Terrestrial Vegetation of California, 3rd Edition 9780520933361

Table of contents :
Contents
Contributors
Preface
1. The History Of Vegetation Classification And Mapping In California
2. Climate, Paleoclimate, And Paleovegetation
3. California Soils And Examples Of Ultramafic Vegetation
4. Nonnative Plants Of California
5. Estuarine Wetlands
6. Beach And Dune
7. Northern Coastal Scrub And Coastal Prairie
8. Sage Scrub
9. The California Channel Islands
10. Forests Of Northwestern California
11. Closed-Cone Pine And Cypress Forests
12. Oak Woodlands And Forests
13. Chaparral
14. Valley Grassland
15. Vernal Pools
16. Riparian Vegetation Of The Great Valley
17. Montane And Subalpine Vegetation Of The Sierra Nevada And Cascade Ranges
18. Southern California Conifer Forests
19. Alpine Vegetation
20. Transmontane Coniferous Vegetation
21. Sagebrush Steppe
22. Mojave Desert Scrub Vegetation
23. Colorado Desert Vegetation
Epilogue
Species Index
General Index

Citation preview

TERRESTRIAL VEGETATION OF CALIFORNIA

Third Edition

This page intentionally left blank

Terrestrial Vegetation of California Third Edition

Edited by

MICHAEL G. BARBOUR rOOD KEELER-WOLF ALLAN A. SCHOENHERR

UNIVERSITY OF CALIFORNIA PRESS Berkeley Los Angeles London

University of California Press, one of the most distinguished university presses in the United States, enriches lives around the world by advancing scholarship in the humanities, social sciences, and natural sciences. Its activities are supported by the UC Press Foundation and by philanthropic contributions from individuals and institutions. For more information, visit www.ucpress.edu. University of California Press Berkeley and Los Angeles, California University of California Press, Ltd. London, England

Library of Congress Cataloging-in-Publication Data Terrestrial vegetation of California / edited by Michael G. Barbour, Todd Keeler-Wolf, Allan A. Schoenherr. - 3rd ed. p. cm. ISBN-13: 978-0-520-24955-4 (case: alk. paper) ISBN-10: 0-520-24955-0 (case: alk. paper) 1. Plant ecology-California. 2. Phytogeography-California. I. Barbour, Michael G. 11. Keeler-Wolf, Todd. Ill. Schoenherr, AllanA. QK149.T442007 581.9 794-dc22 2007002860

© 2007 by the Regents of the University of California

Manufactured in Singapore 10 09 08 07 10 9 8 7 6 5 4 3 2

1

The paper used in this publication meets the minimum requirements of ANSl/N1SO Z39.48-1992 (R 1997) (Permanence of Paper). 00 Cover photograph: Desert agave (Agave deserti) in flower at AnzaBorrego Desert State Park, April 1997. Photograph by Todd Keeler-Wolf.

This third edition is dedicated to the memory of Professor Emeritus Jack Major (1917-2001), who was a seminal force in the creation of the first two editions and served as their coeditor

jack and Mary Major, cross-country skiing in Yellowstone National Park, about 1990. Photograph courtesy of jack's brother, Ted Major. jack excelled at skiing, and he served as a Corporal in the famous Tenth Mountain Division during World War 11.

This page intentionally left blank

CONTENTS

California Soils 71 Soil Taxonomy 79 A Medium for Plant Growth 83 Geography of Soils in California 85 Ultramafic Vegetation 93

CONTRIBUTORS xi PREFACE xiii

The History of Vegetation Classification and Mapping in California 1 Todd Keeler-Wolf Introduction 1 Early Mapping Efforts 4 Large Regional Remote Sensing Efforts 14 Vegetation Classification in California 22 Integrated Vegetation Mapping and Classification 29 Developments and Future Directions 33 2

3

Climate, Paleoclimate, and Paleovegetation 43 Richard A. Minnich Introduction 43 Modern Climate 43 Tertiary Climate Change and Development of California Vegetation 51 Summary of TertiaryQuarternary Climate and Vegetation Change 63 Areas for Future Research 66 California Soils and Examples of Ultramafic Vegetation 71 Anthony T. Q'Geen, Randy A. Dahlgren, and Daniel Sanchez-Mata

4

5

Nonnative Plants of California 107 Carla C. Bossard and John M. Randall Introduction 107 History 108 Invasibility of Californian Habitats 112 Control of Nonnative Invasive Plants and Restoration of California's Native Habitats 116 Summary 119

Estuarine Wetlands 1 24 Brenda f. Grewell, John C. Callaway, and Wayne R. Ferren, Jr. Introduction 124 Abiotic and Biotic Forcing Factors 133 Distribution of Estuarine Wetland Systems 137 Management and Conservation Issues 142 Areas for Future Research 148

6

Beach and Dune 155 Andrea J. Pickart and Michael G. Barbour Introduction 155 Geologic Processes 156 Vegetation of the Beach 157 Vegetation of Nearshore Dunes 158 Vegetation of Backdunes 160 Naturalized Vegetation 162 Physiological Ecology 164 Conservation 168 Areas for Future Research 173

7

Northern Coastal Scrub and Coastal Prairie 180 Lawrence D. Ford and Grey F. Hayes Introduction 180 Northern Coastal Scrub 180 Coastal Prairie 194 Areas for Future Research 205

8

Sage Scrub 208 Philip W Rundel Introduction 208 Environmental Relationships 209 Community Composition and Structure 213 Ecophysiology 220 Disturbance Regimes 223 Restoration 225 Areas for Future Research 225

9

The California Channel Islands 229

Steve Junak Denise A. KnapPI l

f. Robert Haller Ralph PhiIbrick l

l

Allan Sehoenherr and Todd Keeler- Wolf Introduction 229 Overview of Flora and Vegetation 231 Scrub Vegetation Types 232 Island Chaparral 242 Valley and Foothill Grassland 243 Woodland and Forest 243 Wetlands 245 Conservation and Restoration 246 Areas for Future Research 248 l

10

Forests of Northwestern California 253 John O. Sawyer Introduction to the Region's Floristic Richness 253 Conifers with Limited Ranges 255 The Broad Forest Pattern 259

The Parent Material Factor 267 Terrain and Forests 269 Dynamics of Northwestern Forests 281 Summary: Classification, Conservation, Restoration, and Areas for Future

Spatial and Temporal Aspects of Oak Woodland Sustainability 324 Oak Woodland Ecosystem Processes 326 Conservation and Restoration Issues 329 Economic Values and Utilization of Oak Woodlands 332 Areas for Future Research 333 13 Chaparral 339

Jon E. Keeley and Frank W Davis Introduction 339 Biogeographical Patterns 339 Fire 348 Community and Ecosystem Processes 355 Evolutionary and Geological History 360 Areas for Future Research 362 14 Valley Grassland 367

James W Bartolome W James BarrYI Tom Griggs and Peter Hopkinson Introduction 367 Replacement of the Original Grassland 372 The Annual Grassland 373 Conservation and Restoration l

l

Issues 385

Areas for Future Research 388 15 Vernal Pools 394

Ayzik I. Solomesheh Michael G. Barbour and Robert F. Holland Introduction 394 Landforms, Geologic Formations, and Soils 395 Hydrology and Water Chemistry 396 Distribution 397 Evolution of the Vernal Pool Flora 398 Autecology 400 Vegetation Zonation 402 Vegetation Classification 405 Conservation and Management 417 Restoration and Mitigation 418 Areas for Future Research 420 l

Research 284

l

11

Closed-Cone Pine and Cypress Forests 296 Miehael G. Barbour Introduction 296 Cypress Forests 298 Pine forests 303 Areas for Future Research 309

12 Oak Woodlands and Forests 313

Barbara Allen-Diaz Richard Standiford and Randall D. Jaekson Introduction 313 California Oak Woodland Communities 314 l

l

viii

CONTENTS

16 Riparian Vegetation of the Great Valley 425 Mehrey G. Vaghti and

Steven E. Greeo Introduction 425 Ecosystem Processes and Landscape Characteristics 428 Historical and Contemporary Management 433 Vegetation Ecology 435 Conservation and Rehabilitation 447 Directions for Future Research 449 17 Montane and Subalpine Vegetation of the Sierra Nevada and Cascade Ranges 456

Jo Ann Fites-Kaufman PhiI Rundel Nathan Stephenson and Dave A. Weixelman The Sierra Nevada 456 The Cascade Range 487 Productivity and Carbon and Nutrient Cycles 490 Conservation and Restoration 491 Areas for Future Research 493 l

l

18 Southern California Conifer Forests 502

Richard A. Minnieh Introduction to the Distribution and Species Composition of the Region's Forest 502 Historical Descriptions 507 Water Relations and

Microclimate 509 Fire, Forest Dynamics, and Biogeography 511 Disturbance from Insects, Pathogens, and Air Pollutants 526 Catastrophic Dieback in Record Drought 528 Forest Decline or Punctuated Equilibrium 531 Areas for Future Research 532 19 Alpine Vegetation 539

John o. Sawyer and Todd Keeler- Wolf Introduction 539

l

Floristic Relationships of the High Sierra Nevada 540 Treeline 547 Vegetation Patterns 550 Areas for Future Research 569

21

Introduction 587 Characteristics of Vegetation 589 Environmental Relations 589 Succession 592 Native Animals 593 Plant Communities 594 Areas for Future Research 605

20 Transmontane Coniferous Vegetation 574 Robert F. Thorne, Allan A. Schoenherr, Charlie D. Clements,

and lames A. Young Introduction 574 Mountain Juniper Woodlands 578 Pinyon-Juniper Woodlands 579 Desert Montane and Subalpine Woodlands 582 Subalpine Woodland 583 Areas for Future Research 584

Sagebrush Steppe 587

lames A. Young, Charlie D. Clements, and Henricus C.lansen

22

Mojave Desert Scrub Vegetation 609

General Vegetation Types of the Mojave Desert 620 MOjave Desert Vegetation Age, Persistence, and Conservation 646 Areas for Future Research 647 23 Colorado Desert Vegetation 657

Allan A. Schoenherr and lack H. Burk Introduction 657 Vegetation 659 Growth Form and Physiological Response 670 Areas for Future Research 676

Todd Keeler- Wolf

EPILOGUE 683

Introduction 609 Analysis of Mojave Vegetation 613

SPECIES INDEX 685 GENERAL INDEX 701

CONTENTS

ix

This page intentionally left blank

CONTRIBUTORS

BARBARA ALLEN-DIAZ University of California, Berkeley MICHAEL G. BARBOUR University of California, Davis W. JAMES BARRY California Department of Parks and Recreation (retired), Sacramento

GREY F. HAYES Elkhorn Slough National Estuarine Research Reserve, Moss Landing, CA ROBERT F. HOLLAND Institute for Geobotanical Phenomenology Consulting, Auburn, CA PETER HOPKINSON University of California, Berkeley

JAMES W. BARTOLOME University of California, Berkeley RANDALL D. JACKSON University of Wisconsin, Madison CARLA C. BOSSARD St. Marys College of California, Moraga HENRICUS C. JANSEN California State University, Chico JACK H. BURK California State University (retired), Fullerton STEVE JUNAK Santa Barbara Botanic Garden, CA JOHN C. CALLAWAY University of San Francisco, CA CHARLIE D. CLEMENTS USDA Agricultural Research Service, Reno, NV RANDY A. DAHLGREN University of California, Davis

TODD KEELER-WOLF California Department of Fish and Game, Sacramento JON E. KEELEY USGS, Western Ecological Research Center, Three Rivers, CA and University of California, Los Angeles

FRANK W. DAVIS Donald Bren School of Environmental Science and Management University of California, Santa Barbara

RICHARD A. MINNICH University of California, Riverside

JULIE M. EVENS California Native Plant Society, Sacramento

ANTHONY T. O'GEEN University of California, Davis

WAYNE R. FERREN, JR., Maser Consulting P. A., Red Bank, NJ; University of California (retired), Santa Barbara

RALPH PHILBRICK Santa Barbara, CA AND REA J. PICKART US Fish and Wildlife Service, Areata, CA

JOANN FITES-KAUFMAN USDA Forest Service, Nevada City, CA

JOHN M. RANDALL The Nature Conservancy and University of California, Davis

LAWRENCE D. FORD University of California, Santa Cruz

PHILlP W. RUNDEL University of California, Los Angeles

STEVEN E. GRECO University of California, Davis

DANIEL SANCHEZ-MATA Universidad Complutense, Madrid, Spain

DENISE A. KNAPP Santa Catalina Island Conservancy, Avalon, CA

BRENDA J. GREWELL USDA Agricultural Research Service, Davis, CA

JOHN O. SAWYER Humboldt State University, Areata, CA

TOM GRIGGS River Partners, Chico, CA

ALLAN A. SCHOENHERR Fullerton College (retired), Fullerton, CA

J. ROBERT HALLER Santa Barbara Botanic Garden, CA

AYZIK

I.

SOLOMESHCH University of California, Davis

xi

RICHARD STANDIFORD

University of California, Berkeley

MEHREY G. VAGHTI

NATHAN STEPHENSON

US Geological Survey, Three Rivers, CA

DAVE A. WEIXELMAN

F. THORNE Rancho Santa Ana Botanic Garden (retired), Claremont, CA

ROBERT

xii

CONTRIBUTORS

JAMES A. YOUNG

NV

Placer Land Trust, Auburn, CA USDA Forest Service, Nevada City, CA

USDA Agricultural Research Service, Reno,

PREFACE

In 2001, Carol Witham of the California Native Plant Society came to one of us with a request from the Publications Committee: Would we consider revising Terrestrial Vegetation of California? Her proposal was at once reasonable and terrifying. The first edition was published by Wiley-Interscience in 1977. It had only been modestly updated (by placing a supplemental section of new research at the end of the book) for a second edition that was published by CNPS in 1988. Now it was 13 years later, and-given the pace of vegetation research in California since 1977-it was clearly time to provide a new generation of readers with a summary of what was known and unknown, identifying new research directions that might best take us to the next level of understanding. By 2002, the three of us had agreed to serve as editors, a provisional list of chapters and authors had been drafted, and a publishing agreement with the University of California Press had been reached. Although the idea of a third edition was spawned by CNPS, that nonprofit organization could not alone financially see to completion such a large project as this one promised to be. Our objectives in this edition remained the same as for the previous ones: to summarize the physiognomy and species composition of major vegetation types, their patterns of distribution, relationships with environmental factors (including disturbance regimes), changes since the time of Euro-American contact, management and conservation issues, and possible changes for the next century during a time of human population growth and global climate change. Although information at the vegetation type and landscape scales was a first priority, we gave the authors the freedom to include other topics, such as ecophysiology,

population biology, ecosystem attributes (productivity, nutrient cycling), response to fire, geology-soil-plant relations, or plant-animal interactions. This latitude was necessary because ecological research in different parts of California has proceeded in independent directions, depending on the interests of individuals who happened to study there and depending on unique aspects of the habitat that were obvious topics around which research has been organized. Thus, fire regimes and their management occupy major sections in chapters on closed-cone conifers, chaparral, and montane conifer forest; and the impact of exotic, invasive species is an important topic in chapters on grasslands, wetlands, and the cold desert; and plant-soil interactions structure chapters on vernal pools and the forests of northwestern California. All chapters, however, end with sections on conservation and restoration and on "areas for future research." Many vegetation and community types, and some regions, are not covered in this edition, simply because our objective was to summarize and evaluate rather than to be merely descriptive in an encyclopedic way. At the same time, other vegetation types, and regions, are discussed in more than one chapter because they are so widespread or complex that a single chapter's perspective would not do them justice. In such cases, one of the chapters has been assigned the "lead" and provides the most depth of information about that topic. Half a century ago, Professor A.W. Kuchler 1 mapped the potential vegetation of the United States. He identified approximately SS major vegetation types as occurring within the boundaries of the state of California. Harold 1 A.W. Kuchler. 1964. Potential natural vegetation of the conterminous United States. American Geographic SOciety, Special Publication 36.

xiii

TABLE P-l

Area of California Dominated by Vegetation Types Mapped by Kuchler (1964) and Chapters in Which They Are Discussed

Vegetation Type

Acres

Hectares

Chapter

Conifer forests and woodlands Cedar-hemlock-Douglas-firspruce forests (NW)

2,020,700

808,300

Mixed conifer forest

13,641,000

5,522,700

Coast redwood forest

2,320,250

928,100

10

Red fir and Shasta red fir

1,903,500

761,400

10, 17

Mixed subalpine woodland

2,151,000

860,400

10, 17, 18, 20

Closed-cone conifer stands

123,200

49,300

6, 10, 11, 18

Ponderosa/shrub forest

1,695,100

678,000

17,20,21

Great Basin pine forest

49,100

19,600

17,20, 21

Pinion-juniper woodland

2,463,500

985,400

Juniper steppe woodland (Modoc region)

909,700

363,900

Mixed evergreen forest

3,399,200

1,359,700

9, 10, 12, 17, 18

Oak (foothill) woodland

9,554,500

3,821,800

9, 10, 12, 16, 17, 18

8,500,600

3,400,200

9, 11, 12, 13

573,000

229,200

17

Southern coastal (sage) scrub

2,473,500

989,400

7,8,9

Mosaic of oak woodland and southern coastal scrub

641,200

256,500

7,8, 9, 12

Great Basin sagebrush

1,851,400

740,600

17,20

Saltbush-greasewood

3,104,700

1,241,900

21, 22, 23

16,356,000

6,542,395

22, 23

Creosote bush-bur sage

5,330,800

2,132,300

22, 23

Paloverde-cactus scrub (Sonoran/Colorado desert)

1,052,900

421,200

23

878,700

351,500

7,14

13,222,250

5,288,900

1,859,400

743,800

5,16

747,400

299,000

18,19

3,246,000

1,298,400

10 10, 17, 18,20

18, 20 17,20,21

Hardwood forests and woodlands

Scrub vegetation Chaparral Montane chaparral

Creosote bush

Herbaceous vegetation North coastal prairie Interior (Central Valley) grassland Tule marsh (fresh-water marsh) Alpine Sagebrush steppe

9, 12, 14

21

NOTE: Types not mapped by Kiichler, thus not listed above, include: beach and dune (Chapter 6), coastal salt marsh (Chapter 5), northern coastal scrub (Chapter 7), vernal pools (Chapter IS), and Channel Island types (Chapter 9).

San Nicolas

FIG U RE P-1

I.~

Major topographical features of California: rivers, lakes, bays, points, capes, and islands.

124" 120'

llB'

117"

116'

\)

32"

FIG U RE P-2 Major place names of California: counties, mountains, mountain ranges, valleys, and deserts. Numbers along the top are degrees west longitude, and numbers along the side are degrees north latitude.

Heady planimetered those types for the first edition of this book, so that their total areas could be easily compared. We present his figures here (Table I), as a convenient '~ay to rank the importance of each type. Because California's total area is approximately 100 million acres, the percentage of land area for any vegetation type can be obtained by multiplying acres x 10- 6 • As another help to the reader, we include here (Figures 1 and 2) topographic and political maps of the state, showing major geographic features, county boundaries, rivers, and place names. As in any book project, there were many stops and restarts, author changes, delays, miscommunications, and very few instances when deadlines were actually (or even nearly) met. However, the process eventually succeeded. Several chapters contain their own acknowledgment sections. From our perspective as editors, however, we take pleasure in mentioning a few particular individuals and groups. First, we express our debt of gratitude to Chuck Crumly, Publisher in the Science Group at the University of California Press, his assistant Danette Davis, Project Manager Scott Norton, and Joanne Bowser of Aptara, for their consistent encouragement during the lengthy preparation of the manuscript. It was invaluable and greatly appreciated. We thank CNPS for its permission to freely use materials from previous editions of this book, the Koshland Bioscience and Natural Resources Library at UC Berkeley for use of images from the Weislander VTM collection, Kristi Fein for producing some of the maps in this work, Julie Evans for her diverse contributions to many chapters, and

to two ~nonymous reviewers who provided very important 4 ha. There was more of a strict economic focus on these surveys

14

VEGETATION CLASSIFICATION IN CALIFORNIA

to identify timber stands and less information was mapped for nontimber vegetation compared to the VTM survey. However, over 3,000 sample plots were conducted for the survey in a standardized way (Colwell 1977). These plots were chosen to represent the full array of wildland soils and vegetation types encountered during the survey efforts. Standard environmental variables (slope, aspect, elevation, erosion, drainage, and surface rockiness) were collected, and a soil pit was dug with the soil profile described. For wooded plots ocular estimates were made of total woody plant cover, range of height of the understory and overstory, herbaceous species listed; and in commercial forest areas height and age measurements were taken from representative economic species to determine site indices. In nonwoody herbaceous stands, herb cover and species composition were obtained using step-point (Evans and Love 1957) transects and herb cover estimated in 30-cm square point frames for every 10th step point. Data from these 0.4-ha circular plots are archived at the U.S. Forest Service PSW Research Station in Albany, CA (Colwell).

