir forest in Alberta after six years, but the ratio of bacteria to fungi was higher in burned plots. Modern genetic and physiological‐profiling tech­ niques allow investigators to go beyond lumping all bacteria, or all fungi, into a single pool. For exam­ ple, Hart et al. (2005) found that the diversity of the bacterial community dropped one month after a severe wildfire in a ponderosa pine forest, yet the fungal community diversity increased dramatically (Figure 12.15). The response of microbial commu­ nities is both rapid and sensitive to the intensity of fire (Morris et al. 2016). Overall, fire seems to gen­ erally lower fungal diversity for a decade or two,

but to have less (if any) influence on the establish­ ment of mycorrhizal establishment on plant roots (Dove and Hart 2017).

SOIL ACIDITY DECLINES AND pH RISES AFTER FIRES Soil pH typically increases immediately after fire, then declines back to prefire levels over a period of months, years, or decades. Soil pH influences the availability of some nutrients, both through direct geochemical effects and indirect effects on microbial activity. For example, Montes and Christensen (1979) raised the pH of incubated soils from beneath several different vegetation types by 0.5–1.7 pH units, and found that nitrate production increased by several‐fold. Soil pH depends upon the equilibrium between the exchange complex and the soil solution (see Chapter 7). An exchange complex dominated by H+ and Al3+ maintains a low (strongly acidic) pH in the soil solution, whereas domination of the exchange complex by so‐called base cations (K+, Ca2+, and

4 3

4 Bacteria p