Large Regional Remote Sensing Efforts During the 1970s while work continued through the statewide efforts of the S-V Survey, additional research began on making use of the first imaging satellites to

produce vegetation maps in the state. Because of the relatively low-cost and high frequency of repeat images provided by the orbiting satellites, this medium quickly replaced the aerial photo interpretation methods for large regional or statewide efforts in mapping of vegetation. In 1972, ERTS-I (Earth Resources Technology Satellite), later called LANDSAT-1, was launched to beconle the first of a series of satellites especially designed to obtain remotely sensed data for use in the inventory and monitoring of natural resources. Because California has a wide range of cli111atic and topographical conditions, it proved to be an ideal testing site for conducting integrated remote sensing research in water supply, timber, range, chaparral fuels, agriculture, and recreation (Colwell 1977). Many such studies were conducted by the Remote Sensing Research Program of the l)epartlnent of Forestry and Conservation, and by the Space Sciences Laboratory, both of the University of California, Berkeley, under National Aeronautic and Space Ad111inistration (NASA) grants, and cooperative agreements with the Forest Service and with the Bureaus of Land Management, Reclamation, and Outdoor Recreation. LANDSAT imagery could be enhanced from magnetic and color composite processes to show different features correlated with plant productivity and structure. In these early studies LANDSAT imagery served as the first stage of stratification in multilevel sampling designs. In conjunction with this process, aerial photography (both high-altitude color-infrared and low-altitude color) and syste111atically collected ground data were used to provide additional salnpling levels. The information collected was used in equations to estimate the condition and volume of a vegetative resource. These estimates were made for timber volUlne, range productivity (Carneggie et a1. 1974), or fuel hazard ratings (Nichols 1974, DeGloria et a1. 1975). The LANDSAT-I multi spectral scanner detected differences among trees, shrubs, grass, and bare ground. Sonle distinctions could also be made of shrub Inaturity classes, such as pioneer, immature, and Inature. In some homogeneous areas, types could be separated by species composition. Major forest types and some broad timber volume classes could be separated (Gialdini et a1. 1975). In its early years, imagery from spacecraft was used prilnarily as the first stage of stratification for various sampling designs requiring additional photography for ground truth and lnore detailed sampling. In the early phases of its use, satellite irnagery was considered useful in the separation of broad vegetation types, condition classes, and other contrasting land conditions.

Hardwood Rangeland Mapping

()ne of the first regional approaches cOlnbining satellite and aerial photography was the Pillsbury Hardwood Rangeland Map (Pillsbury et a1. 1991, Pacific Meridian Resources 1994) Hardwood range lands below 5,000-foot elevation were originally 111apped by Dr. Norm Pillsbury (CaI Poly SLO) under

contract by California Departlnent of Forestry and Fire Protection (CDF). Polygons were delineated on 1981 1:24,000 scale black-and-white air photos, transferred to 1:100,000 scale base maps, and digitized. The data were updated by Pacific Meridian Resources under contract froln CDF using 1990 LANDSAT TM imagery. This GRID format data represent the base classification data used to update delineated polygons (polygons are provided as an additional layer). Each pixel was coded based on species group, tree size, and canopy closure class.

CALVEG and the U.S. Forest Service Remote Sensing Lab

The Classification and Assessment with Landsat of Visible Ecological Groupings (CALVEG) system was originally developed by the U.S. Forest Service Pacific Southwest Region's Ecology Prograln in 1978 (Parker and Matyas 1979, USDA Forest Service 1981). CALVEG mapping of the entire state was done between 1979 and 1981 by U.S. Forest Service personnel by photo interpretation of 1:250,000 scale color infrared prints of LANDSAT Multispectral Scanner iInagery acquired between 1977 and 1979 (Parker and Matyas 1979). Image interpretation was guided by existing soil and vegetation maps, field checking, and by personal contact with vegetation experts throughout the state (Parker and Matyas 1981). The minimum mapping unit was 400-800 acres, but the spatial resolution of the resulting map was very coarse. Average polygon size of the map mosaic is 38,000 acres for the entire state. The resulting map provides statewide land cover/land use polygons mapped froIn 1977 LANDSAT imagery and then digitized from 1: 1,000,000 scale maps. The data contain vegetation attributes for series-level species groups only. With the development of several future generations of satellite based scanners vegetation (LANDSAT 1 through 7, and the French 10-m pixel SPOT satellite) Inapping of vegetation in California in the 1980s and 1990s becalne largely a satellite-based operation. Satellite-based mapping had the distinct advantage over aerial photography of having repeat inlages of the land surface captured regularly throughout the year for a relatively low cost given the size of the area covered. With an ever-increasing technological effort put into the refinement of computer image processing in the 1980s, it became feasible to develop Inapping protocols that relied less on the trained eyes of air-photo interpreters and 1110re on the interaction of cOlnputer image-recognition software and ecological models to develop refined vegetation mapping rules. The most influential and widely used vegetation rnapping program relying largely on LANDSAT iInagery and image processing is the ongoing mapping conducted through the Remote Sensing Laboratory (RSL) of the U.S. Forest Service in Sacramento. The original CALVEG product formed the basis for the development of the coordinated mapping prograrTI within the California Region of the U.S. Forest Service, whose specific mission was to maintain and update vegetation I11apS

VEGETATION CLASSIFICATION IN CALIFORNIA

15

o

o

Not Mapped

Cl

1999 2000 2001

1997 _1998

D

0 Cl 2002 FIG U RE 1.6 Location of coordinated update

mapping area of California mapped by V.S. Forest Service Remote Sensing Lab in cooperation with CDF Fire and Resource Assessment Program.

for the jurisdictional portion of the state shared by the U.S. Forest Service and the California Department of Forestry and Fire Protection (Schwind and Gordon 2001). Since 1988, the RSL has had an active and scheduled program for creating and maintaining existing vegetation data layers of National Forest lands. These layers have been developed by the classification and modeling of a variety of remotely sensed and ancillary data. The RSL has developed the use of computer-assisted image processing to develop LANDSAT into multipurpose vegetation maps for all of the major wooded parts of the state (Fig. 1.6). The main purposes of these maps are to assess wildlife habitats, late successional old growth, forest health, mortality, growth, and standing forest volumes for the needs dictated by the Natural Resource Planning Act, Forest Resource Management plans, Northwest Forest Monitoring Plan, Sierra Nevada Framework Monitoring Plan, and other bioregional assessments, and more localized watershed and county planning efforts. The RSL keeps a coordinated schedule of updating their products with the goal of having vegetation maps no older than 5 years. Updates are made where changes to vegetation and surface fuels occur from various causes. In these same changed areas, inventory plots using a nationally

16

VEGETATION CLASSIFICATION IN CALIFORNIA

established system of forest inventory plots (PIA 2002) are re-measured. To achieve a coordinated cycle, baseline vegetation maps and PIA grid inventory plots need to be completed to a common standard and common source dates within a province as much as possible, balancing workloads and budget constraints. The CALVEG mapping classification is being maintained and updated at RSL and currently has 220 distinct vegetation and land use types (H. Gordon personal communication December 2006). CALVEG alliances are similar in resolution (though not always in formal definition) to those in the uppermost floristic level (Le., alliance) of the National Vegetation Classification hierarchy (see below). Both are based on dominant and existing vegetation components in a given area, but mapping units stress woody economically important types and tend to aggregate or overlook vegetation that typically does not attain stands of the minimum mapping unit of 2.5 ha. Other themes in the database include forest stand characteristics such as tree size and density and relative percentages of conifers and hardwoods in mixed stands (Table 1.3). The methods for developing classified images for CALVEG mapping follow a prescribed sequence of steps including

TABLE 1.3

CALVEG Mapping Criteria and Attributes Used by Region 5 Forest Service Remote Sensing Lab Minimum Mapping Size

2.5 acres for contrasting vegetation conditions based on vegetation type, tree canopy closure, and overstory tree size No minimum mapping unit for lakes and conifer plantations

Life-Form Classes

Conifer: greater than 10% conifer cover as the dominant type Mix: greater than 10% tree cover and 2045'

>24"

6

Multi-Layered Tree

Size class 5 trees over a distinct layer of size class 4 or 3 trees, total tree canopy exceeds 600/0 closure.

Mayer and Laudenslayer (1988).

.I

Standards For Canopy Closure

1

NOTE:

6

TABLE 1.8

Comparison of Tree-dominated Habitats by Wildlife Habitat Relationship Type and Selected Other California Classification Systems

WHR type

Cheatham and Haller(1975) CNDDB(1986)

Sub-Alpine Conifer (SCN)

Subalpine Coniferous Forests(8.6)

Eyre (1980)

Sierran Mixed Sub- Mountain Hemalpine Coniferous lock (205) Forest (86200) Englemann Foxtail Pine Forest Spruce(86300) Subalpine Fir (206) Bristlecone Pine (86400) Southern California Subalpine Forest (86500)

Whitebark Pine (208) Bristlecone Pine (209)

Whitebark pine Forest (86600)

Western White Pine (215)

Limber Pine Forest (86700)

Limber Pine (219)

Kiichler(1977) Upper MontaneSubalpine Forests(17) Southern Montane Subalpine Forest (18)

Munz and Keck (1973)

Parker and Matyas(1981)

Paysen et al. (1980)

Proctor et al. (1980)

Subalpine Forest (17)

Bristlecone Pine

Bristlecone Pine Limber Pine

True Fir Zone: Subalpine Second coniferous forGrowth Forest est (14c) (1.1.2E) Bristlecone True Fir Zone: Pine Woodland (18a) Old Growth Forest(I.I.2F)

no corresponding vegetation type

no corresponding vegetation type

Bristle-cone Pine Forest (18)

Englemann Spruce-Alpine Fir Foxtail Pine Limber Pine Lodgepole Pine

Great Basin Subalpine Forest (19)

Thorne(1976)

Mountain Hemlock Western White Pine Whitebark Pine

California Mixed Subalpine (256) Red Fir (RFR)

Red Fir Forest (8.531)

Red Fir Forest (85310)

Red Fir (207)

Coast Range Montane Forest (14)

Red Fir Forest (15)

Red Fir Fed Fir-Noble Fir

Red for forest (14b2) (continued)

TABLE 1.8 (continued)

WHR type

Cheatham and Haller(1975) CNDDB(1986)

Eyre (1980)

Kiichler(1977)

Lodgepole Pine (LPN)

Lodgepole Pine(8.61)

Lodgepole Pine Forest (86100)

Lodgepole Pine(218)

Whitebark Pine

Whitebark PineLodgepole Pine Forest (86220)

Sierra Nevada Mixed Conifer (243)

Upper Montane- Lodgepole Subalpine Forests Forest (16) (17)

Sierran Mixed

Ultramatic White

Conifer Forest (8.423)

Pine Forest (84160)

Pacific Ponderosa PineDouglas-fir (244)

Sierra Big Tree (8.425)

Ultramafic Mixed Coniferous Forrest (84180)

Lodgepole Pine Forest(8.622) Sierran Mixed Conifer (SMC)

NOTE:

Munz and Keck (1973)

Parker and Matyas(1981)

Paysen et al. (1980)

Proctor et al. (1980)

Lodgepole Pine

Lodgepole Pine

no corresponding vegetation type

Lodgepole Pine forest (14cl)

Big Tree

Mixed Conifer

not applicable

Yellow Pine forest (14a2)

Southern Montane Subalpine Forest (18)

Sierran(14)

] effrey Pine (NCM, Yellow Pine NI, NS) Forest Mixed Conifer-Fir

Sierran Mixed Coniferous Forest (84230)

Pine (NMC, NI, NS, SS)

Big Tree Forest (84250)

Ponderosa Pine(NI, SS)

Mayer and Laudenslayer (1988).

Thorne(1976)

Mixed Conifer-

Mixed conifer forest (14a3)

grazing, recreational use, and other human activities. To accomplish this goal they recognized that classification of ecosystems was a high priority. Much of their effort in the 1970s and 1980s was directed toward classification (Paysen et a1. 1980; Paysen 1982; Paysen, Derby, and Conrad, 1982; Hunter and Paysen 1986; AlIen 1987). The purpose of these efforts was to develop a land classification that could be applied to both research and management activities. Their approach differs fundamentally from earlier classifications of California's vegetation because it relies on field samples (plots) of the vegetation to build the bottom layers of the hierarchy. Consequently the basic taxonomic units of the system were not identified by arbitrarily assigning a name to what is generally perceived as a unique type. Instead, the basic units are defined after a number of plots are analyzed over a large area. Rather than qualitative and anecdotal, this classification is quantitative and driven by the availability of data.

Quantitative Classifications

Until recently, California has lagged behind Oregon, Washington, Idaho, and Montana in developing data-driven vegetation classifications. Efforts by Daubenmire (1952); Daubenmire and Daubenmire (1968); Franklin, Dyrness, and Moir (1971); Pfister and Arno (1980); and others set the standards for the USDA Forest Service, Pacific Southwest Region classification of forests, woodlands, and scrublands in California. Because the focus on quantification of vegetation on Forest Service lands was largely for management of forested landscapes for timber production, the concept of "potential natural vegetation" was used to define the vegetation types. Potential natural vegetation (PNV) is the vegetation that would become established if all successional sequences were completed without human interference under present climatic and edaphic conditions (Brohman and Bryant 2004; Winthers et a1. 2004). The larger floristic units of these efforts became known as series and the smaller basic units as associations. Series were identified by the dominant plants in the overstory, which would be likely to occur in mature (PNV) stands; whereas the associations were identified by characteristic species in the understory layers. There is an ecological basis for grouping associations into a series. For example, although there are many associations of white fir forest in the Sierra Nevada, all occupy sites that are warmer than the red fir-dominated forests. Red fir typically occurs at higher elevations or on cooler exposures, so the dominant overstory species reflect broad-scale environmental differences. The presence of certain understory species reflects more localized differences related to microclimate and soil. However, the implications of "potential" vegetation were difficult to come to grips with in many cases, particularly ib light of more recent studies suggesting that there is usually an array of natural processes including fire,

flood, disease, climatic shifts, and so on, which alter detailed long-term linear predictability of the chronological evolution of a given stand of vegetation (Westoby, Walker, and Noy-Meir 1989; Briske, Fuhlendorf, and Smeins 2003). The framework of the classification system adhered to by the USDA Forest Service in California is described by AlIen (1987). This project, most active in the early 1990s, entails eight different Forest Service zone ecologists developing classifications of targeted areas. The work includes extensive sampling of the vegetation, soils, and other environmental characteristics followed by quantitative analysis using multivariate statistical programs, which became more common and useful during the 1980s. Classifications are completed for Port Orford-cedar forests Gimerson 1994; Jimerson and Daniel 1999). Douglas-fir and tanoak forests Gimerson 1993; Jimerson et a1. 1996) and serpentine vegetation Gimerson et a1. 1995) of the western Klamath Mountains. Classifications for blue oak woodlands and redwood forests in the Central Coast region have been developed (Borchert, Segotta, and Purser 1988; Borchert et a1. 1993), for mixed conifer forests of the northern Sierra Nevada (Fites 1993) and red fir forests (Potter 1998) of the central and southern Sierra Nevada, for eastside pine forests of the Cascade Range, Modoc Plateau, and northern Sierra Nevada (Smith 1994), and for chaparral types of the Transverse and Peninsular Ranges (Gordon and White 1994). Recently, classifications for the herbaceous upland vegetation of the North Coast Ranges Gimerson et a1. 2000), the northern oak woodlands of the inner north Coast Ranges Gimerson and Caruthers 2002), the chaparral and coastal scrub of the Los Padres National Forest (Borchert et a1. 2004) and the riparian vegetation of the Sierra Nevada (Potter 2005) have also been completed. All of these efforts relied on quantitative plot-based sampling and analysis of representative stands of mature vegetation (again in keeping with the notion of potential vegetation as opposed to existing vegetation, which would include stands subject to recent disturbance). Regular meetings and discussion among the zone ecologists with reference to standards for classification and nomenclature assisted in developing a unified resolution and set of definitions for each of the types defined.

The Manual of California Vegetation

In 1991, the California Native Plant Society convened a committee to develop a standardized vegetation classification system for the state. The underlying assumption: that a unified classification treating the entire state with quantitative rules can provide the necessary means to develop defensible definitions for vegetation and natural communities, and these could be used to further statewide multispecies conservation goals. This notion was collectively agreed to by the committee, which included many state and federal agencies and academics, and several organizations with conservation

VEGETATION CLASSIFICATION IN CALIFORNIA

27

goals. The advent of new state legislation such as the Natural Community Conservation Planning Act in 1991 furthered the importance of this committee. In 1995 under the auspices of the CNPS Vegetation Committee (Keeler-Wolf 1993; Barbour 1995; Keeler-Wolf and Barbour 1997), the first edition of the Manual of California Vegetation (Sawyer and Keeler-Wolf 1995) was published. This classification was a synthetic effort to compile all information known about the state's vegetation into a unified framework that relied on quantitative defensible descriptions to arrive at an unequivocal definition of the major vegetation types (called "series" in the first edition). The classification relied heavily on the U.S. Forest Service's and the budding National Vegetation Classification System's (Bourgeron and Engelking 1994) notion of dominance by layer (tree, shrub, or herb) to define the major types of vegetation. However, it differed philosophically, in that it attempted a classification of existing vegetation regardless of its potential outcome. Thus, several of the series identified by Forest Service Ecologists were treated somewhat differently within the MCV. For example the Tanoak (Lithocarpus densiflorus) alliance was treated very broadly by Jimerson et al. (1995) to include many plant associations that were strongly dominated by Douglas-fir (Pseudotsuga menziesii). They were so included because of the premise that Tanoak is really the predominant late seral species and indicative of the PNV environment. In the MCV a Tanoak, a Tanoak-Douglas-fir, and a Douglas-fir alliance were recognized to indicate the range of current conditions expressed by dominance in the tree layer of the vegetation. Individual plant associations defined within the Jimerson et al. descriptions could easily be translated into the appropriate MCV alliance based on the quantitative descriptions stating such things as overstory dominance and constancy of major species. Although the first edition of the MCV defined 245 series

Principal among them was a category known simply as "habitats." These included such complex fine-scale matrices as alpine vegetation, montane and subalpine meadow and wetland scrub habitat, and subalpine upland shrub habitat. All of these types were thought to be composed of individual alliances and associations, which had not yet been further refined by quantitative vegetation analysis. Complex patterning of vegetation in the wide range of the state's vernal pools was also left open to broader interpretation, as insufficient information was available at the time to define individual patterns of floristic variation analogous to series. The MCV relied on the existing Holland (1986) classification for ve~nal pools with some minor modification. One final category was also defined: "unique stands." Unique stands are defined as one-of-a-kind types of vegetation that are not known to have one of the classic features of real vegetation-they do not display stand redundancy-repeating stands within similar ecological settings. Unique stands have been defined for locally dominant populations of rare species, such as several cypress species (Cupressus goveniana, C. forbesii, C. arizonica subsp. arizonica, C. arizonica subsp. nevadensis) only known from a few groves. Unique stands also have been defined for Unique stands structural situations such as Holly-leaf cherry "forests" where due to very long intervals between fire, Prunus illicifolia stands, typically shrubby, have grown into stands of tree-size individuals in rare instances. The California Natural DiverSity Database considers all types of unique stands as having conservation value. Philosophically, the MCV was a combination of a "topdown" and a "bottom-up" classification with emphasis on the finer resolution series and associations. The members of the CNPS vegetation committee and the authors believed it was most important to set up a standardized floristic basis for quantitatively defining the full variation of vegetation within the state. The inductive approach supported by most phytosociologists (e.g., see Major 1988) emphasizes

and other categories, it did not purport to be a final classifi-

the fine-scale associations as the basis of the classification

cation. In reality, the authors realized that it might take many years for a "final" classification for the complex vegetation of the state to be completed. The main purpose of the first edition was to provide a framework on which further descriptions could be built. The vegetation of the state had not been fully quantified; thus it was impossible to develop quantitative rules for all of the types in the state. In most cases the best quantitative descriptions could be developed for the wooded (forest and woodland) vegetation because of the ongoing focus by the U.S. Forest Service on the economically important forested vegetation types. Some shrub types had also been defined quantitatively, particularly chaparral of the south coastal area, and some herbaceous types, as well. However, in many cases the rules of dominance or the natural range of variation was not well enough understood to develop series-level descriptions for them. Thus, there were several other categories besides series that were used as provisional "place holders" for widely recognized, but quantitatively poorly understood vegetation.

and these are aggregated into larger floristic units (the series). However, some of the series and the other units such as the habitats, vernal pools, and unique stands had no quantitative data and were defined tentatively based on expert opinion, using the basic rules of dominance to define them. These rules identify dominant species as comprising greater than 50% relative cover of the top-most layer of vegetation. If there are several co-dominants, these species are listed together defining the name of the type. The authors expected these and other categories to be considered "place-holders" until further quantitative data and analysis replaced them.

28

VEGETATION CLASSIFICATION IN CALIFORNIA

The National Vegetation Classification System

Among the criticisms of the first edition of the MCV was the fact that it did not identify a classification hierarchy within which the relatively fine-scale classification units of series and associations could be placed. This classification hierarchy

was being worked out by a consortium of ecologists working with a panel of the Ecological Society of America (ESA Panel 2004). In 1997 a prototype national vegetation classification was unveiled by The Nature Conservancy, which drew on the advice of the ESA panel. This classification system was also adopted by the Federal Geographic Data Committee (FGDC) on vegetation as the standard for all national and federal projects relating to vegetation (FGDC 1997). A year later the first version of the entire National Vegetation classification system (NVCS) complete from associations up through all seven levels of the hierarchy was published by The Nature Conservancy (Grossman et al. 1998). The lowest level of the NVCS hierarchy is the fundamental level of the association, defined in the same way as it was in the MCV. The alliance was the next level up and was defined in the same way as the series was in the MCV. The principal difference was semantic. Series was thought by the ESA Panel to have too much of a connection with the notion of Potential Natural Vegetation, since the term was first popularized by forest ecologists working with potential vegetation concepts. Alliance was a term that arose from the European school of phytosociology (Braun-Blanquet 1932/ 1951, Muller-Dombois and Ellenberg 1974) and implied existing, not potential, vegetation. In the way that the term was applied in the MCV, there was no real philosophical difference between series and alliance. Above the level of the alliance the national classification shifted from a floristic to a physiognomic basis for classification. Starting with the next level up from alliance-the formation, all other classification units were based on a combination of life form of the predominant species in the stands and subdivisions of those life forms based on plant size, leaf morphology, phenology (e.g., deciduousness), or hydrology (Table 1.9). Current revisions of the MCV are well underway, and every effort was made to make the upcoming second edition compliant with the National Classification. For example, all series described in the first edition will be re-described as alliances, and the alliances and associations treated will be discussed within the framework of the NVCS. The advent of the new state and national vegetation classification standards spurred a number of significant events within the state. These included the first integrated mapping and classification projects (Keeler-Wolf, Lewis and Roye 1996; Keeler-Wolf, Roye, and Lewis 1998; Keeler-Wolf, Vaghti, and Kilgore 2000). It also enabled agreement and cooperation between state and federal agencies doing mapping and classification work within the state to agree to a set of standards for mapping and classification of vegetation (Vegetation Memorandum of Understanding 2001) adhered to by all major agencies and organizations involved in state vegetation classification and mapping.

Integrated Vegetation Mapping and Classification Within a few months following the publication of the MCV, a project was started at Anza-Borrego Desert State Park,

TABLE 1.9

Example of the National Vegetation Classification Hierarchy of a Serpentine Coniferous Woodland Vegetation Type Found in the Western Klamath Mountains

Example

Category Physiognomic Categories Class

Open Tree Canopy

Subclass Group

Subgroup Formation

Evergreen Open Tree Canopy Temperate or Subpolar Needleleaved Evergreen Open Tree Canopy Natural/Seminatural Rounded-crowned temperate or subpolar needle-leaved evergreen open tree canopy

Floristic Categories Alliance Association

Pinus jeffreyi Woodland Alliance Pinus jeffreyi/Quercus vaccinifolia-Arctostaphylos nevadensis/Festuca idahoensis Association

which for the first time integrated a field-based vegetation sampling and classification scheme with a detailed GISbased map of the vegetation using the new standards for quantification of vegetation set forth in the MCV and the National Vegetation Classification system (Keeler-Wolf, Roye, and Lewis 1998). Over 500 reieve samples were taken using a gradient-directed approach (Austin and Heyligers 1991, Gillison and Brewer 1985). These samples were analyzed using classification programs such as TWINSPAN (Hill 1979) and a quantitative key was developed which enabled all types of vegetation in the park to be identified. This latter asset was particularly valuable when applied to an accuracy assessment of the concomitantly developed map of the park. Some 95 individual vegetation-mapping units were defined and mapped. Mapping was done using traditional air-photo interpretation methods with the hand-delineated polygons digitally transferred and orthocorrected to a standard map base derived from orthorectified satellite imagery. About 28,000 individual polygons were retained in the mapping area of approximately 930,000 acres. A partial accuracy assessment of the mapping product showed mean accuracy among the types to be 82% (range 78~)-1 00 % ) using a standard binomial distribution algorithm (Cochran 1977, Meidinger 2003).

VEGETATION CLASSIFICATION IN CALIFORNIA

29

SoquoIaI KIng.

canyon NF "OR/CULTURE 1II

..-c:.. I.

PoInt lleyu

MA (BLII)

c;

0

100

FIGURE 1.7 Location of integrated mapping and classification projects in California as of late 2004. These projects relied upon a data-driven classification, extensive field data, and a manual photo-interpretation of either digital or hard-copy aerial photography. Figure produced by A. Mahaney 200S.

One of the most important developments came as a result of a well-funded initiative by the National Park Service to develop a new standard for mapping and classifying vegetation throughout all national park units (http://biology.usgs.gov/npsveg/standards.html). From 1996 through the present, several National Park units within California began active mapping and classification programs. The first to finish were Point Reyes National Seashore, Golden Gate National Recreation Area, and Yosemite National Park. Other similar projects were instituted in within the past several years. These included mapping and classifying the vegetation of the majority of the Mojave Desert (Thomas et al. 2004), Western Riverside County (Klein and Evens 2005), and Suisun Marsh (Keeler-Wolf, Vaghli, and Kilgore 2000). Each of these projects used a gradient driven sampling program, quantitative analysis, and air-photo interpretation-based vegetation mapping in a GIS environment (Fig. 1.7). Standardized data collection procedures were developed through state and national efforts. These included reieve sampling protocols developed cooperatively by state and national programs (see both the national park vegetation mapping program website and the CNPS Vegetation Program sampling protocol Web site: http://www.cnps.org/programs/vegetation/protocol.htm) There are several significant advantages to these integrated mapping and classification projects. From the sampling and classification standpoint they provide significant new baseline field data for long-term monitoring as each sample point is located using GPS at a minimum and in

30

VEGETATION CLASSIFICATION IN CALIFORNIA

some cases have additional permanent markers. The sample allocation process uses the gradient-driven approach to develop a series of samples representative of the full array of vegetation in a given area. Thus, each area is thoroughly sampled. These samples are analyzed using a standardized protocol that includes the following steps: a. Screen all sample-by-species data for outliers. Samples that are more than 3 standard deviations (SOs) away from the mean are removed, and species that are in fewer than three samples are removed. b. Run presence-absence Cluster Analysis (typically Sorensen's flexible beta method) to determine general arrangement of samples. c. Run cover class category Cluster Analysis to display a more specific arrangement of samples based on species presence and abundance. d. Run Indicator Species Analysis (Dufrene and Legendre 1997) at each of the successive group levels in the Cluster Analysis output, from two groups up to the maximum number of groups (all groups have at least two samples). e. Settle on the final representative grouping level of each Cluster Analysis to use in the preliminary labeling. f. Preliminary label alliance and association for each of the samples, and denote indicator species from the rSA. g. Develop decision rules for each association and alliance based on most conservative group member-

FIGURE 1.8 Side-by-side comparison of Landsat Thematic Mapper-based CalVeg classification vegetation map 2002 version (right) and hand-delineated and attributed NVCS-based vegetation map derived from 1:15,840 scale Color infrared aerial photography delineated in 2001. Area shown is the vicinity of Lake Elanor and Cherry Lake, Tuolumne County. Area on right side of CalVeg image has been merged with generalized version of hand-delineated map. Scale is 1:30,000.

ship possibilities based on review of species cover on a plot-by-plot basis. h. Relabel final alliance labels for each sample and arrange in table of database. i. Use decision rules developed in the new data to assign alliance and association names to all analyzed data and all outlier samples removed from data set. In some cases, ordination techniques such as Bray-Curtis or Nonmetric Multidimensional Scaling (McCune and Grace 2002) are also used to further define the environmental correlations with the vegetation classification units described. Keys to the vegetation are written and standardized descriptions are also written for each of the newly defined associations or alliances. These descriptions are sent to the western regional office of NatureServe where they are subjected to a review process and if ratified and considered new, entered into the National Vegetation Classification System (NatureServe 2004). Because the projects are usually funded over a time frame of 1 to 3 years, the data are collected, entered, and analyzed in relatively short order, becoming available to the parks and to other organizations relatively qUickly. Recent projects such as The Santa Monica Mountains National Recreation Area (Keeler-Wolf and Evens 2006) and Western Riverside County (Klein and Evens 2005), have taken advantage of the CNPS Rapid Assessment sampling approach for a number of vegetation types that tend to have relatively low within stand species diversity, such as chaparral and coastal scrub types. This technique allows for salient environmental variables to be collected along with a stand-based plotless estimate of cover of all major plant species in different strata (CNPS 2005). The number of

samples has increased to well over 1,000 for projects using the Rapid Assessment technique. Higher sample sizes for each vegetation type enables richer description of the variation of all vegetation types and also allows for a more detailed accuracy assessment of the map, since the Rapid Assessment technique also doubles as an accuracy assessment protocol. Releve samples using the CNPS reieve protocol (CNPS 2004), are used for vegetation with more floristic diversity such as woodlands, forests, grasslands, and wetlands. From the mapping point of view, the map products are very detailed and spatially accurate compared to most other currently available products (Fig. 1.8). The minimum map unit is usually 0.25-0.5 ha. The maps line-up well with standard digital orthophoto quarter quadrangles and therefore proVide a satisfying fine scale resolution product suitable for use with standard USGS mapping products such as 7.5minute quadrangles (Fig. 1.9). The process of developing the maps accompanying these detailed classifications has taken two directions depending upon whether stereo aerial photo pairs or digital orthophoto-graphs are the basis of the imagery. Classic photo-pair interpretation using the high-resolution diapositives of either true calor or calor infrared photos usually of from 1:9600 to 1:24,000 still proVide the best baseline for these detailed maps. These have been the basis for all of the National Park Service mapping done in the state since 1996. However, the process of taking hand-delineated polygons, usually drawn on acetate or mylar affixed to the photo image, then georeferencing, scanning, and orthorectifying the polygon boundaries and mosaikking those individual air-photo scannings into a GIS layer with attributes for each polygon takes much time and money.

VEGETATION CLASSIFICATION IN CALIFORNIA

31

1000

o

1000

2000

3000

4000

5000 Feet

Portion of the vegetation map for the Golden Gate National Recreation Area, Marin Co. overlaid on a standard USGS 7.S-minute topographic map series depicted at 1:10,000 scale. Acronyms are alliance labels for each polygon. CoBr = coyotebrush; CoLO = Coast Live Oak; POak = Poison oak; CaGrN= California annual grassland with Native component (a map unit aggregation of several vegetation types); PNeGr = Purple Needlegrass; MxBm = Mixed Broom (nonnative Genista and Spartium spp.); CaBay = California Bay; AnGr = Annual Grassland; DoFi = Douglas-fir; Euca = Eucalyptus sp.; MoCy = Monterey Cypress (planted); BIB= Bluebush (Ceanothus thyrsif/orus), Urb = urbanized (development). Map produced by A. Kilgore, Department of Fish and Game 2001.

FIGURE 1.9

An alternative approach is to rely on already orthorectified imagery such as USGS Digital Orthophoto Quadrangles (DOQQ) or independently developed orthorectified photos that already meet USGS specifications for spatial accuracy. These may be loaded as a base image within a computer's GIS project. The photo interpreter simply delineates polygons using a "head's-up" approach where on-screen digitizing is accomplished using a mouse or a digitizing tablet, while visually tracing the cursor over the boundaries of vegetation stands. The clear advantage of this approach is the reduction of time in processing and piecing together the hand delineations. The disadvantage is the loss of the three-dimensional stereo-pair perspective, valuable for gleaning subtle information on vegetation structure often necessary to define habitat stage (Mayer and Landenslayer 1988) or floristic associations or alliances. Recent work on regional planning efforts such as Napa County (Thorne et al. 2004), Western Riverside County (Klein and Evens 2005), and in San Diego County (Evens and San 2005) has made use of this technique. These projects have maintained attribute information on height and structure of the vegetation by relying more heavily on field sampling and verification to take the place of three-dimensional stereo photo interpretation. Using the head's-up approach a photo interpreter can attribute the individual polygons directly into a computer database, which can be linked with the GIS. The maps produced can be customized based on the many attributes stored and linked within the GIS, thus doing away with the complex polygon notation and symbolism for such earlier products as the S-V maps (see Fig. 1.4). On some of these projects the use of both DOQQs (as the basis for on-screen delineation) and ancillary stereo pairs (set up with a stereoscope and light table inlmediately adjacent to the computer work-station) assists in making the fine-level distinctions whenever necessary. Recent technological advances such as split-computer-screen digital paired photography and stereoscope lenses allow for threedimensional viewing of the digital imagery and head's-up digitizing may make this a more fluid process making the best of both worlds. Currently the detailed mixed mapping and classification process is being undertaken on a project-by project basis throughout the state. The National Park Service, National Forest Service, Bureau of Land Management, Bureau of Reclamation, U.S. Geological Survey, California Department of Fish and Game, University of California, and California State Parks, have all been involved in such projects in the past 9 years. Currently approximately 200/0 of the state has been mapped and classified using this approach (see Fig. 1. 7). The California Native Plant Society has been influential, both through its Vegetation Committee and its Vegetation Program, in instituting standards in both classification and mapping for these projects (Evens and Keeler-Wolf 2003).

Developments and Future Directions Cooperative Mapping and Shared Standards

Traditionally, one of the most difficult issues for mapping and classification of vegetation has been the widely varying needs and mandates driving these projects. Colwell (1977) concluded his discussion of California vegetation mapping by stating that SO years of different styles of mapping serving different agencies and different needs has not produced a completely compatible and uniformly applicable system for understanding the dispersion of vegetation across the state. In addition to the CNPS Vegetation committee's efforts to standardize and refine the state's vegetation classification, other attempts have been made in the past 30 years to coordinate efforts among the various agencies including one preceding the CNPS vegetation committee in the mid-1980s where several state and federal agencies convened in meetings for several years to attempt adoption of a standard vegetation classification. The advent of the NVCS and its national acceptance over the last few years has facilitated an integrated approach to classifying and mapping through a federally mandated (FGDC 1997) acceptance of this as the standard. Over the past few years under the direction of the State Biodiversity Council (http://ceres.ca.gov/biodiversity), the Vegetation Mapping Memorandum of Understanding was developed. A committee representing the state and federal agencies and other organizations involved in California vegetation mapping was convened. The interagency group has agreed upon the use of the NVCS as the standard classification system. This committee has developed a set of standards and guidelines (http://ceres.ca.gov/biodiversity/ vegmou.html). Important among these were a list of attributes to be used by cooperators (Table 1.10). A mapping project using these attributes could address many of the needs of the cooperating agencies including wildlife predictive modeling, identification of floristic vegetation units relating to the NVCS, and fire and fuels modeling and monitoring. Currently the Committee is cooperating to test the efficient use of these attributes and the ability of the agencies to cooperate on statewide vegetation mapping projects. Due to different mandates and different funding sources between the various state and federal agencies cooperative mapping and classification remains to this day an elusive goal. However, there is general agreement on which agencies and organizations are involved in mapping and classification of different parts of the state and a more cooperative spirit has been engendered through the committee, which has reduced duplication of effort and increased compatibility of products. It is unlikely that fully cooperative mapping and classification will become entirely developed until long-term stable funding for coordinated efforts exists.

VEGETATION CLASSIFICATION IN CALIFORNIA

33

TABLE 1.10

Attributes Recommended by 2003 State Biodiversity Council Vegetation Mapping MOU Group

Map Unit Attribute

Core at All Scales

Life Form (Cover Type)

.I

Ecological Unit

.I

MCV Hierarchy (derived from other fields) -this consists of multiple fields, one for each hierarchical level

.I

CALVEG Hierarchy (derived from other fields) -this consists of multiple fields, one for each hierarchical level

.I

Land Use/Land Cover-Anderson Level 1

.I

Birdseye Total Cover (max value 1000/0)

.I

0/0 of Birdseye Total Cover by trees (canopy

.I

Optional at All Scales

Notes

closure-sum of conifers and hardwoods 0/0 cover) 0/0 of Birdseye Total Cover by conifers

.I

0/0 of Birdseye Total Cover covered by

.I

hardwoods (and not covered by overstory trees) 0/0 of Birdseye Total Cover covered by shrubs

.I

(and not covered by trees) 0/0 of Birdseye Total Cover covered by

.I

herbaceous (and not covered by trees or shrubs) Map Unit Aggregation Type (changed name from Internal Diversity)

.I

Attribution Method (field-based, modeled, etc)

.I

Cause and Date of Record Change (fire, error correction, etc.)

.I

Broad scale: does not apply Medium scale: to species for trees, to genus level for other life forms Fine scale: to species level

Dominant Species (visible from above, usually 1 to 3 species) by Layer

Size Class (DBH)

.I

Broad scale: does not apply

Height class and/or Vertical Structure

.I

Broad scale: optional; see life form Medium scale: optional; use structural classes for trees (1, 2, or multiple layer Fine scale: core

Land Use/Land Cover-Anderson Level 2

WHR Type (derived from other fields)-this consists of multiple fields, one for each hierarchical level Shrub Structural Diversity (includes live/dead fuel ratio)

.I

Broad scale: optional Medium scale: optional Fine scale: core Broad scale: optional Medium scale: core Fine scale: core

.I

.I

TABLE

Map Unit Attribute

1.10 (continued)

Core at All Scales

Disturbance Index (roads, exotics, erosion, other impacts)

Optional at All Scales

Notes Broad scale: derived from other data; coarse (low) level of detail Medium scale: derived from other data or remotely sensed; medium level of detail Fine scale: observed or derived from other data; high level of detail

Groundlevel Total Vegetation Cover (max value 4000/0-a sum of all cover values from up to four different structural layers) Groundlevel Total Conifer Cover (max value 1000/0) Groundlevel Total Hardwood Cover (max value 1000/0-all hardwood cover, whether covered by overstory trees or not) Groundlevel Total Shrub Cover (max value 1000/0-all shrub cover, whether covered by trees or not) Groundlevel Total Herbaceous Cover (max value 1000/0-all herbaceous cover, whether covered by trees/shrubs or not) Percent Mortality

Broad scale: coarse level Medium scale: by type of tree

Special Habitat Elements (related to vegetation only-snags, downed logs, etc; not caves, cliffs, etc.)-Observed

Broad scale: does not apply Medium scale: does not apply

Age Class (difficult to capture using remote sensing)

Broad scale: does not apply

NOTE: Broad scale = range of 1:7,500,000 to 1:250,000, typical polygon size of 6AOO to 64,000 acres, and MMU of 50 acres; medium scale = range of 1:250,000 to 1:24,000, typical polygon size of 1,000 to 10,000 acrease, and MMU of 10 acres; fine scale = range of 1:24,000 to 1:6,000, typical polygon size of less than 1,000 acres, and MMU of 5 acreas.

Image Segmentation and Combined Segmentation and Delineation

Despite the recent successes of the integrated classification and photo-interpretation-based mapping projects, the costs of these projects remain relatively high, and the number of expert aerial photo interpreters familiar with the national vegetation classification system are few. One hope for addressing these issues has come through advancements of image segmentation of more fine-scale imagery products, coupled with machine-based learning programs. Several projects have recently been undertaken to test the efficiency of these approaches compared with more traditional photointerpretation approaches. A test of image segmentation versus photo interpretation was done under the auspices of the Biodiversity Council's Vegetation MOU committee

(http://ceres.ca.gov/biodiversity/Meetings/Special/ILCMP_0 6.03.pdf) with the results described by Evens and KeelerWolf (2006). The U.S. Forest Service along with UC Davis researchers also recently conducted a pilot study of vegetation mapping using image segmentation in the Lake Tahoe Basin Management Area (Dobrowski et a1. 2006; Greenberg 2006). Image segmentation has also been used in vegetation mapping of Portola and Butano State Parks in the Santa Cruz Mountains of San Mateo County (Roy Woodward personal communication 2004), and is currently being investigated for the Point Reyes National Seashore as a means to update and revise the map based on aerial photography. Preliminary feedback from these projects suggests that with proper field calibration and aggregation of false reflectance and shading polygons reasonable results may be obtained.

VEGETATION CLASSIFICATION IN CALIFORNIA

3S

FIGU RE 1.10 Side-by-side location of 1933 VTM and 1998 Yosemite National Park and Environs vegetation maps for the Twin Lakes area of the Eastern Central Sierra ca 9 km southwest of Bridgeport, Mono County. Note similar resolution of map polygons. Shaded polygons on the right image are Artemisia tridentata-Purshia tridentata mixed alliance stands, analogous to the mapping unit"Atr-Pt" on the left image.

However, full accuracy assessment has not been completed on any of these projects. Without extensive field verification mapping below basic physiognomic units such as the National Vegetation Classification "Class" level remains elusive (Greenberg et al. 2006).

Change Detection

One of the most valuable tools yet to be fully developed is the detailed assessment of change based on current and past vegetation conditions. As discussed previously, Levien et al. (1998, 1999) have developed a version of change detection using large pixel satellite imagery for wooded areas. In addition, Herwitz, Sandler, and Slye (2000) have quantified local crown change of oaks in woodlands and savannas in the central coast range using detailed measurements of repeat aerial photography, and Vaghti and Keeler-Wolf (2004) have developed change detection for fine-scale, largely herbaceous vegetation in the Suisun Marsh, Solano County. How-

36

VEGETATION CLASSIFICATION IN CALIFORNIA

ever, much remains to be clarified and for a midscale view of vegetation throughout the state. The existence of the Wieslander (1961) VTM maps proVides a tempting basis for a comparision for about 40% of the state. The University of California Davis Information Center for the Environment has been experimenting with methods to efficiently scan and orthorectify the original hard copy VTM maps Oeff Kennedy andJim Thorne personal communication 2004). Comparison of the VTM and current mid- to fine-scale-mapping efforts seems possible, because of similar resolution and rules governing the depiction of overstory dominants (Fig. 1.10). Recent work by Thorne et al. (2006) has substantiated an efficient methodology for digitizing the old VTM maps and has underscored some promising comparisons between current and 1930's era conditions. However, using existing old maps compared to current maps always raises questions about the underlying methodological differences that could strongly skew interpretation of results. Careful interpretation and likely broadening of scales of assessment to allow for better

classification matching may be necessary in some cases (e.g., fine-scale patterning of nonwoody vegetation). The uncertain value of comparative studies using Wieslander (1961) plot data and relocated up-to-date samples, especially for difficult to locate stands of chaparral and other scrub, has been recently discussed by Keeley (2004). The practice of Ifretrospective" mapping of vegetation has not been fully experimented with, but is a promising avenue of future research. The concept of uniformitarianism (the present is the key to the past) could also be applied, and the photo signatures of old photos could be as easily interpreted as those of new. Existing detailed maps using the integrated approach of classification and mapping could lay the groundwork for the use of historical aerial photographs to step back in time and be photo-interpreted using the same techniques applied to the current photos, assuming a detailed field-based classification for the same area is conducted and that most of the signatures are analogous and interpretable. Because aerial photos were flow in many parts of California well back into the 1930s, a much more quantitative assessment of change in many different parts of the state could be made.

Melding of European Phytosociology and American Approaches

Some contentious debates have developed over the years between European and North American classification schools (e.g., Rejmanek 1997). Much of this debate has to do with the notion of whether it is more useful to work from the typically American top-down or the typically European bottom-up in the classification hierarchy (Ponomarenko and Alvo 2001). However, in California much cooperation has proceeded through the association of western vegetation ecologists and European phytosociologists in recent years. This work has been particularly fruitful in certain vegetation types such as serpentine grassland (Rodriguez-Rojo et al. 2001a, 2001b) and vernal pools (e.g., Barbour et al. 2003, 2005). As the California and the national vegetation classification systems grow using the plant association as the basis of vegetation description, there will be less and less to argue about, particularly as reanalysis of an ever-growing data set of comparable releve samples is developed.

Archiving and Analysis of Vegetation Plots

One of the most promising pathways for further refinement of the vegetation classification in the state involves the digital archiVing and dissemination of vegetation data. Several efforts are underway to develop Web-based or other data-archiving and retrieval systems. A national effort is being spearheaded through VegBank (http://vegbank.org/vegbank/index. jsp). VegBank is the vegetation plot database of the Ecological Society of America's Panel on Vegetation Classification. It currently contains over 21,000 plots nationwide including approximately 1,000 plots from California (mostly from

National Park mapping projects). This system enables qualified users to download, enter, and analyze plots most of which are releves. The California Department of Fish and Game in cooperation with the California Native Plant Society maintains a system of databases of all of their vegetation projects. Collectively called the California Vegetation Information System (CVIS), it currently houses about 15,000 samples collected within the past 10 years and includes releves, rapid assessment, and point intercept transect data. Information about CVIS can be obtained by contacting the author at ([email protected]). Previously collected data including the U.5. Forest Service Ecology Plots collected by U.5. Forest Service California Regional Ecology staff include approximately 7,000 plots (contact [email protected]), about 12,000 Forest Inventory and Analysis plots (archived through the U.5. Forest Service's Remote Sensing Lab archivist, Kama Kennedy; [email protected]), and virtually all of the known 18,000 original VTM plots collected by Wieslander crews in the 1930s have been entered by the combined forces of the Wieslander data group (http://vtm.berkeley.edu/data/). Cooperative efforts are under way to link these various databases and provide these data for further classification analysis and other uses. These data are beginning to be used for a number of statewide and national efforts including the LandFire (http://www.landfire.gov/index.html) analysis cooperative program with the U.5. Forest Service and NatureServe to develop national data driven fire models for all vegetation, and an effort to use vegetation data for modeling global climate change a. Thorne personal communication, January 2005). There is significant potential for these archived vegetation samples to be used for reference conditions for restoration efforts, modeling the occurrence of rare plant and animal species (in conjunction with vegetation maps), and many other purposes.

Filling in "Holes" in the Statewide Classification

Within a few years it will be possible to display all the locations and data for the thousands of vegetation samples taken statewide. With this will come the ability to physically see which parts of the state have been adequately represented or not by samples and determine which parts of the state are more in need of further basic sampling to refine the statewide vegetation classification. At this point the largest geographic gaps in our knowledge appear to be in the largely privately held lands of the outer and middle North Coast Ranges, and the outer and middle southern Coast Ranges from Santa Barbara County north to the San Francisco Bay. For the South and central Coast Ranges, federal- and state-managed lands are now undergoing vegetation assessments, but much has yet to be learned of the vegetation of the Santa Cruz Mountains, the Diablo Range, and the non-federally managed lands of the Temblor, the Cuyama, and other mountains that make up the central and South Coast Ranges. In the North Coast

VEGETATION CLASSIFICATION IN CALIFORNIA

37

Ranges west of the Mendocino National Forest and north of Point Reyes National Seashore there are large gaps in our classification knowledge until we reach Redwood National Park and the larger Redwood State Parks of Humboldt County. However, in the 30 years since the first edition was written, there has been exceptional activity in the state, now surpassing the work done in the 1930s by Wieslander's efforts. Given the current emphasis that vegetation classification and mapping is receiving, it should take less than 10 years before a relatively complete classification and detailed map of the state's natural vegetation exist.

References AlIen, B. H. 1987. Ecological type classification for California. General Technical Report PSW-98. USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. AlIen, B.H., B.A. Holtzman, and R.R. Evett. 1991. A classification system for California's hardwood rangelands. Hilgardia 59: 1-45. AlIen, B. H., R. R. Evett, B. A. Holzman, et al. 1989. Report on rangeland cover type description for California hardwood rangelands. Department of Forestry and Fire Protection and Department of Forestry and Resource Management, Berkeley, CA. Austin, M. P., and P. C. Heyligers. 1991. "New approach to vegetation survey design: Gradsect sampling." in: C. R. Margules and M. P. Austin (eds.), Nature conservation: Cost-effective biological surveys and data analysis, pp. 31-36. Commonwealth Scientific and Industrial Research Organization. Melbourne, Australia. Barbour, M.G. 1995. The story behind the Manual of California Vegetation. Fremontia 23: 19-22. Barbour, M.G., and]. Major (eds.). 1977. Terrestrial vegetation of California. Wiley-Interscience. New York. Barbour, M.G., A. Solomeshch, C. Witham, et al. 2003. Vernal pool vegetation of California: Variation within pools. Madrofio 50: 129-146. Barbour, M. G., A. Solomeshch, C. Witham, et al. 2005. Vernal pool vegetation of California: communities of long-inundated deep habitats. Phytocoenologia 35 (1): 12-34. Bechtold, W.A. and S.]. Zarnoch. 1999. Field Methods And Data Processing Techniques Associated With Mapped Inventory Plots. USDA Forest Service Proceedings RMRS-P-12, Rocky Mountain Research Station, 421-424. Borchert, M., D. Segotta, and M. D. Purser. 1988. Coast redwood ecological types of southern Monterey County, California. General Technical Report PSW-107. USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. Borchert, M.I., N.D. Cunha, P.C. Krosse, et al. 1993. Blue oak plant communities of southern San Luis Obispo and northern Santa Barbara Counties, California. General Technical Report PSWGTR-139. USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. Borchert, M., A. Lopez, C. Bauer, et al. 2004. Field Guide to Coastal Sage Scrub and Chaparral Series of Los Padres National Forest. Department of Agriculture-U.S. National Forest Service. Region Five Ecological Field Guide. Bouldin, ].R. 1999. Twentieth-century changes in forests of the Sierra Nevada, California. Ph.D. dissertation. University of California, Davis, CA. Bourgeron, P.S., and L. D. Engelking. 1994. A preliminary vegetation classification of the western United States. Unpublished report.

38

VEGETATION CLASSIFICATION IN CALIFORNIA

The Nature Conservancy, Western Heritage Task Force, Boulder, CO.

Braun-Blanquet, ]. 1932/1951. Plant sociology: the study of plant communities. McGraw-Hill, New York. Brewer, W. 1966. Up and down California. University of California Press, Berkeley. Briske, D. D., S. D. Fuhlendorf, and F. E. Smeins. 2003. Vegetation dynamics on rangelands: A critique of the current paradigms. Journal of Applied Ecology 40: 601-614. Brohman, R., and L. Bryant. 2004. Existing Vegetation Classification and Mapping Technical Guide. USDA Forest Service, Washington Office, Ecosystem Management Coordination Staff. Cheatham, N. H., and]. R. HaIler. 1975. An annotated list of California habitat types. Unpublished report. University of California, Berkeley, CA. Clements, F. E. 1916. Plant succession: An analysis of the development of vegetation. Publication 242. Carnegie Institute of Washington, Washington, DC. Clements, F. E. 1920. Plant indicators: The relation of plant communities to process and practice. Publication 290. Carnegie Institute of Washington, Washington, DC. CNPS. 2004. Releve sampling protocol April 2, 2004. Downloadable from: http://www.cnps.org/cnps/vegetation/protocol.php#instructions CNPS. 2005. Vegetation rapid assessment sampling protocol November 25, 2005. Downloadable from: http://www.cnps.org/cnps/ vegetation/protocol.php#instructions Cochran, W.G. 1977. Sampling techniques, 3rd edition. Wiley, New York. Collins, ]. B., and C. E. Woodcock. 1996. An assessment of several linear change detection techniques for mapping forest mortality using multitemporal Landsat TM data. Remote Sensing of Environment S6 (1): 66-67. Colwell, W. 1977. The status of vegetation mapping in California today. M.G. Barbour and ].A. Major (eds.) 1977. Terrestrial vegetation of California. Wiley-Interscience, New York. Colwell, W. 1974. Soil-vegetation maps of California. USDA Forest Service Resource Bulletin PSW-13. Cooper, W. S. 1922. The broad-sclerophyll vegetation of California. Publication 319. Carnegie Institution of Washington, Washington, DC. Critchfield, W. B. 1971. Profiles of California vegetation. Research Paper PSW-76. USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. Daubenmire, R. 1952. Forest vegetation of northern Idaho and adjacent Washington, and its bearing on concepts of vegetation classification. Ecological Monographs 22: 301-330. D'Erchia, T., C. Edwards,]. Ulliman, et al. 1993. Gap analysis: A geographic approach to protecting biological diversity. Wildlife Monographs No. 123:141 (Supplement to the Journal of Wildlife Management 57 (1)). Daubenmire, R., and]. B. Daubenmire. 1968. Forest vegetation of eastern Washington and northern Idaho. Technical Bulletin No. 60. Washington Agriculture Experiment Station, Pullman, WA. Davis, F. W., D. M. Stoms, A. D. Hollander, et al. 1998. The California Gap Analysis Project-Final Report. University of California, Santa Barbara, CA. Davis, F. W., P. A. Stine, D. M. Stoms, et al. 1995. Gap analysis of the actual vegetation of California 1. The southwestern region. Madrofio 42 (1): 40-78.

Dobrowski. S.Z., ].A. Greenberg, C.M. Ralnirez, and S.L. Ustin. 2006. Improving image derived vegetation maps with regression based distribution modeling. Ecological Modelling 192: 126-142. Dufrene, M., and P. Legendre. 1997. Species assemblages and indica-

Lake Tahoe Basin using hyperspatial image analysis. Photogrammetric Engineering and Remote Sensing 72: 581-589. Griffin, ]. R., and W. B. Critchfield. 1976. The distribution of forest trees in California. Research Paper PSW-82. USDA, Forest Service,

tor species: the need for a flexible asymmetrical approach. Ecological Monographs 67: 345-366.

Pacific Southwest Research Station, Berkeley, CA. Grossman, D. H., K. Goodin, M. Anderson, et al. 1998. International

ESA Panel. 2004. Jennings, M., O. Loucks, D. Glenn-Lewin, et al. 2003. Guidelines for describing associations and alliances of the U.5. National Vegetation classification. The Ecological Society of

classification of ecological communities: Terrestrial vegetation of the United States. The Nature Conservancy, Arlington, VA. Hanes, I.L. 1976. Vegetation Types of the San Gabriel Mountains.

Arl1erica Vegetation Classification Panel, Version 2.0 March 28,

pp. 65-76 in: ]. Latting, ed. Plant communities of Southern Cali-

2003.

fornia. California Native Plant Society, Berkeley, CA.

pling-A practical tool in range management. Jour. Range Management 10: 205-212. Evens, J., and S. San. 2005. Classification and mapping of vegetation

Hanes, I.L. 1977. Chaparral. Pages 417-469 in: M.G. Barbour and]. Major, editors. Terrestrial vegetation of California. Wiley-Interscience, reprinted by the California Native Plant Society 1988, Sacramento, CA.

in the San Dieguito River Park, San Diego County, California.

Haultain, S., I. Keeler-Wolf, J. Evens. 2004. Preliminary classification of

Evans, R.A., and R.M. Love. 1957. The step-point method for sam-

Unpublished report available from California Native Plant Soci-

vegetation of Sequoia and Kings Canyon National Parks. Draft report

ety, Sacramento.

on file at Sequoia National Park headquarters, Three Rivers, CA.

Evens, J., and I. Keeler-Wolf. 2003. Importance of vegetation in

Herwitz, S. R., B. Sandler, and R. E. Slye. 2000. Twenty-one years of

regional conservation planning. The Sampler (online) newslet-

crown area change in the Jasper Ridge Biological Preserve based

ter of the California Native Plant Society Vegetation Program

on georeferenced multitemporal aerial photographs. Interna-

5: 10-1 S. (http://www.cnps.org/programs/vegetation/PDFs/sampler_2003_final_version_12-2. pdf) Evens,]., and I. Keeler-Wolf. 2006.Vegetation classification, descriptions, and mapping of the Clear Creek Management Area, Joaquin Ridge, Monocline Ridge, and environs in San Benito and western Fresno Counties, California. Final report to Resources Legacy Fund. Grant #: 2004-0173. FGDC. 1997. Vegetation Classification Standard. FGDC-STD-005. Vegetation Subcommittee, Federal Geographic Data Committee. FGDC Secretariat, U.S. Geological Survey. Reston, VA. 58p. Fites, J. 1993. Ecological guide to mixed conifer plant associations of

tional Journal of Remote Sensing 21: 45-60. Hill, M.O. 1979. TWINSPAN: a Fortran program for arranging mu1tivariate data in an ordered two-way table by classification of the individuals and attributes. Section of ecology and systematics, Cornell University, Ithaca, NY. Holland, R. F. 1986. Preliminary descriptions of the terrestrial natural communities of California. Unpublished document, California Department of Fish and Game Natural Heritage Division, Sacramento, CA. Holland, V. L., and D.]. Keil. 1995. California vegetation. Kendall/ Hunt DubuqueIowa. 516pp.

the northern Sierra Nevada and southern Cascades. Technical

Hunter, S. C., and I. E. Paysen. 1986. Vegetation classification system

Publication. USDA, Forest Service, Pacific Southwest Region, San

for California: User's guide. General Technical Report PSW-94.

Francisco, CA. Franklin, ]. E, C. I. Dyrness, and W.H. Moir. 1971. A reconnaissance method for forest site classification. Shinrin Richi 13: 1-14. Gialdini, M.]., S. Titus, J. P. Nichols, et al. 1975. The integration of o1anual and automatic image analysis techniques with supporting ground data in multistage sampling framework for timber resource inventories. Three examples. Proe. NASA Earth Resources Survey Symp. 11 pp.

USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. Jensen, H.A. 1947. A system for classifying vegetation in California. California Fish and Game 33: 199-266. Jimerson, I. M. 1993. Preliminary plant associations of the Klamath province, Six Rivers and Klamath National Forests. Unpublished report. USDA, Forest Service, Six Rivers National Forest, Eureka, CA. Jimerson, I. M. 1994. A field guide to Port Orford cedar plant associ-

Gilison, A.N., and K.R. W. Brewer. 1985. The use of gradient directed

ations in northwest California. Technical Publication R5-ECOL-

transects or gradsects in natural resource survey. Journal of Envi-

TP-002. USDA, Forest Service, Pacific Southwest Region, San Fran-

ronmental Management 20: 103-127. Gopal, S., and C. E. Woodcock. 1994. Theory and methods for accuracy assessment of thematic maps using fuzzy sets. Photogrammetric Engineering and Remote Sensing 60: 181-188. Gordon, H.]., and I.C. White. 1994. Ecological guide to southern California chaparral plant series. Technical Publication R5-ECOLTP-005. USDA Forest Service, Pacific Southwest Region, San Francisco, CA.

cisco, CA. limerson, I. M., L. D. Hoover, E. A. McGee, et al. 1995. A field guide to serpentine plant associations and sensitive plants in northwestern California. USDA Forest Service, Pacific Southwest Region, publication R5-ECOL-TP-006. Jimerson, I.M, E.A. McGee, D.W. lones, et al. 1996. A field gUide to the tanoak and the Douglas-fir plant associations in Northwestern California. RS-ECOL-TP-009. USDA Forest Service, Pacific Southwest Region, San Francisco, CA.

Cordon, H., and B. Schwind. April 2000. The CalVeg Geobook-A Visual Tool for Describing Vegetation Information Derived From Rerl10tely Sensed Data. USDA Forest Service. Available at: http://www.fs.fed.us/r5/rsl!publications/. Goudy, C. B., and D. W. Smith. 1994. Ecological units of California: sub-

Jimerson I. M., and S. Daniel. 1999 Supplement to A field guide to Port Orford cedar plant associations in northwest California. Technical Publication R5-ECOL-TP-002. USDA, Forest Service, Pacific Southwest Region, San Francisco, CA.

sections [a 1:1000000-scale map]. USDA Forest Service, Pacific Southwest Region and USDA Natural Resource Conservation Service.

Jimerson, I. M, ]. W. Menke, S. K. Carothers, et al. 2000. Field Guide to the Rangeland Vegetation Types of the Northern Province.

Greenberg, ].A., S.Z. Dobrowski, C.M. Ramirez, J.L. Tuil, and S.L.

General Technical Report R5-ECOL-TP-014. USDA Forest Service,

Ustin. 2006. A bottom-up approach to vegetation mapping of the

Pacific Southwest Region, Berkeley, CA.

VEGETATION CLASSIFICATION IN CALIFORNIA

39

Jimerson, TM., and S.K. Carothers. 2002. Northwest California Oak Woodlands: Environment, Species Composition, and Ecological Status. USDA Forest Service Gen. Tech. Rep. PSW-GTR-184. Kauth, R.]., and G.S. Thomas. 1976. "The Tasseled Cap-A Geographic Description of the Spectral-Temporal Development of Agricultural Crops as Seen by Landsat." In Proceedings of the 2nd International Symposium on Machine Processing of Remotely Sensed Data, Purdue University, West Lafayette, IN, 4641-4651, June 1976. Keller-Wolf, T 1993. Rare Community Conservation in California. In J. Kelley (ed.), Proceedings of the Symposium: Interface between ecology and land development. Southern California Academy of Sciences. Keeler-Wolf, T, K. Lewis, and C. Roye. 1996. The definition and location of sycamore alluvial woodland in California. Unpublished report for the California Department of Water Resources, Sacramento. Keeler-Wolf, T, and M.G. Barbour. 1997. Conservation and classification of vegetation in California: a symposium. Fremontia 25: 17-27. Keeler-Wolf, T, C. Roye, and K. Lewis. 1998. Vegetation mapping and classification of the Anza-Borrego Desert State Park, California. Unpublished report on file at California Dept Fish and Game, Natural Diversity Database, Sacramento. Keeler-Wolf, T, M. Vaghti, A. Kilgore, et al. 2000. Vegetation mapping of Suisun Marsh, Solano County, California: a report to the Department of Water Resources prepared by Wildlife and Habitat Data Analysis Branch, California Department of Fish and Game, Sacramento. Keeler-Wolf, T, P. Moore, M. Schindel, et al. 2002. Vegetation classification and description of Yosemite National Park. NatureServe and Environmental Systems Research Institute. Unpublished report on file at Wildlife and Habitat Data Analysis Branch. 265pp. Keeler-Wolf, T, M. Schindel, S. San, et al. 2003. Vegetation classification of Point Reyes National Seashore and Golden Gate National Recreation Area. Unpublished report on file at Point Reyes National Seashore and at California Department of Fish and Game, Sacramento. Keeler-Wolf, T, S. San, and D. Hickson. 2005. Vegetation classification of Joshua Tree National Park, Riverside and San Bernardino Counties, California. Unpublished report on file at Wildlife and Habitat Data Analysis Branch, Califlornia Department of Fish and Game, Sacramento. Keeler-Wolf, T., and J. Evens. 2006. Vegetation Classification of the Santa Monica Mountains National Recreation Area and Environs in Ventura and Los Angeles Counties, California-Version 1-Association-level and specific alliances. A report submitted to National Park Service Santa Monica Mountains National Recreation Area by California Department of Fish and Game, Wildlife and Habitat Data Analysis Branch and The California Native Plant Society, Vegetation Program Sacramento, California Keeler-Wolf, T and J. Evens. 2006. Vegetation classification and descriptions for the Santa Monica Mountains National Recreation Area and environs, Los Angeles and Ventura counties, California. Report to the National Park Service and Earth Sciences Research Institute. On file at Wildlife and Habitat Data Analysis Branch, California Department of Fish and Game, Sacramento. Keeley, J. 2004. VTM plots as evidence of historical change: Goldmine or landmine? Madrono 51 (4): 372-378. Kellogg, A., and E. L. Greene. 1889. Illustrations of west American oaks. Published from funds provided by James N. McDonald, San Francisco, CA.

40

VEGETATION CLASSIFICATION IN CALIFORNIA

Kelly, M., B. AlIen Diaz, and N. Kobzina. 2005. Digitization of a historic dataset: the Wieslander California Vegetation Type Mapping Project. Madrofio 52: 191-201. Kelly, M., B. Allen-Diaz, and N. Kobrina. 2004. Making a Historic Vegetation Dataset Digital and Available to the World. forestry@berkeley July 2004, Volume 5, Issue 2. Kennedy, J., and J. Thorne. 2003. University of California Davis, Information Center for the Environment, personal communication. Klein, A., and J. Evens, 2005. A quantitative standardized vegetation classification and mapping database for Western Riverside County. California Native Plant Society and California Department of Fish and Game. Unpublished report on file at California Department of Fish and Game Wildlife and Habitat Data Analysis Branch, Sacramento. Lee, C. 2003. Vegetation Alliances and Associations of the Whiskeytown National Recreation Area. MS, Humboldt State University, Arcata, CA. Levien, L.M., C.S. Fischer, P.D. Roffers, et al. 1998. Statewide change detection using multitemporal remote sensing data. Presented at the First International Conference on Geospatial Information in Agriculture and Forestry, Lake Buena Vista, FA, 1-3 June 1998. Levien, L.M., P. Roffers, B. Maurizi, et al. 1999. A machine-learning approach to change detection using multi-scale imagery. Presented at the American Society of Photogrammetry and Remote Sensing 1999 Annual Conference, Portland, Oregon, May 20, 1999. Major, ]. 1988. "The Alpine chapter update." (pp. 1015-1016). ed. M. G. Barbour and]. Major, Terrestrial Vegetation of California New expanded edition California Native Plant SOCiety, Sacramento. Mayer, K., and W. Laudenslayer. 1988. A guide to wild-life habitats of California. State of California, The Resources Agency, Department of Forestry and Fire Protection, Sacramento, CA. McCune, B., and J. B. Grace. 2002. Analysis of Ecological Communities. MjM Software, Gleneden Beach, Oregon. Meidinger, D. V. 2003. Protocol for accuracy assessment of ecosystem maps. Res. Br., B.C. Min. For., Victoria, B.C. Tech. Rep.011. Milliken, ]., S. Gill, B. Beardsley, et al. 1997. A report of accuracy assessment methods and results for the Lassen - Modoc Northeastern California Cooperative Vegetation Mapping Project (Sacramento, CA: Region 5 USFS and California Department of Forestry and Fire Protection). Minnich, R.A. 1976. Vegetation of the San Bernardino Mountains. ed. J. Latting, Plant Communities of Southern California, Pages 99-125. California Native Plant Society. Berkeley, CA. Minnich, R.A., M.G. Barbour, ].H. Burk, et al. 1994. Sixty years of change in conifer forests of the San Bernardino Mountains. Journal of Conservation Biology. Minnich, R.A., M.G. Barbour, J.H. Burk, and R.E Fernau. 1995. Sixty years of change in Californian conifer forests of the San Bernardino Mountains. Conservation Biology 9:902-914. Mooney, H.A. 1977. Southern coastal scrub. Pages 471-489. Ed. M. G. Barbour and J. Major, Terrestrial vegetation of California. Wiley-Interscience, reprinted by the California Native Plant Society 1988, Sacramento, CA. Muller-Dombois, D., and H. Ellenberg. 1974. Aims and methods in Vegetation Ecology. Wiley, New York, 547pp. Munz, P. A. 1968. A California flora and supplement. University of California Press, Berkeley, CA. Munz, P.A., and D.D. Keck. 1949. California plant communities. Aliso 2: 87-105. Munz, P. A., and D. D. Keck. 1950. California plant communities; part two. Aliso 2: 199-202.

National Parks mapping website. http://biology.usgs.gov/npsveg/ NatureServe. 2003. NatureServe Explorer: An online encyclopedia of life [Web application]. Version 1.8. (http://www.natureserve.org/ explorer). NatureServe, Arlington, VA (Accessed August 2004). Nichols, J. D. 1974. Mapping of wildland fuel characteristics of the Santa Monica Mountains of Southern California. Fiscal rep., Cooperative Agreement 21-274, U.S. Forest Serv. Remote Sensing Res. Program, Univ. Calif., Berkeley. 26pp. Ornduff, R. 1974. California plant life. University of California Press, Berkeley, CA. Ornduff, R., P. Faber, and T. Keeler-Wolf. 2004. California plant life, 2nd edition, University of California Press, Berkeley, CA. Pacific Meridian Resources. 1994. California Hardwood Rangeland Monitoring Final Report. CDF-FRAP, 1920 20th St. Sacramento, CA. 85pp. Parker, 1., and W. Matayas. 1979. Calveg: A classification of Californian vegetation. Unpublished map and report. USDA, Forest Service, Pacific Southwest Region, Ecology Group, San Francisco, CA.

Parker, 1., and W.J. Matyas. 1981. CALVEG. Mosiac of Existing Vegetation of California. Regional Ecology Group. USDA Forest Service, San Francisco. Paysen, T. E. 1982. Vegetation classification-California. ed. C. E. Conrad and W. C. Oechel, Dynamics and management of Mediterranean-type ecosystems. General Technical Report PSW58. USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. Paysen, T. E., J. A. Derby, and C. E. Conrad. 1982. A vegetation classification system for use in California: its conceptual basis. General Technical Report PSW-63. USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. Paysen, T. E., J. A. Derby, H. Black, et al. 1980. A vegetation classification system applied to southern California. General Technical Report PSW-45. USDA, Forest Service, Pacific South-west Research Station, Berkeley, CA Peinado, M., F. Alcaraz, J. Delgadillo, et al. 1994. The coastal salt marshes of California and Baja California. Vegetation 110: 55-66. Peinado, M., J. 1. Aguirre, and J. Delgadillo. 1997. Phytosociological, bioclimatic and biogeographical classification of woody climax communities of western North America. J. Veg. Sci. 8: 505-528. Pfister, R. D., and S. F. Arno. 1980. Classifying forest habitat types based on potential climax vegetation. Forest Science 26: 52-70. Pillsbury, N., et. al. 1991. Mapping and GIS Database Development for California's Hardwood Resources. Available from: CDF-FRAP. 1920 20th St., Sacramento, CA, 62pp. Ponomarenko, S., and R. Alvo. 2001. Perspectives on Developing a Canadian Classification of Ecological Communities. Information Report ST-X-18E, Science Branch, Canadian Forest Service, Natural Resources Canada, Ottawa. Potter, D. A. 1998. Forested communities of the upper montane in the central and southern Sierra Nevada. USDA Forest Service, Pacific Southwest Region, General Technical Report PSW-GTR-169. Potter, D. A. 2003. Riparian community type classification for the west slope central and southern Sierra Nevada, California. DRAFT. Pacific Southwest Region. Berkeley, CA. Potter, D.A. 2005. Riparian plant community classification: west slope, central, and southern Sierra Nevada, California. R5-TP-022 USDA Forest Service Pacific Southwest Region, Vallejo, California. Rejmanek, M. 1997. Vegetation classification: shortcuts lead nowhere. Global Ecology and Biogeography Letters 6: 164-165.

Rivas-Martinez, S., D. Sanchez-Mata, and M. Costa. 1999a. North American boreal and western temperate forest vegetation. (Syntaxonomical synopsis of the potential natural plant communities of North America, 11). Itinera Geobot. 12: 5-316. Rivas-Martinez S., D. Sanchez-Mata, and M. Costa. 1999b. North American new phytosociological classes. Itinera Geobot. 13: 349-352. Rodriguez-Rojo, M.P., D. Sanchez-Mata, R. G. Gavilan, et al. 2001a. Typology and ecology of Californian serpentine annual grasslands. Journal of Vegetation Science 12: 687-698. Rodriguez-Rojo, M.P., D. Sanchez-Mata, S. Rivas Martinez, et al. 2001 b. Syntaxonomical approach for classification of the Californian serpentine annual grasslands. Lazaroa 22: 83-94. Rowlands, P. G. 1995. Vegetational attributes of the California Desert Conservation Area. ed. J. Latting and P. Rowlands. The California Desert: An introduction to natural resources and man's impact. June Latting Books, Riverside, CA. Ryherd, S. L., and C. E. Woodcock. 1990."The Use of Texture in Image Segmentation for the Definition of Forest Stand Boundaries." In 23rd International Symposium on Remote Sensing of Environment, Bangkok, Thailand, pp. 1209-1213, 18-25 April 1990. Sawyer, J., T. Keeler-Wolf, and J. Evens. 2006. A Manual of California Vegetation Second Edition. Unpublished Manuscript. Sawyer, J.O., and T. Keeler-Wolf. 1995. A Manual of California Vegetation. California Native Plant Society. Sacramento, CA. Schwind B, C. Curlis, and S. Daniel, 1999. Creating a consistent and standardized vegetation database for northwest forest plan monitoring in California. White paper with USDA Forest Service, Remote Sensing Laboratory, Sacramento, CA. USDA. Schwind, B., and H. Gordon. 2001. Calveg geobook: A comprehensive information package describing California's wildland vegetation, version 2. USDA Forest Service, Pacific Southwest Region, Remote Sensing Lab, Sacramento, CA. Unpublished CD ROM. Scott, J.M., T.H. Tear, and F. Davis. eds.1996 GAP analysis: A landscape approach to biodiversity planning. Bethesda MD. American society of Photogrammetry and Remote Sensing. Scott, J. M., F. Davis, B. Csuti, et al. 1994. Ecological guide to eastside pine plant associations: Northeastern California. San Francisco, CA. Stoms, D. M., F. W. Davis, and C. B. Cogan. 1992. Sensitivity of wildlife habitat models to uncertainties in GIS data. Photogram Metric Engineering and Remote Sensing 58: 843-850. Story, M., and R. G. Congalton. 1986. Accuracy assessment: a user's perspective. Photogrammetric Engineering and Remote Sensing 52: 397-399. Sudworth, G.B. 1908. Forest trees of the Pacific slope. 1967 reprint, Dover Press, New York. Taylor, R. 2004. A natural history of coastal sage scrub in southern California: regional floristic patterns and relations to physical geography, how it changes over time, and how well reserves represent its biodiversity. Ph.D. dissertation. University of California, Santa Barbara. 215pp. The Nature Conservancy and Environmental Systems Research Institute. 1994. USGS-NPS Vegetation Mapping Program: Field Methods for Vegetation Mapping. The Nature Conservancy. Arlington, VA (Web publication. http://biology.usgs.gov/npsveg/ index.html). The Nature Conservancy and Environmental Systems Research Institute. 1994a. Standardized national vegetation classification

VEGETATION CLASSIFICATION IN CALIFORNIA

41

system. Arlington, Virginia and Redlands, CA (Web publication. http://biology.usgs.gov/npsveg/classification/index.html). The Nature Conservancy. 1998. An environmentally-driven approach to vegetation mapping at Yosemite National Park, USGS-NPS Vegetation Mapping Program. (Web publication http://biologogy.usgs .gov/npsveg/yose/enviro.html) Thomas, K., T Keeler-Wolf, ]. Franklin, et al. 2004. Mojave Desert Ecosystem Program: Central Mojave Vegetation Database. USGS Western Ecological Research Center and Southwest Biological Science Center, Sacramento, CA. Thorne, J. H., TR. Kelsey, ]. Honig, and B. Morgan. The development of 70-year old Wieslander Vegetation Type Maps and an assessment of landscape change in the central Sierra Nevada (http://cain.nbiLorg/repository/WP(Yo20DR°/c)2010-16-06.doc). California Energy Commission, PIER Energy-Related Environmental Program CEC 500-2006-107 Thorne, J. H., ]. A. Kennedy, ]. F. Quinn, et al. 2004. A vegetation map of Napa county using the manual of California vegetation classification and its comparison to other digital vegetation maps. Madrono 51 (4): 343-363. Thorne, R. F. 1976. "The vascular plant communities of California." Pages 1-31. ed. ]. Latting, Plant communities of southern California. California Native Plant Society, Sacramento, CA. U.S. Forest Service. 1949. California Forest and Range Experiment Station. U.S. Forest Service. 1954. California Forest and Range Experiment Station. USDA Forest Service. 1981. CALVEG: A Classification of California Vegetation. Pacific Southwest Region, Regional Ecology Group, San Francisco CA. 168pp. USGS. 1997a. Field Methods for Vegetation Mapping (complete document available at following Web site: httpll: biology.usgs.gov/ npsveg/fieldmethods.html Vaghti, M., and T Keeler-Wolf. 2004. Suisun Marsh Vegetation Mapping Change Detection 2003. Unpublished report on file at Wildlife and Habitat Data Analysis Branch. VegBank Web site. http://vegbank.org/vegbank/index.jsp Vegetation Memorandum of Understanding. 2000. Memorandum of understanding for cooperative vegetation and habitat mapping and classification. Available at: http://www.cnps.org/cnps/vegetation/pdf/vegMOU. pdf.

42

VEGETATION CLASSIFICATION IN CALIFORNIA

Vogl, R.]. 1976. An Introduction to the Plant Communities of the Santa Ana and San Jacinto Mountains. Pages 77-98. ed.]. Latting, Plant communities of Southern California. California Native Plant Society. Berkeley, CA. Warbington, R. 2004. US Forest Service Remote Sensing lab views on the cost of mapping using manual and image segmentation approaches. Walker, R. E. 2000. Investigations in vegetation map rectifications and the remotely sensed detection and measurement of natural vegetation changes. Ph.D. thesis. Geography Department, University of California, Santa Barbara. Weaver, ]. E., and F. E. Clements. 1938. Plant ecology. McGraw-Hill, New York. Western Riverside County (DFG-CNPS 2004). Westhoff, V., and E. van der Maarel. 1978. The Braun-Blanquet approach. ed. R.H. Whittaker, Classification of plant communities. Pages 278-399. Junk, The Hague. Westoby, M., B. Walker, and I. Noy-Meir. 1989. Opportunistic management for rangelands not at equilibrium. Journal of Range Management 42: 266-274. Wieslander, A. E. 1935a. A vegetation type map of California. Madrofio 3: 140-144. Wieslander, A. E. 1935b. First steps of the forest survey in California. ]. For. 33: 877-884. Wieslander, A. E. 1961. California's vegetation maps: Recent advances in botany. Univ. Toronto Press. 4pp. Wieslander, A. E., and H. A. Jensen. 1946. Forest areas, timber volumes, and vegetation types in California. Release Number 4 with map. USDA, Forest Service, Pacific Southwest Research Station, Berkeley, CA. Wieslander, A. E., and R. E. Storie. 1952. The vegetation-soil survey in California and its use. Journal of Forestry 50: 521-526. Winthers, E., D. Fallon, ]. Haglund, et al. 2004. Terrestrial Ecological Unit Inventory Technical Guide. USDA Forest Service, Washington Office-Ecosystem Management Coordination Staff, 125pp. Woodward, R. 2004. Personal communication re: development of California State Parks Image Segmentation-based vegetation maps in Portola and Butano State Parks, San Mateo County. Zinke, P.]. 1950. The soil-vegetation survey as a means of classifying land for multiple-use forestry. Proceedings Fifth World Forestry Congress, Seattle, WA.

TWO

Climate, Paleoclimate, and Paleovegetation RICHARD A. MINNICH

INTRODUCTION MODERN CLIMATE

Temperature Winter Precipitation The North American Monsoon Potential Evapotranspiration and Runoff Climatic Variability in the Instrumental Record El Nino-Southern Oscillation Pacific Decadal Oscillation TERTIARY CLIMATE CHANGE AND DEVELOPMENT OF CALIFORNIAN VEGETATION

nied by perhaps the greatest regional species diversity in temperate North America. The vegetation has comparable richness in life forms that includes tall conifer forests, riparian deciduous hardwoods, broadleaved evergreen woodlands and savannas, shrublands of evergreen chaparral and drought-deciduous shrubs, and fields of annual and perennial forbs and grasses, all of which cover extensive areas in the state. The vegetation is an overprint of floras of different age and regional origins that are vestiges of Tertiary vegetation and climate history. This chapter first summarizes the modern climate of California for background to the reviews of vegetation in this volume. What follows is a summary of paleoclimate and paleobiogeography of the California flora since the late Cretaceous.

Cenoloic Global Cooling Plate Tectonics and Topography of the Western United States Paleovegetation Late Cretaceous to Early Eocene Middle Eocene and Oligocene Miocene to Early Pleistocene Late Pleistocene and Holocene

Late Pleistocene Early and Middle Holocene Mid-Holocene Late-Holocene Drought Little Ice Age SUMMARY OF TERTIARY-QUARTERNARY ClI MATE AN D VEG ETATION CHANG E AREAS FOR FUTURE RESEARCH

Introduction The extraordinary variety of climate in California-ranging from high rainfall regimes in the northwestern mountains to Death Valley, the driest place in North America, is accompa-

Modern Climate California's mediterranean climate of winter precipitation and protracted summer drought is an outcome of seasonal changes in global circulation. In Winter, strong latitudinal temperature gradients focus the polar-front jet stream and onshore flow of moist surface westerlies around surface low pressure in the Gulf of Alaska into northern California and the Pacific Northwest. In summer, reduced latitudinal temperature gradients weaken the jet stream that shifts poleward to western Canada. Gulf of Alaska low pressure is replaced by surface high pressure covering the northeast Pacific Ocean. A vital component of the mediterranean climate is the coastal marine layer, a steady-state feature associated with the cooling and moistening of the tropospheric boundary layer overlying the cold, upwelling California Current. It is capped by a strong thermal inversion that divides the layer from warm and dry subsiding air masses aloft. The marine layer virtually precludes convective precipitation of Pacific Ocean air masses from May to September. Summary climate data are given in Table 2.1.

43

TABLE 2.1

Climatological Data of Selected Stations in California Snowfall (cm)

Mean Temp te)

Mean Precipitation (cm) Region/Station

Lat./Long. (deg and min)

Alt.

(m)

Jul

Aug

Sep

Oct

Nov

Dec

Jan

Feb

Mar

Apr

May

Jun

Total

Days Mean >2.5cm

- - - - - - Frost 2 free Jan Jul days

Pacific Coast

S. F. Int'l AP

37 37

122 24

2

0.1

0.1

0.5

2.1

6.1

8.6

11.7

9.1

8.4

3.3

1.0

0.3

50.9

o o o

Monterey

36 36

121 54

116

0.2

0.3

0.7

2.4

6.0

7.5

10.9

8.4

8.2

3.9

1.2

0.5

50.2

0.1

Santa Maria

34 55

120 28

79

0.1

0.1

0.6

1.2

3.4

4.6

6.4

7.1

6.0

2.7

0.7

0.1

32.6

Los Angeles Int'l AP

33 56

118 24

30

0.1

0.2

0.4

0.8

3.8

4.3

7.0

6.6

5.1

2.0

0.4

0.2

30.9

San Diego Int'l AP

3244

11710

4

0.1

0.2

0.5

1.2

2.5

4.5

5.3

4.9

4.4

2.0

0.5

0.2

24.3

o o o

o o o o o o o o

Gasquet

41 51

123 58

117

0.9

2.6

5.8

15.4

37.9

40.4

37.7

30.7

32.6

16.3

9.6

3.6

233.5

o

o

Pit River

40 59

121 59

445

0.9

1.5

5.3

11.6

28.7

30.5

33.1

27.1

27.6

13.6

6.4

2.3

188.6

Shasta Dam

40 43

122 25

327

0.5

1.2

3.4

7.7

21.0

25.5

31.2

25.9

22.9

10.7

6.1

3.0

159.1

12.7

Orland

39 45

122 12

77

0.2

0.4

1.0

2.8

7.2

8.2

11.0

8.8

7.5

3.1

1.9

1.0

53.1

1.0

Sac'to Exec. AP

38 30

121 30

4

0.1

0.2

0.7

2.3

5.4

7.1

9.5

7.9

6.0

2.8

1.2

0.4

43.6

o

Fresno Air Terminal

3647

11943

109

0.0

0.1

0.5

1.3

3.0

3.8

4.9

4.9

4.9

2.5

0.9

0.4

27.2

0.2

Bakersfield

35 26

11903

149

0.0

0.1

0.3

0.7

1.5

1.9

2.7

2.9

3.0

1.7

0.5

0.2

15.5

0.2

Riverside

33 57

117 23

275

0.1

0.3

0.7

0.7

2.4

3.1

5.8

4.7

4.5

2.0

0.6

0.2

25.1

Eureka

4049

124 10

6

0.4

0.9

2.2

6.9

14.6

16.4

17.3

13.5

13.5

7.2

4.3

1.6

98.8

0.5

Fort Bragg

39 31

123 26

35

0.3

0.9

1.6

6.6

13.7

17.4

19.8

16.2

14.8

7.3

3.5

1.1

103.2

Fort Ross

38 31

123 15

33

0.3

0.7

1.3

6.3

13.1

15.6

20.9

15.5

13.5

6.7

2.4

1.1

97.3

8.8

13.9

335

8.7

13.9

293

9.5

13.9

9.4

17.0

365

10.9

15.5

365

10.6

17.1

272

13.3

20.6

365

13.4

20.8

365

o

7.4

27.5

7.2

25.8

273

7.5

24.1

321

7.7

27.5

303

8.8

28.7

287

o

o o o o o

12.1

25.7

257

o

5.8

23.4

170

6.9

24.6

225

4.5

21.3

Siskiyou/ Cascades

Cen. Val./ Inter. SCA

N. Sierra Nevada Placerville

38 42

120 49

563

0.2

0.2

1.5

5.5

11.6

16.1

18.1

17.2

14.7

7.8

3.9

1.2

97.3

7.6

Colfax

39 07

120 57

731

0.3

0.5

2.0

5.8

16.7

20.3

22.8

19.4

18.3

9.1

4.3

1.5

120.6

36.1

Nevada City

39 15

121 00

847

0.2

0.3

1.8

7.0

16.3

23.0

26.0

24.8

20.0

10.9

5.3

1.6

137.2

50.8

13

Strawberry Valley

39 34

121 06

1174

0.4

0.8

3.3

11.6

26.5

35.0

42.0

33.8

30.0

14.9

7.8

2.4

208.5

277.6

51 140

Lake Spaulding

39 19

120 38

1571

0.5

0.8

2.8

10.0

20.9

28.6

31.2

29.0

24.9

14.1

8.3

3.0

174.1

650.2

Tamaracka

38 36

119 56

2457

1.0

0.5

1.8

7.4

10.7

18.9

23.0

28.0

21.3

5.6

4.2

3.3

125.7

1098.8

Truckee RS

39 20

120 11

1834

0.9

1.2

1.9

4.2

10.0

13.5

16.2

12.6

10.7

5.5

3.5

1.8

82.0

527.5

Tahoe City

39 10

12008

1898

0.7

0.8

1.6

4.6

9.3

13.7

15.7

14.0

10.4

5.4

3.0

1.8

81.0

Lemon Cove

3623

11902

156

0.0

0.1

0.7

1.7

4.1

5.0

6.9

5.9

6.3

3.3

1.3

0.4

35.7

Three Rivers

3628

11852

346

0.2

0.2

1.5

2.8

6.9

7.8

11.9

10.7

11.8

4.6

2.0

0.8

1179

0.5

0.5

1.8

5.3

10.6

14.3

15.4

14.7

13.3

8.2

4.6

3.9

20.1

1.2

17.8

-3.8

13.4

101

139

-2.8

16.4

479.8

131

-1.8

16.1

o o

8.0

27.1

61.2

o o

8.5

27.8

2.1

91.3

165.6

33

3.5

21.6

180 121

42

S. Sierra Nevada

Hetch Hetchy

37 57

119 47

Grant Grove

36 44

118 58

2011

0.6

0.3

3.1

5.2

11.6

14.7

20.3

17.9

19.1

7.7

3.7

3.7

107.8

493.5

147

1.2

17.2

Lodgepole a

36 36

118 44

2051

1.1

0.8

3.7

5.0

11.3

16.0

25.7

23.0

19.1

7.8

3.2

1.8

118.5

644.8

167

-2.7

15.1

Ellery Lake a

37 56

119 14

2926

1.5

1.7

1.3

3.6

5.1

11.0

11.2

11.2

8.0

4.9

2.6

2.1

64.2

664.5

-4.8

12.7

33 14

116 46

823

0.8

1.3

1.3

2.0

6.1

8.1

13.1

11.9

12.3

5.1

1. 7

0.3

64.0

3.8

o

6.9

22.3

34 15

117 11

1586

0.4

0.9

2.3

3.8

10.7

14.1

22.3

20.2

17.3

8.0

3.4

0.5

103.9

123.6

14

2.9

20.5

34 18

11804

1740

0.2

0.5

1.9

2.8

9.7

11.3

19.3

19.6

16.0

7.0

2.1

0.5

90.9

68.6

6.6

23.6

279

Southern California Mountains Henshaw Dam Lake Arrowhead Mt. Wilson Idyllwild Big Bear Lake Damb Big Bear Lake Mt. San Jacinto C

33 45

11642

1637

1.8

2.1

2.2

2.4

6.0

8.7

13.1

11.3

10.9

5.0

2.0

0.4

65.9

100.1

4.5

20.1

34 15

11659

2073

1.1

1.9

1.9

3.5

7.7

13.4

17.9

18.1

16.3

6.1

2.5

0.4

91.4

313.1

-1.9

16.3

3415

11653

2070

1.9

2.5

1.4

1.8

5.5

7.5

11.0

10.6

8.7

3.4

1.3

0.4

56.0

161.5

0.9

17.5

33 48

11638

2566

1.4

4.1

3.3

1.7

9.4

12.0

12.0

9.2

7.4

2.7

1.4

0.2

64.8

0.0

15.9

41 30

120 33

1341

0.8

0.9

1.2

2.4

3.7

3.8

3.9

3.6

3.5

2.7

3.4

2.5

33.4

82.2

21

-1.6

18.8

39 58

120 05

1338

0.8

0.7

1.1

1.7

3.5

4.6

5.3

3.6

3.1

1.6

1.9

1.4

29.0

57.9

22

0.8

22.6

1.0

1.3

1.4

22.8

106.9

41

-3.4

16.4

-9.3

7.4

17

Modoc Plat./Mono L. Alturas Doyle Bridgeport White Mountain

38 15

119 14

1972

1.2

1.2

1.2

0.7

2.3

2.7

3.6

3.8

2.4

37 35

118 14

3801

2.7

2.8

2.2

2.6

3.1

6.6

6.3

4.3

5.5

4.8

5.0

2.1

48.0

402.6

68

(continued)

TABLE 2.1

(continued)

Snowfall (cm)

Mean Temp roC)

Mean Precipitation (cm) Region/Station

Lat./Long. (deg and rnin)

Alt. (nl)

Jul

Aug

Sep

Get

Nov

Dec

Jan

Feb

Mar

Apr

May

Jun

Total

Days Mean >2.5cm

- - - - - - Frost 2 free Jan Jul days

Mojave Desert Bishop AP

37 22

118 21

1250

0.4

0.3

0.5

0.5

1.3

2.1

2.8

2.4

1.3

0.8

0.7

0.4

13.5

20.5

5

2.9

24.8

Palmdale

33 35

118 06

791

0.1

0.5

0.5

0.7

1.8

3.5

3.9

3.9

3.3

1.2

0.3

0.1

19.8

3.8

7.3

27.3

Mojave

3503

11810

832

0.3

0.5

0.6

0.5

1.6

1.9

3.0

2.9

2.3

0.8

0.3

0.1

14.8

3.8

o o o o o

7.4

28.1

Daggett-Barstow AP

3451

11648

585

1.0

1.0

0.9

0.4

0.6

1.1

1.6

1.1

1.1

0.6

0.2

0.2

9.8

2.0

Needles

34 46

116 47

278

0.9

1.7

1.2

0.8

0.9

1.1

1.6

1.2

0.9

0.3

0.2

0.1

10.9

0.8

Death Valley

36 28

116 52

-59

0.3

0.4

0.5

0.3

0.5

0.4

0.7

1.2

0.9

0.3

0.2

0.1

5.8

o

Wildrose RS

36 16

117 11

1249

0.9

1.6

1.2

0.7

1.1

1.4

2.4

3.1

2.8

0.8

1.1

0.4

17.5

6.0

Mitchells Cavern

34.57

115 33

1325

2.0

3.8

2.1

1.6

1.8

2.4

3.2

3.9

3.6

1.4

0.7

0.3

25.8

8.6

o o

Mountain Pass

35 28

115 33

1441

2.7

3.2

1.6

1.2

1.6

1.5

2.5

2.1

2.3

1.1

0.8

0.6

21.2

23.3

6

Twenty-nine Palms

34 08

11602

602

1.5

1.8

1.2

0.6

0.6

1.0

1.3

0.9

0.9

0.3

0.3

0.0

10.4

2.3

Thermal AP

33 38

116 10

-34

0.5

0.7

0.7

0.4

0.7

0.8

1.4

1.2

0.9

0.2

0.2

0.0

7.6

o

Imperial

3251

11534

-19

0.3

0.7

0.7

0.6

0.5

0.9

1.1

0.8

0.7

0.2

0.1

0.0

6.6

o

Iron Mountain

34 08

115 08

281

0.7

1.0

0.7

0.9

0.6

1.1

1.4

0.8

1.0

0.4

0.2

0.1

8.9

o

Blythe

33 37

11436

81

0.7

1.7

1.0

0.7

0.5

1.1

1.2

1.0

0.9

0.5

0.1

0.1

9.5

Gold Rock Ranch

32 53

114 52

147

0.7

1.4

0.9

0.9

0.7

1.3

1.6

1.1

0.9

0.3

0.2

0.0

10.0

o o

o o o o o o

9.0

31.6

307

11.5

35.4

316

11.4

38.3

4.9

26.3

7.8

28.0

9.6

31.3

266

12.6

32.9

317

13.1

33.1

317

12.0

34.8

12.2

34.8

13.8

34.1

Mojave Desert Mtns

Sonoran Desert

290

NOTE: Western Regional Climate Center, Desert Research Institute. Station record 1948-2001. Web site: www.wrcc.drLedu/index.html aSource: V.S. Departn1ent of Commerce. Weather Bureau. Climates of the States, California. bDepartment of Commerce. Weather Bureau. Climatic summary of the United States, Supplement for 1931-1952. CClimatological Data, California. Monthly summaries. National Oceanic and Atmospheric Administration. Environmental Data and Information Service. National Climatic Center. Asheville, North Carolina.

Temperature

The position of the jet stream is coupled with broadscale latitudinal temperature gradients in California. In January nlean sea-level temperatures range from 9°C in northern California to 14°C in southern California. Atmospheric lapse rates result in temperatures decreasing to O°C at ca. 1,600 m in northern California and 2,200 m in southern California, with Illeans as low as -6°C along the crest of the Sierra Nevada. Many low-lying basins contain persistent temperature inversion layers resulting from radiational cooling and low insolation. In the Central Valley, where ground inversions are maintained by reflective ground fogs, Il1ean January temperature averages 8°C. Surface inversions result in means as low as -3°C in the high northeastern plateaus from Modoc to Lake Tahoe. From May to September, California is dominated by strong onshore flows from the Pacific anticyclone to thermal low pressure over the hot desert interior. Northwesterly gradient winds combined with sea breezes and anabatic circulations transport the marine layer air inland to the coastal ranges usually within 100 km of the coast. Farther inland, the nlarine layer dissipates from diabatic heating and mixing with warm air aloft (Glendening, Ulrickson, and Businger 1986). Mean July temperatures along the coast reflect local sea surface temperatures, ranging from 14°C-16°C in strong upwelling zones north of Point Conception to 18°C-22°C in the southern California bight. Temperatures increase to 24°C-28°C in the inland valleys of southern California and along the Central Valley. The deserts beyond the reach of the marine layer average 26°C-35°C, and 38°C in Death Valley. Temperatures in mountains reflect ambient lapse rates, decreasing to 14°C-18°C at 2,500 m and 10°C at 3,000 Ill. Mountain and desert temperatures and lapse rates above the marine layer are isoclinal with latitude across California from June to September. The frost-free period varies primarily with elevation and distance from the Pacific Ocean. Freezing teIl1peratures are rare along the Pacific coast south of San Francisco and infrequent in December and January as far north as the Oregon border. Mild freezes occur from November to February in the Central Valley, interior valleys of southern Cali-

fornia, and the Salton Sea trough. The frost-free period is less than 100 days above 1,500-2,000 m in the Modoc Plateau, Sierra Nevada, and mountains of southern California.

Winter Precipitation

Winter precipitation results primarily from cold fronts of extratropical cyclones and associated troughs of the jet stream ITIoving into the region from the North Pacific Ocean. Because the mean position of the jet lies from the Pacific Northwest to northern California, the frequency of storms decreases southward in the state. The deepening of the marine layer before the passage of cold fronts gives rise to weather conditions similar to frontal occlusions, with extensive cloud shields and long periods of steady precipitation in

stable air. Winds aloft are southwesterly because cold fronts precede the passage of the associated trough aloft. Low-level winds are south to southeasterly (Minnich 1984). In the postfrontal phase, steady precipitation is replaced by convective showers concentrated over high terrain, with the winds at the surface and aloft shifting to westerly. Clouds and showers dissipate with the onset of subsidence following the passage of troughs, with winds turning northwesterly. The interaction between prefrontal circulation and terrain results in strong gradients in mean annual precipitation throughout California. Because storm air Illasses are stable, the variation in local precipitation is Illore influenced by mechanical (physiographic) lift over ITIountain barriers rather than from thermal convection. PhysiographiC lift is most intense on the south- to southwest-facing escarpments that lie at right angles to storm winds. Amounts decrease downwind to inland Illountain ranges-regardless of altitude-due to depletion of storm air mass moisture and descending airflow in rain shadows. The average annual precipitation (AAP) along the northern California coast varies from 50 cm at San Francisco to 100 cm north of Ft. Ross, with locally higher aIllounts where mountains skirt the coastline. The coastal Illountains north of Eureka and near Cape Mendocino receive >250 CIll. Totals in the North Coast Ranges decrease downwind to 150-200 cm in the Salmon and Siskiyou Ranges, and decrease southward with the declining general altitude of the mountains to 100 cm near Santa Rosa. Rain shadows from the North Coast Ranges decrease AAPs to 35-60 cm in the Sacramento Valley. However, low-level convergence of southerly prefrontal air masses against the narrowing rift between the Sierra Nevada and North Coast Range increases AAP to 60-80 cm north of Red Bluff. Orographic lift along the undissected western slope of the northern Sierra Nevada increases AAP from 60 cm along the lower foothills to 150-200 cm at the crest. To the east, rain shadows result in AAPs of 20-60 cm in the Modoc Plateau, and 60-80 cm in the Lake Tahoe Basin. In the South Coast Ranges, the AAP is 100-150 cm on the steep coastal escarpments of the "lead" Santa Cruz and Santa Lucia Mountains but amounts in the "downwind" Diablo Ranges seldoITI exceed 50 cm. Intervening basins including Salinas Valley receive 30-40 CIll. Rain shadows extending from the South Coast Ranges into the San Joaquin Valley produce an AAP of 30 cm near the Sacramento delta, lowering to 15 cm near Bakersfield and the Carrizo Plain. Amounts then increase with orographic lift to 30-50 cm in the Sierra Nevada foothills. The topographic complexity of the southern Sierra Nevada results in large variability of AAP along the west slope. Steep southwestern exposures have AAP of 100-150 cm, as at Yosemite, the upper San Joaquin drainage, Kaiser Ridge, and £IOIll Sequoia National Park to the Great Western Divide and Greenhorn Mountains. Leeward slopes on the coastal front receive 50-100 cm, including the upper Tuolumne River, Mono Creek Basin, the upper Kings River, and the Kern River

CLIlvfATE, PALEOCLIMATE, AND PALEOVEGETATION

47

plateau northward to Mt. Whitney. The AAP seldom exceeds 50 cm south of the Greenhorn Mountains due to low altitude of these ranges and their leeward position to the western Transverse and South Coast Ranges. In southern California, precipitation is highest on the steep southern escarpments of the Transverse Ranges (AAP, 80-110 cm). Amounts decrease to 60-80 cm in the "downwind" San Rafael Mountains, Pine Mountain Ridge, and leeward slopes of the San Gabriel and San Bernardino Mountains. Farther inland, the relatively high ranges that include Mt. Pinos, San Emigdio, Tehachapi and Liebre Mountains, and drainages north and east of Big Bear Basin receive 35-60 cm. The AAP is only 40-60 cm on the coastal slopes of the Peninsular Ranges because storm winds frequently parallel the escarpment. Amounts reach 80-100 cm on local southern escarpments of the Santa Ana Mountains, Palomar Mountain, and Cuyamaca Peak. The high San Jacinto Mountains receive only 40-70 cm and the Santa Rosa Mountains only 40-50 cm due to their leeward position to the Santa Ana and Palomar Mountains. Amounts in the southern California plains vary from 25-35 cm at the coast to 40-50 cm at the base of the mountains. The AAP in the southeastern deserts and Owens Valley is mostly 10-15 cm, with totals of 6-10 cm in the Salton Sea trough and 5.8 cm in Death Valley. Amounts reach 30 cm in the Panamint and Inyo Mountains, and the higher ranges in the northeast Mojave Desert. The White Mountains above 3,000 m receive 50 cm. The peak of the winter precipitation season shifts from December/]anuary in northern California to January/February in the south. This trend reflects a gradual equatorward shift of the jet stream due to the cooling of the North Pacific SSTs through the winter. The precipitation maximum reverts to December/January in the southeastern Deserts when rare cyclones or cutoff lows with southerly wind trajectories advect moisture from tropical Pacific and along the Gulf of

1,750 m, 750/0 at 2,750 m, and 1000/0 at 3,000 m. Average snow lines are about 200-400 m lower in the Sierra Nevada (Barbour et al. 1991), and ca. 400-600 m lower in northern California. The water content of solid precipitation depends on both the S/AAP ratio and on the total AAP. In general, frozen precipitation exceeds 50 cm above 2,300 m in southern California, 1,900 m in the central Sierra Nevada, and 1,400 m at Mt. Lassen. Amounts exceed 100 cm above 3,000 m in southern California, 2,500 m in the Sierra Nevada, and 2,000 m in northern California. Average snow lines decline through the rainy season. In southern California, average snow lines decrease from 2,700 m in November to 2,300 m in February and 1,700 m in March and April (Minnich 1986). Seasonal trends have not been documented for northern California. Late fall and winter storms advect moisture from the subtropical Pacific, whereas storms in March and April are strongly influenced by the marine layer overlying the ever-cooling California current. Average snow levels in California tend to rise with increasing total annual precipitation due largely to the advection of moist subtropical storm air masses during El Nifio events. In southern California, average annual snow lines vary from 2,000 m in years with 700/0 normal precipitation to 2,400 m in years with 1400/0 of normal. Extraordinary snow accumulations during very wet years are frequently limited to the highest elevations, with middle elevations producing storm runoff. The average snowline during the southern California floods of January 1969 was 2,700-3,000 m. The snowline during the 1998 New Years flood at Yosemite was >2,500 m. Snowmelt is dependent primarily on solar and infrared radiant loading (Miller 1981). In winter, melt rates exceed snow accumulation rates on south-facing exposures lying close to right angles to the sun below ca. 2,000 m in southern California (Minnich 1984) and ca. 1,300 to 1,700 m in the Sierra Nevada and North Coast Ranges. With increasing

California. Precipitation declines rapidly throughout the

elevation, snowmelt is less sensitive to slope aspect because

state in April due to the weakening of the jet stream. Weak disturbances extend the rainy season into May and June in the northern Sierras, the Modoc Plateau and other interior basins south to the White Mountains when relatively warm Pacific air masses compared to winter reduce the Sierra Nevada rain shadow, resulting in convective showers on leeward slopes under high sun. Extratropical precipitation virtually ceases throughout the state by late June when the jet stream assumes its summer position near the U.S.-Canadian border. An accumulating winter snowpack melt serves to increase dry season soil moisture in high mountain watersheds. The ratio of the water content of frozen precipitation to the average annual precipitation (S/AAP) increases uniformly with altitude reflecting atmospheric lapse rates (Minnich

the upslope snowpack retreat is phased with ever-higher sun angles in spring and summer. The North American Monsoon

From July to September, the western margin of the North American monsoon, a deep layer of moist, unstable tropical air periodically causes afternoon thunderstorms in California, especially in the eastern mountains and deserts. The monsoon moves into the region around a mid-tropospheriC anticyclone sustained by intense convective heating of high elevation land surfaces of the southwestern United States and Mexican plateau (Hales 1974). Tropical moisture of the monsoon arrives from the equatorial eastern Pacific Ocean by way of the Gulf of California. When the anticyclone cen-

1986). The lower limit of reliable snowfall of 1,000-1,200 m

ter lies over northern Mexico, monsoon moisture is steered

in California approximates the moist adiabatic lapse rate from mean sea-surface temperatures to storm freezing lines. In southern California, the S/AAP ratios increase to 250/0 at

northeastward from the Gulf of California into northwestern Mexico and Arizona. Dry southwesterly flow over California results in clear skies, although deep troughs in the westerlies

48

CLIMATE, PALEOCLIMATE, AND PALEOVEGETATION

produce infrequent thunderstorms that source Pacific air masses in the Cascades/Siskiyou Mountains and far northern Sierra Nevada. When the anticyclone moves northward into the southwestern United States, southeasterly winds aloft transport monsoon moisture into southwestern California and the Sierra Nevada. Moisture arrives at upper levels (3-4 kill) as convective debris from thunderstorms or mesoscale convective systems over the Sierra Madre Occidental of northwestern Mexico and Gulf of California. Below 2 km, ill0ist air masses (derived from convective outflows of thunderstorms over Mexico) surge into the Salton Sea trough, Colorado River Valley, and occasionally as far north as Owens Valley (Hales 1974; Stensrud, Gall, and Nordquist 1997). Moisture surges also result from the lift of the trade wind layer overlying the Gulf of California by passing east Pacific tropical cyclones. Convection is most frequent in the mountains exposed to low-level moist air masses from the Gulf of California, primarily east of a line from the Peninsular Ranges and eastern San Bernardino Mountains to the eastern escarpment of the Sierra Nevada northward to Lake Tahoe. Total July to September precipitation averages 5-10 cm at most with peak amounts along the crest of the southern California Peninsular Ranges, eastern San Bernardino Mountains, and the Sierra Nevada divide from Mt. Whitney to Lake Tahoe. Amounts decrease westward to 100 cm in the northwest coast and northern Sierra, 40-100 cm in the Santa Cruz and Santa Lucia Mountains, central Sierra Nevada, and southern California Coastal Ranges. During summer PET exceeds precipitation over the entire state. Summer thundershowers have limited effect on soil moisture recharge. Lysimeter data for the Sierra San Pedro Martir in Baja California and the San Jacinto Mountains show that summer rain is countered by high summer ET, with limited wetting of the root zone (Franco-Vizcaino et a1. 2002). Short-term rainfall intensities are higher than in winter, increasing the runoff component of the water budget. In mediterranean climate, the available soil water is dependent on recharge of the previous winter because plant growth and transpiration demand are out of phase with seasonal precipitation distribution. In water surplus watersheds (precipitation > ET) soils are saturated virtually every rainy season. Most precipitation variability is expressed in runoff, with surpluses unavailable for ecosystem use. In waterdeficit watersheds (ET > precipitation), the relationship between precipitation and ET becomes stronger along a decreasing precipitation gradient (Franco-Vizcaino et a1. 2002). Hence, interannual variation in fire hazard is more sensitive to precipitation in water deficit watersheds than water surplus watersheds because the vegetation utilizes most precipitation. Plant water supply in water deficit watersheds also depends on soil field capaCities, which vary greatly in mountainous terrain. Winter rains may saturate steep slopes with low field capacities (e.g., serpentinite > schist> granite. Organic matter preservation in soils is enhanced by formation of day-humus and calcium-humus complexes. Soils formed on gabbro have

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

73

140 120

E ~

..c::

.,

'5.

"0

'0

lithosequence of soils formed in northern California on acid igneous granite (A-I), basic igneous gabbro (B-1), serpentine (Serp), and metasedimentary (schist) bedrock parent materials. BI and schist are highly weatherable parent materials and show greater A horizon thickness, clay content, and organic carbon. A-I is highly susceptible to physical weathering, but chemical weathering is slow; thus soils are deep and clay content is low.

Cl)

FIGURE 3.2 A

20

70 r-

-

-

-

r-

r-

E

100 r-

15

~

~ 50

0. .,

U 40

..c::

80

r 40 r

60

"0 t:

>,

'"

10

E :::l E 30

0

60

.~

r-

i5

r-

'x

::E'"

..c::


0), soluble components (e.g., salts, carbonates) will be leached from the soil profile and colloidal materials (e.g., clays) will be translocated downward. In contrast, in arid regions (effective precipitation = 0), soluble elements will acculnulate in the soil profile and translocation of colloidal materials is very slow. In arid climates there may be a net upward movement of water in the soil profile, due to high evapotranspiration rates, which results in the upward movement of soluble materials such as salts. Temperature and moisture also have a strong influence on the type, amount, and net primary productivity of vegetation communities. The timing of precipitation patterns also plays an important role in soil development. Soils that receive precipitation during the growing season with very little precipitation during the winter months (ustic moisture regime) often display evidence of limited deep percolation in the form of an accumulation of calcium carbonate or salts in the subsoil. This occurs because a majority of the soil moisture reservoir is lost from the system through ET, leaving little water for deep percolation. A soil receiving the same amount of precipitation during the winter months with a dry growing season (xeric moisture regime) will have distinctly different soil properties. In this scenario, much of the pore space is filled with water when ET is low; thus a surplus of water is available for deep percolation and the leaching of soluble constituents beyond the root zone.

Soil development along a climosequence driven by an elevation gradient on the western slope of the central Sierra Nevada illustrates the effects of climate on soil properties (Dahlgren et al. 1997; Fig. 3.3). The transect of seven soils formed on granitic bedrock spans elevations from 198 to 2,865 m with mean annual temperature and precipitation differences of 13°C (3.9-16.7) and 94 cm (33-127), respectively. Snow is the primary form of precipitation above 1,600 m, rainfall dominates the lower elevations. Soil pH decreases by about two units with rising elevation, due to greater leaching associated with increasing precipitation. Organic carbon shows a strong increase with increasing elevation to about 1,000 m, after which organic carbon pools remain more constant (12-15 kg/m 2). Organic carbon accumulation is reflected in the thicker A horizons found above 1,000 m. Peak biomass production and litterfall occur between 1,000 and 2,000 m, but colder soil temperatures there preserve a greater fraction of soil organic matter. Soil depth and B horizon thickness reach a maximum at midelevations having intermediate levels of precipitation and temperature. Sufficient clay translocation to form Bt horizons (clay-rich horizons) is apparent only below 1,600 m. The degree of chemical weathering, based on clay production, increases in a near linear fashion to about 1,600 m and displays a pronounced decrease at higher elevations. We interpret this pattern in rates of chemical weathering as reflecting

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

7S

moisture-limited weathering at low elevations and temperature-limited weathering at high elevations. The explanation for the pronounced decrease in clay production occurring at > 1,600 m is not known; however, it coincides with the approximate elevation of the present-day average winter snow line, and it is only 200 m below the average freezing elevation during winter storms, which in turn corresponds to a major vegetation ecotone between lower and upper montane zones (Barbour et a1. 1991). This pattern for chemical weathering based on clay production is consistent for all parent materials examined in the western Sierra Nevada (Fig. 3.3).

RELIEF

The relief or topography factor of soil formation refers to the shape of the landscape (slope angle and slope length; convex vs. concave) and which direction the slope faces (aspect). Slope characteristics may have a strong influence on soil moisture and hydrologic flowpaths. On steep slopes, water may run off along the soil surface, creating the potential for soil erosion, or move laterally down slope, translocating colloidal and soluble constituents. Surface runoff results in dryer soil conditions at upland slope positions, whereas surface runoff and subsurface lateral flow to down-slope positions result in wetter conditions and even poor drainage (saturation leading to anoxic conditions) in lower slope positions. Surface runoff may also transport topsoil from the hillslope to lower slope positions, producing thin soils on steep slopes and deeper soils at the base of the slope. Slope and aspect also have a strong influence on soil temperature that further affects soil moisture and vegetation characteristics. In the northern hemisphere, south-facing slopes receive more solar radiation resulting in warmer soil temperature and greater evapotranspiration leading to less available moisture. At the oak woodland-grassland ecotone in California, it is common to see grasslands occupying south-facing slopes and oak woodlands on adjacent northfacing slopes. The shape of a slope (convex vs. concave) is another factor that can strongly affect soil moisture conditions. Convex surfaces lead to divergent water flow and are drier, while concave surfaces lead to convergent water flow and are wetter. These differences in soil moisture can have important effects on vegetation type and productivity. A hillslope sequence (toposequence) from Santa Barbara County illustrates the role of topography on soil development (Gessler et al., 2000; Fig. 3.4). Soil erosion and deposition, coupled with redistribution of water along a hillslope, are the most influential topographic factors altering soil properties. The soil profile at the summit position occupies a stable landscape position experiencing neither erosion nor deposition. As a result, soil developmental processes, such as clay translocation, are well expressed at this position, but soils in the backslope and toeslope positions show only weak profile development as erosion and deposition continually reset the soil-forming clock. Soil in the backslope position

76

experiences long-term erosion leading to shallow soils over bedrock, sharply in contrast with deep soil formed from longterm deposition in the toeslope position. In the toeslope position, A horizons are particularly thick (80 cm), because organic-rich materials are deposited from erosion along the backslope. The continuous deposition of materials masks clay translocation; because the degree of deposition is greater than the rate of clay translocation, a distinct clay-enriched B horizon (Bt) is absent in the toeslope position. Profile accumulation of organic carbon closely follows A horizon thickness, which is strongly regulated by erosion and deposition. Soil pH is elevated by nearly one unit in the toeslope position reflecting the downslope leaching of soluble base cations, from upslope positions. Net primary productivity by the annual grasses that dominate these soils was highest for the toeslope position, slightly less for the summit position, and much lower for the backslope position. Furthermore, the toeslope position has the most available water due to its thick soil profile and inputs of water from upslope positions, whereas the thin soils of the backslope provide little water storage and surface runoff results in redistribution of water to lower slope positions (Fig. 3.4).

ORGANISMS

The organisms living on and in soils (vegetation, animals, microorganisms, and humans) comprise an ecosystem. Terrestrial vegetation has been considered "soil engineers" in that their regulation of organic matter quantity and quality coupled with nutrient cycling processes can greatly affect soil properties. Biota play a major role in both physical and chemical weathering processes. In particular, tree roots are able to apply tremendous force to break apart rocks exposing new surface area and porosity to enhance chemical weathering processes. Upon decomposition of organic matter, organic acids and chelates (an organic compound that forms strong bonds with metals) may be produced which increase chemical weathering reactions. As only one example, carbonic acid is produced from CO 2 released from root and microbial respiration. Roots are also notoriously "leaky" contributing organic acids, sugars, amino acids, and complex organic molecules into the rhizosphere. It is estimated that biota increase chemical weathering from 3 to 10 times relative to abiotic processes alone (Drever 1994). Plants and animals mix the soil by burrowing and root growth. These activities create biopores in the soil that can greatly affect the transport of water and gases through the soil profile. Biota promote soil structure development through addition of organic matter and compression of particles. For example, earthworms ingest soil materials and produce soil pellets that greatly affect soil porosity, aeration, water movement, and water-holding capacity (Hole 1981). Humans influence soil formation through their impact on the natural vegetation, such as agricultural practices and urban development. Removal of vegetation and compaction by heavy machinery decrease the rate of water infiltration into the soil, thereby increasing surface runoff and erosion.

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

Summit

(cm) 15 "'''" o~'

A

50

100

\~w

I- Y I-

-

0

50 AB

R

/

Bl2

f J

R

I I/

/

/

/

/

/

/

/

100

150

.I

200

~

'"'""c

-'" 0

:c

I-y l-

Y

iI

250

Bwl

l-

I- Bw2

Y

Y

II-

v

E

Al A2

Bw

50

BI!

f

""",

Toeslope

Backslopc

0

Y

I-

Bw3

Y

y

400

7.0

300

6.5

200

6.0

100

5.5

:t

0-

I-

0

5.0 Summit

Backslopc

Tocslope

Summit

Backslope

Toeslope

600

20 ME

Co

15 400

~

U

i=:"

FIG U RE 3.4 Atoposequence of soils formed from a hillslope catena in Santa Barbara County, California. The summit position is the most stable; thus additions and losses are minimal and translocations and transformations are maximized. The backslope position is an unstable surface where losses through erosion are high, resulting in little opportunity for other pedogenic processes to occur. As a result, this soil is shallow and shows slight evidence of pedogenesis. The toeslope position illustrates the effect of rapid accumulation of material from upslope. The frequent additions mask transformations and translocations. As a result, the soil is deep, carbon content is high, and net primary productivity is high.

ME

~

""-

10

z

£

200

Summit

Backslope

Toeslope

Swnmit

Backslope

Tocslol>c

Atmospheric deposition (e.g., acids, nitrogen, and sulfur) and agricultural management practices (e.g., irrigation, fertilizer, lime, pesticides, tillage, and fire) may have tremendous effects on soil properties. A biosequence of soils consisting of Douglas fir, grassland, Douglas fir invading grassland and oak woodland all growing on the same parent material with the same climate and topography was described by Popenoe et al. (1992; Fig. 3.5). The A horizon thickness was greatest for the grassland and Douglas fir-invaded grassland, intermediate for oak woodland, and very thin in the Douglas fir forest. Large inputs of organic matter from grass roots and the mixing of materials by soil fauna are important for forming the thick grassland A horizon. The 50 years since invasion of Douglas fir into grassland have not been long enough to result in any appreciable change in A horizon thickness. Soil pH and exchangeable potassium concentrations are highest for the Douglas fir

and oak woodland soils because of more rapid cycling of base cations. Organic C and N concentrations are highest in the A horizons of the grassland and invaded-grassland soils, reflecting greater organic matter enrichment by herbaceous vegetation. The C/N ratio is highest for Douglas fir, intermediate for oak and lowest for grasslands and invaded grasslands due to litter quality differences. Conifer vegetation typically has the highest concentrations of C-rich, difficultto-decompose organic materials (lignin, tannins, and waxes) compared to woody deciduous and hebaceous vegetation.

TIME

The extent of soil-forming reactions depends on the time period over which they have operated. The greater the time, the more developed a soil will become due to increased time for soil-forming processes to take place (weathering, humus

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

77

Mixed coniferous forest

Conifer invaded meadow

Meadow

Oak woodland

0.8

70 r--

E

60



.., " £ 0"

50

-'" u

40

~

N

«

-

6.0

-

on

S o

5.5

-

:I:

C-

30

5.0 .---

~ :>G ~

.---

.D

"

10

,i

.---

I·'

Ul

4.0

Gr Inv-G OF Oak

-

~ 0.2

,.

4.5

n

0

0.4

'""

"?

, ,

0.6

..,'"00

"

20

r--

E

r--

.§ ..c:

-

';

0.0

Gr Inv-G OF

Oak

70

Gr Inv-G OF Oak 30

-

4 .--60

on ~ ~

-

-

.--.---

50

-

U

.S!

"'"00

on ~ ".., 00

g

40

25

3

.---

5 20

2

.---

.---

Z

(5

30

r--

15 .---

i'

I~


30(Y(l slope on the older landforms. Thus, SOll1e soils on older landfonns that have been highly dissected by erosion are weakly developed soil profiles because rates of soil erosion exceed rates of soil forn1ation. Soils forn1ed in alluvial environments have parent materials with a wide range of particle-size distribution, depending on the velocity of the water carrying thel11. Fine-textured alluvium (silts and clays) is generally deposited in basin positions occurring behind the natural levees of rivers. Thus, basin soils are generally fine textured (clay loams, silty clay loa ms, and clays) and display poor drainage conditions (e.g., Sacramento clay and Merced clay-Haplaquolls). The fine textures and poor drainage lead to accumulation of organic matter giving these soils a very dark color. Soils forn1ed on young alluvial fans, in contrast, display minimal soil development and often retain evidence of vertical stratification from multiple deposits during past flood events. Soils originating from igneous alluvium from the Sierra Nevada often have sandy textures, whereas those originating from sedimentary alluvium from the Coast Ranges often have loamy textures. The sandy nature of the Sierra Nevada alluvium in the southeastern portion of the San Joaquin Valley has resulted in relatively large areas of wind-modified soils (Delhi dune soils-Xeropsamments). The young alluvial soils in the Central Valley display minimal soil development (Entisols and Inceptisols) with small accumulations of organic matter in topsoil horizons and development of weak soil structure throughout. In spite of their youth, these soils are among the most productive agricultural soils, as they have near-neutral pH and no indurated layers (e.g., clay pans) to impede root growth and water movement. Their natural riparian vegetation has typically been replaced by agricultural crops. With increasing soil development over time, clay translocation leads to development of argillic horizons (clay-rich B horizons). Argillic horizons often require a minimum of 30,000-50,000 years to become distinguishable in Central Valley alluvium (Harden 1987). In the absence of erosion, argillic horizons become thicker (>50 cm) and accumulate greater concentrations of clay (>50(Y9.5) and sodium content promote clay translocation and fonnation of argillic horizons and a silica-calcium carbonate cemented hardpan (Fresno-Duric Natrixeralf). These saline-sodic soils once supported only salt-tolerant vegetation due to vegetation stress resulting from direct toxicities (e.g., sodium and boron), micronutrient deficiencies, very high pH values, and physiological drought. Many saline-sodic soils have been reclaimed by ripping the hardpan, displacing the sodium with calcium (gypsum [CaS0 4 2Ff zO] additions), leaching of salts with irrigation, and sometimes artificial drainage (drain tiles) to provide for removal of saline drainage waters. Saline-sodic soils are also comlnon in gently undulating topography fonned in alluvium from the Coast Ranges. The Coast Ranges are uplifted oceanic sediments and therefore

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

contain lnuch higher concentrations of sodium and salts than other parent l1laterials. Coast Range alluvium is often finer textured, which results in slow vertical drainage and higher leaching requirements. In areas with undulating topography, salts accurnulate in depressions during the winter months and are drawn into the slightly higher zone surrounding the depression by evapotranspiration (Whittig and janitzky 1963). As a result, soils in the basin rim position, as well as SOlne of those in depressions, may accumulate appreciable alnounts of sodium and salts. Winter rainfall in the Central Valley is generally effective in leaching salts and carbonates in well drained soils, especially in the northern portion of the Central Valley. As rainfall decreases in the San joaquin Valley, there is no longer sufficient leaching to remove calcium carbonate from the soil profile. The depth of calcium carbonate accumulation increases from about 2 m below the surface in the northern San joaquin Valley to near the surface in the extreme southern part of the valley. The low rainfall in the southern and western portions of the San joaquin Valley «20 cln) results in an aridic soil moisture regime and classification of many of these soils in the Aridisol soil order. Vegetation ecologists have recognized for sorne time that the climatic aridity and unique soils in the southern San joaquin Valley never supported grassland, but instead was dominated by an Atriplex desert scrub. The Delta region, located at the confluence of the Sacralnento and San joaquin Rivers, forms the lowest part of California's Central Valley. The Delta is composed ahnost entirely of peat and is the second largest contiguous body (about 100,000 ha) of peat outside the Florida everglades. The peat was deposited following the most recent glaciation, when melting continental glaciers contributed to a sealevel rise (---100 m) and the flooding of the previous channels of the Sacramento River. As sealevel slowly rose, a dense growth of tule provided detritus that accumulated to depths of 9 to 15 m. The construction of dams and diversion of water to agriculture led to the disappearance of the tule 111arsh. Beginning in the 1850s, reclamation projects drained tracts in the Delta by constructing levees and purnping waters to allow agricultural production on the rich organic soils. Today, much of the land surface in the Sacralnento-San joaquin Delta has subsided as much as 10 m below sealevel. Current subsidence prilnarily is the result of lnicrobial oxidation of the organic soils (commonly ranging frOl1l 3 to 6 crn yr I), enhanced by the introduction of oxygen into otherwise anoxic, water-saturated soils. The organic soils (Histosols) are used extensively for crops, such pears, asparagus, turf grass, and corn. A series of poorly drained soils are present on deltas, basins, basin rims, and alluvial fans along the edge of the Delta. Vertisols, clay-rich Mollisols, Inceptisols, and Histosols are common in this region. Many of these soils are saline, have aquic soil moisture regimes, and are strongly acid (pH SO(1() serpentine is called serpentinite. As exposure to water under pressure converts more and more magnesium silicate minerals to serpentine, we can say that the rock becomes more and more "serpentinized." Serpentine is only one of many minerals that are rich in magnesium and iron (hence called "mafic" or "ultramafic"). Ultramafic rock is an umbrella category that includes (besides serpentinite) peridotite, ophiolite, gabbro, dunite, pyroxenite, and hornblendite, among others. Important minerals and elements in these rocks are olivine, chrome, nickel, antimony, cobalt, mercury, and gold (Bates and .lackson 1984; Coleman 1977; Coleman and .love 1992). All of these ultramafic rock types bring similar stresses to plants, and all are in California, so in this chapter we more often use the broader term "ultramafic" instead of the narrower term "serpentine." Serpentinite is unique and relatively easy to determine in the field because of its shiny or silky luster, slightly soapy feel, conchoidal fractures, and greenish surface. California has >3,200 km 2 of ultramafic outcrops. The outcrops are typically associated with fault lines, which is why they occur in strips trending southeast to northwest, parallel to nearby fault lines (Fig. 3.12). Nearly all of this ultramafic material was metamorphosed while being emplaced in, or rising up through, the continental crust.

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

93

120°

liS·

FIG URE 3.12 Biogeographical map of

40·

California, with the distribution of major ultramafic outcrops. The biogeographical limits are adapted from Rivas-Martinez (1977) and Rivas-Martinez et al. (1999a). Lithological information comes from the California Division of Mines and Geology (Kruckeberg 1984). I = Californian region, 1I = Great Basin region (NE California) and Warm Desert region (SE California), 1lI = Pacific Northwest region. l.l.a = Klamath sector, l.l.b = Sierra Nevada sector, l.l.c. = North Coast Ranges sector, l.l.d = GreatValley sector; l.2a = Transverse Ranges sector, l.2.b = South Coast and Channel Islands sector. The bar (above 3soN latitude line) represents approximately 67 km.

35°

Soils on ultramafics are usually shallow and skeletal, with little profile development. Often, the steepness of the slopes and sparseness of the vegetation allow for continual erosion. Unweathered rock particles and rock fragments are very common in the soil profile. The most extensive soils within California's ultramafic areas are Lithic Argixerolls, Lithic Haploxerolls, and Pachic Argixerolls. Ultramafic soils have exceptionally low concentrations of calcium, an essential nutrient for plants. Its low availability decreases the ability of plants to grow (Kruckeberg 1951, 1954; Mason 1946a, 1946b; Walker 1948, 1954; Walker, Walker, and Ashworth 1955; Whittaker 1954). High magnesium levels and the presence of several heavy metals add additional stresses. In nonultramafic soils, the Ca:Mg ratio is 3, but in mafic soils it is 0.3. Metals reach high values in the tissues of certain plants growing on ultramafics (Bargagli 1998). Such plants are called "hyperaccumulators," and they can be used as bioindicators (Brooks 1998b). The ecological value of hyperaccumulation may be deterrence of herbivores and avoidance of competition. Plants on ultramafics possess heightened drought resistance (Walker 1954), as well as tolerance for heavy metals (Antonovics, Bradshaw, and Turner 1971).

94

To summarize, uItramafic soils impose the following stresses on plants: imbalance of calcium and magnesium, magnesium toxicity, low availability of molybdenum, toxic levels of heavy metals, sometimes high alkalinity, low concentrations of such essential nutrients as nitrogen and phosphorus, and low soil water storage capacity (Kruckeberg 2002). The vegetation on soil derived from serpentinite has been classically called "serpentine vegetation." Although this term still persists, we will use the broader term "ultramafic vegetation" (Brooks 1998a). Ultramafic vegetation is unique in physiognomy as well as in floristic composition. Very often, a dramatic landscape shift occurs across abrupt discontinuities between ultramafics and country rock. For example, regional stands of dense conifer forests are replaced by stunted and open stands of other conifers, by chaparral or even by barrens on which woody vegetation is absent. Kruckeberg (1969) described the folloWing attributes of ultramafic vegetation: (a) dwarfing, (b) reduced cover, (c) lowered abundance of some species, (d) altitudinal extensions of regionally widespread species, (e) reduced species richness, (f) increased patchiness of species distributions at a local scale, and (g) the presence of ultramafic

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

endemics. Many publications have focused on California's ultramafic endemics (Koenigs, Williams, and ]ones 1982; Koenigs, Williams, ]ones, and Wallace 1982; Kruckeberg 1954, 1969, 1992a, 1992b, 1999,2002; Mason 1946a, 1946b; Proctor and Woode1l1975; Raven 1964; Stebbins 1942; Stebbins and Major 1965; Walker 1954; Walker, Walker, and Ashworth 1955; Whittaker 1960), but few have had a vegetation-scale focus.

tion, and they occur on steep sites where the parent rock is serpentinized. They support an open herbaceous cover rich in endemic taxa. In general, ultramafic areas, especially barrens and rock outcrops, are refugia for endemics and native taxa because they are islands of native flora little modified by exotic taxa that are invasive elsewhere, but intolerant of ultramafic soil (Anderson, Fralish, and Baskin 1999; Harrison et al. 2006; Safford et al. 2005). NORTH COAST RANGES

Selected Vegetation Types

California's ultramafics extend west-east from the North Coast Ranges to the west-facing slopes of the Sierra Nevada and north-south from the Klamath-Siskiyou Mountains to the South Coast Ranges (see Fig. 3.12). Forests, woodlands, chaparral, and grasslands are the main vegetation types that occur on ultramafics (Sanchez-Mata, Rodriguez-Rojo, and Barbour 2004). Conifer forests always are climatic climax vegetation, but conifer-pine-oak woodlands can be either climatic or edaphic, and some chaparral types are a xeroedaphic climax. The pine-oak woodlands on ultramafics at low elevations are related to formations dominated by Quercus douglasii and Pinus sabiniana, which are widely distributed throughout the foothills that surround the Great Valley; they are included in the phytosociological association Pino sabinianae-Quercetum douglasii (Rivas-Martinez 1997). At middle elevations, conifer forests and woodlands on ultramafics contain Pinus ponderosa, P. jef(reyi, and sometimes closed-cone conifers such as Pinus attenuata. At high elevations, conifer forests are the climatic climax rather than edaphic vegetation. Chaparral is a complex of many shrub-dominated vegetation types that cover 8.50/0 of California (Barbour and Major 1988; Keeley and Soderstrom 1986). Chaparral can be climax or seral; that is, it occurs on sites that cannot support large forested communities because of climatic or edaphic factors. Ultramafic chaparral is open and low, and it grows on strongly weathered (serpentinized) parent rocks widely distributed from Santa Barbara County north through the Coast Ranges and Sierra Nevada foothills. The shrubs exhibit xeromorphic traits: they are dwarfed, compact, and close to the ground; they have low productivity; and the evergreen leaves are often reduced in area, curled, or thickened. Common dominants include Adenostoma fasciculatum, Arctostaphylos spp., Ceanothus spp., Frangula californica s. I., Fremontodendron californicum, Fremontodendron decumbens, Garrya congdonii, Heteromeles arbutifolia, Quercus durata, and Rhamnus crocea. Scattered trees, sometimes growing as tall shrubs, include Pinus sabiniana, Umbellularia californica, Pseudotsuga menziesii, Quercus chysolepis, and Q. agrifolia. There is considerable open intershrub space occupied by perennial bunch grasses, subshrubs, and annual herbs. Ultramafic grasslands are very diverse, and they have most recently been described by Rodriguez-Rojo et al. (2001a, 2001b). Barrens are also devoid of woody vegeta-

Ultramafic areas are a common element at low elevations in the North Coast Ranges (McCarten 1987). The most extensive vegetation types are Sargent cypress woodlands (Cupressus sargentii) toward the coast, chaparral in the most continental areas, pine woodland that from Sonoma County to Mendocino and Humboldt Counties, and riparian scrub. All are xero-edaphic climax types. Sargent Cypress Woodland

Sargent cypress woodlands often occupy oceanic, maritime sites (e.g., San Francisco Bay region sensu lato and coastal portions of Humboldt, Mendocino, Sonoma, Santa Clara, San Mateo, Monterey, San Benito, San Luis Obispo, and Santa Barbara Counties). In addition, these woodlands extend as a riparian formation close to Brewer willow scrub (Salicetum brewed) into more inland, protected areas such as canyons, interior valleys, at the head of streams, and where summer fog is of frequent occurrence. In wet years some groves receive a surprising amount of rain (Hardham 1962). The floristic composition of Sargent cypress woodlands growing on ultramafics is very consistent from stand to stand. Cupressus sargentii is usually the overwhelming or sole dominant tree in the overstory, but Pinus sabiniana can co-occur in northern and central California and P. coulteri in central and southern California. In the most mesic situations, scattered elements from mixed evergreen forest are also present (e.g., Pseudotsuga menziesii, Umbellularia californica, Arbutus menziesii, Quercus chysolepis, Lithocarpus densiflorus var. densiflorus, and Quercus agrifolia). Quercus durata is the most common shrub. Adjacent herbaceous communities include the perennial species Calamagrostis ophitidis, Elymus multisetus, Galium andrewsii, Melica californica, M. torreyana, and Polygala californica, plus many ephemerals. Adjacent rocky places are dominated by Allium falcifolium, Asclepias solanoana, Calochortus raichei (The Cedars, Sonoma County), Cymopterus terebinthinus var. californica, Galium californicum subsp. californicum, Streptanthus morrisonii s.1., S. tortuosus var. suf(rutescens, and Cheilanthes gracillima. Riparian Scrub

Communities dominated by Salix brewed occupy ultramafic streambeds, creeks, and canyons in lowlands throughout the Coast Ranges (see Fig. 3.14, later). They are vicariants of

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

9S

TABLE 3.1

Releves of 12 Stands of Brewer Willow Riparian Scrub (Salicetum breweri)

Releve (altitude) 1

2

3

4

5

6

7

8

9

10

11

12

5

(420m) (420m) (240m) (420m) (420m) (830m) (460m) (480m) (240m) (830m) (350m) (370m) (440m)

Characteristics

Salix brewed

5

5

5

Frangula tomentella

+

+

+

Baccharis salicifolia

+

+

5

5

5

5

+

5

+

5

5

+

III III

+

+

+

Calycanthus occidentalis

v IV

+ +

Aquilegia eximia

5

+

+

+

Stachys albens

5

+

11

+

Carex mendocinensis

+

+

Others

Toxicodendron diversilobum

+

+

+

+

+

+

+

+

IV

Equisetum braunii

+

+

+

Datisca glomerata

+

+

+

+

III

+

+

11

+

+

11

Artemisia douglasiana

+

Clematis ligusticifolia

+ +

Rhododendron occidentale s. 1.

+

III

+

Cupressus sargentii (52) Umbellularia californica (52)

+

+

Lonicera vacillans

+

+

+

+

11

+

+

+

11

+

11

+

Cypripedium californicum

+

Melica torreyana

+

+

Iris macrosiphon

+

+

10

11

Total number of taxa represented

NOTE:

4

5

6

6

8

8

9

9

10

10

8

From Sonoma (11, 12), Colusa (I, 2, 4, 5, 7, 8), Lake (3, 9), and San Benito (6, 10) Counties. The numbers + and 1-5 are standard phyto-

sociological cover-abundance classes, + meaning < 10/0 cover and 5 meaning> 750/0 cover. The column of Roman numerals at the right side denotes classes of constancy, I meaning < 200/0 and V meaning> 80%. Accepted taxonomic rank for abbreviated taxa: Equisetum braunii, Equisetum telmateia subsp. braunii; Frangula tomentella, Frangula californica subsp. tomentella; Lonicera vacillans, Lonicera hispidula var. vacillans. Holotypus association: Releve 9*, Lake Co.: Between Aetna Springs and Middletown, Butts Canyon Road, Butts Creek (N38°42'·W122°27'), 600/0 cover, 200 m 2 area, serpentinite boulder beds and sandy soils on avalanche sites [reg. DSM 85a/4-02]. Others, not shown: Melica californica + in 3; Cornus glabrata + in 5; Perideridia kelloggii I, and Rupertia physodes + in 9; Sisyrinchium bellum I, Encameria linearifolia +, Yucca whipplei +, Erysimum capitatum +, Eschscholzia californica +, Distichlis spicata +, and Melica imperfecta + in 10; Epipactis gigantea + in 12. Abbreviations: S2 = second stratum; S = synthesized column; between brackets = [reg. = registration number/month-year].

Salix delnortensis groves on Klamath-Siskiyou ultramafics. Table 3.1 summarizes 12 releves of this homogeneous scrub community compiled from Sonoma, Colusa, Lake, and San Benito Counties in the new association Salicetum breweri, which we propose as the type for the new alliance Salicion breweri [Salicetum breweri Sanchez-Mata & Barbour ass. nova hoc loco, holotypus associatio: Table 3.1, releve 9 from Lake

96

County; Salicion breweri Sanchez-Mata & Barbour all. nova hoc loco has Salicetum breweri as alliance type and Salix breweri as main characteristic species]. Another riparian scrub type is dominated by Frangula tomentella and Rhododendron occidentale s. I., and related with the Sargent cypress woodland. Associated shrubs include Salix breweri, Salix lasiolepis, Baccharis salicifolia, and

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

Calycanthus occidentalis. Springs and peaty communities contain such rare and threatened plants as Brodiaea peduncularis, Carex mendocinensis, Cypripedium californicum, Delphinium llliginosum, Epipactis gigantea, Helenium bigelovii, and Parnassia californica. GREAT VALLEY AND SIERRA NEVADA

The most extensive vegetation characterizing ultramafic areas within the Great Valley are pine-oak woodlands on deep soils developed from scarcely weathered ultramafic parent rocks throughout the inner Coast Ranges foothills, diverse chaparral types in the most continental areas and spreading into the foothills on soils derived from strongly serpentinized parent rocks, and conifer woodlands at higher elevations that also grow on strongly serpentinized soil. Pine and Pine-Oak Woodlands

Pine-oak woodlands are open formations dominated by Qllercus douglasii and Pinus sabiniana. They occur to the northwest of Glenn County (to the Klamath region), east into the montane zones of the Sierra Nevada, and south to midelevations in Monterey, San Benito, San Luis Obispo, Fresno, and Kern Counties. They belong to the association Pino sabinianae-Quercetum douglasii Rivas-Martinez 1997 included in the phytosociological class Heteromelo arbutifoliae-Quercetea agrifoliae Rivas-Martinez 1997, which can be considered to be climax vegetation. We present, in Table 3.2, a summary of 13 releves taken from Shasta, Napa, Yolo, Contra Costa, Alameda, Amador, El Dorado, and Santa Barbara Counties. Ponderosa pine woodlands occur mainly on low-elevation ultramafics in the northwestern part of the Great Valley and along some west-facing slopes of the Sierra Nevada foothills. They grow on strongly serpentinized soil, and they are in the vicinity of ponderosa pine-California black oak woodland that grows on deep, nonultramafic soil. Usually these woodlands have three canopy layers. The tree dominated by Pinus ponderosa associated with Caloce-

drllS decurrens, Pinus attenuata, Pinus lambertiana, P. ponderosa x P. jeffreyi (Armstrong 1977; Callaham 1956), P. sabiniana, and Quercus chysolepis. The shrub layer is dominated by Arctostaphylos viscida subsp. viscida, and Ceanothus cuneatus var. cuneatus, associated with Arctostaphylos patula, A. viscida x A. patula, Ceanothus lemmonii, Cupressus macnabiana (local), Eriodictyon californicum, Heteromeles arbutifolia, Pickeringia montana, and Quercus garryana var. breweri (northwest border of Sacramento Valley). The herb layer is a mix of sparse perennials (Allium campanulatum, Ceanothus prostratus, Elymus elylnoides, Eriophyllum lanatum s. 1., Monardella villosa s. I., Penstemon laetus s. I., Polygala cornuta, and Salvia sonomensis) and many annual grasses and forbs. ]effrey pine woodlands are also present on ultramafics, but they are range up to higher elevations, especially in the Sierra Nevada. At mid elevations, they are present in Mon-

terey, San Benito, San Luis Obispo, Fresno, and Kern Counties, and also in the Sierra Nevada. Hybrid populations between Pinus ponderosa, P. jeffreyi, and P. coulteri (central and south Coast Ranges) can be found in some localities (Callaham 1956; Glooschenko 1961; Armstrong 1977). Associated trees include Calocedrus decurrens, Pinus coulteri, Pinus lambertiana, and Pinus attenuata. The floristic composition of higher elevation ]effrey pine stands depends on the type of ultramafic substrate, whether serpentinite (Bl) or diorite/granodiorite (B2). Associated trees include Abies Iowiana, Calocedrus decurrens, Lithocarpus densiflorus var. echinoides, Pinus contorta var. murrayana, P. jeffreyi x P. ponderosa, P. Iambertiana, P. monticola, P. sabiniana (local), Quercus chrysolepis, Q. kelloggii, and scattered individuals of Pseudotsuga menziesii and Umbellularia californica. Shrub species are many: Amelanchier alnifolia s. 1. (B1, B2), Arctostaphylos nevadensis (B2), A. patula (B2), A. patula x A. nevadensis (A. xbarbouri, B2), A. viscida subsp. viscida (B1), Ceanothus cordulatus (B2), C. cuneatus var. cuneatus (B1), Frasera albicaulis var. nitida (B1), Prunus emarginata (Bl, B2), Quercus vaccinifolia (B2), and Ribes roezlii (B2). There is a high diversity of perennial grasses and forbs. ]effrey pine woodland is a xero-edaphic climax and this vegetation type belongs to the phytosociological class Calocedro decllrrentisPinetea jeffreyi Rivas-Martinez and Sanchez-Mata 1997. ]effrey pine woodlands on diorite or granodiorite-derived soils may have a shrub layer dominated almost monospecifically by huckleberry oak (Quercus vaccinifolia) (association Querco vaccinifoliae-Pinetum jeffreyi Rivas-Martinez and SanchezMata 1997; Fig. 3.13). Chaparral

Ultramafic chaparral types within the Great Valley are diverse, but all are dominated and structured by the ultramafic endemic shrub Quercus durata. It is also a California endemic. Scattered trees of Pinus sabiniana and Pinus coulteri can exist in central and southern areas. Quercus durata s. str., a Californian endemic distributed northward of Santa Barbara County, is the only American species of Quercus that has a restricted distribution on ultramafics. Sufficient bioclimatical, biogeographical, and floristic differences exist to define four ultramafic vegetation types: one in the northwest (Tehama, Glenn, Lake, Colusa, Napa, Sonoma, and Yolo Counties; Table 3.3: the association Ceanotho albiflori-Quercetum duratae Sanchez-Mata, Barbour and Rodriguez-Rojo [in Rivas-Martinez] 1997) (see Fig. 3.14, later), another in the central area (Contra Costa, Alameda, and Santa Clara Counties), a third in the south (foothills throughout Monterey, San Benito, San Luis Obispo, Fresno, and Kern Counties), and a fourth in the Sierra Nevada foothills. Klamath-Siskiyou Region

The ultramafic vegetation in this unique area is very diverse and more closely related to those of the Sierra Nevada than

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

97

TABLE 3.2

Releves of 13 Stands of Pine-oak Vegetation (Pino sabinianae-Quercetum douglasii)

Releve (altitude)

7 11 12 13 9 10 1 2 4 5 6 8 5 3 (960m) (620m) (620m) (300m) (310m) (290m) (300m) (660m) (960m) (430m) (440m) (460m) (890m) (S60m)

Trees

Quercus dOllglasii

3

2

3

Pinus sabiniana

3

2

2

4

3

Qllerclls wislizenii

2

+

2

4

4

2

3

2

4

2

V

4

2

2

IV

+

+

11

+

+

2

Qllercus kel/oggii Qllerclls agrifolia

4

4

V

+

11

Characteristics

Pinus sabiniana (52)

IV

2

Ceanothlls cuneatus

2

+

Heteromeles arbutifolia

+

+

+

+

+

+

+

+

IV

2

IV IV

Phoradendron villosllm Rhamnus crocea

+

+

+

+

+

+

IV

+

IV

Phoradendron vil/osum (P)

+

Aesculus californica

+

IV

+

2

III

Melica torreyana Lonicera vacil/ans

+

Querclls berberidifolia

+

Adenostoma fasciculatum

+

+

Galium andrewsii

+

+

+

Arctostaphylos pulchella

+

+

+

III

+

III

+

11

+

11

+

+

+

Diplacus aurantiaclls

III

2

2

+

11 11

2

Symphoricarpos mol/is

+

Fritil/aria affinis +

Rhus trilobata

II

+

II

+

+

Arctostaphylos mariposa

+

+

2

2

11

Others

2

+

Tox icodendron diversilobum

Pentagrarnma triangularis

2

2

+

+

+

+

+

V IV

+

Galillm porrigens Festuca californica

2

IV

+

IV

+

Sanicula crassicaulis

III

Elymus glazlclls

III +

Umbelllllaria californica (52)

+

Poa secunda Marah fabaceus Chlorogalum pomeridianum

+

+ +

2

+

+

+

II 11

+ +

11

+

II

TABLE

3.2 (continued)

1 2 3 4 5 6 7 8 9 10 11 12 13 5 (960m) (620m) (620m) (300m) (310m) (290m) (300m) (660m) (960m) (430m) (440m) (460m) (890m) (S60m)

Releve (altitude)

Iris douglasiana Eriodyction californicum Nassella pulchra Achillea californica s. I. Eriophyllum lanatum s. I. Dichelostemma capitatum Artemisia californica

+

Ranunculus californicus Total number of taxa represented

13

23

Trees: Quercus garryana 1 and Juniperus australis + in 1; Quercus xmorehus + in 6, 1 in 7; Quercus chrysolepis 2 in 9 and 13; Pinus ponderosa + in 10 and 11; Umbellularia californica 2 in 13. Characteristic taxa: Yucca whipplei 1 and Calystegia collina s. I. + in 2, 1 in 3; Juniperus californica + in 8; Ptelea crenulata + in 13. Others: Pellaea andromedifolia + , Comandra californica 2, Cercocarpus montanus s. 1.2 in 1; Salvia spathacea + in 2;

Salvia sonomensis + in 2 and 3; Lomatium macrocarpum 1 in 1, + in 3; Leptodactylon californicum + in 3; Baccharis pilularis s. I. + in 5 and 8; Lupinus succulentus + in 6 and 7; Melica californica + in 6, 7, and 9; Stachys ajugoides 1 in 6 and 7, + in 9; Triteleia laxa 1 in 6,2 in 7, and 1 in 11; Sysirinchium bellum 1 and Lithophragma bolanderi 1 in 7; Ceanothus leucodermis + and Wyethia helenioides + in 8; Cercocarpus montanus s. I. 1 and Calochortus pulchellus 1 in 8 and 9; Bromus anomalus 1 in 10 and 11, and + in 12; Frangula tomentella + in 10, 11, and 12; Rhamnus ilicifolia + in 10 and 12; Sambucus mexicana + in 8 and 11; Ericameria arborescens 1 and Holodiscus aggr. microphyllus 1 in 9; Lathyrus jepsonii + in 11 and 12; Quercus chrysolepis (52) 2 in 13; Clematis ligusticifolia 1, Dodecatheon hendersonii 1, Delphinium nudicaule 1, Lomatium californicum + , Cercocarpus betuloides s. I. + , Zigadenus micranthus + , Holodiscus discolor s. I. + , Ceanothus sorediatus + , Sanicula bipinnata + , Lupinus albifrons + , and Calochortus pulchellus 1 in 13. NOTE: From Shasta (1), Napa (6), Yolo (5), Contra Costa (9, 13), Alameda (8), Amador (4, 7), El Dorado (l0, 11, 12), and Santa Barbara (2, 3) Counties. Table format is the same as that described for Table 3.1. Holotypus association: Releve 1*, Shasta Co.: Between Wengler and Big Bend, Big Bend Road-Pit River, Shasta National Forest, N40058'-W121°55', S face, 20 % slope, 200 m 2 area, andosols on basaltic bedrocks [reg. RM 1996b: 133/ 9-96] [Rivas-Martinez 1997: 30]. Accepted taxonomic rank for abbreviated taxa: Achillea californica: Achillea millefolium var. califomica; Arctostaphylos mariposa: Arctostaphylos viscida subsp. mariposa; Arctostaphylos pulchella: Arctostaphylos viscida subsp. pulchella; Ceanothus sorediatus: Ceanothus oliganthus subsp. sorediatus; Comandra californica: Comandra umbellata subsp. califomica; Frangula tomentella: Frangula californica subsp. tomentella; Juniperus australis: Juniperus occidentalis var. australis; Rhamnus ilicifolia: Rhamnus crocea subsp. ilicifolia. Abbreviations: P = parasitic species.

2

2

2

FIG U RE 3.1 3 Landscape and vegetational scheme (catena) on ultramafics in Sierra Nevada highlands: 1. White fir forest on granitic rock-derived soils (Cas-

tanopsio sempervirentis-Abietetum lowianae Rivas-Martinez and Sanchez-Mata 1997), 2. Jeffrey pine woodland on granodiorite-derived soils (Querco vaccinifoliae-Pinetum jeffreyi Rivas-Martinez and Sanchez-Mata 1997).

to those of the North Coast. Typically, several main ultramafic vegetation types are recognized: ]effrey pine woodland; Shasta red fir forest with Brewer spruce, noble fir, and mountain hemlock; subalpine woodland; and whitebarkpine woodlands at rocky summits. ]effrey pine (Pinus jeffreyi) and whitebark-pine (Pinus albicaulis) woodlands can

be considered xero-edaphic climax vegetation, while both fir forest formations are climatic climax vegetation. In addition to these main forest types we will remark some peculiar vegetation types, commonly very rich in endemics, and restricted to the ultramafic Klamath-Siskiyou high mountain areas, very often with a local character.

CALIFORNIA SOILS AND ULTRAMAFIC VEGETATION

99

TABLE 3.3

Releves of 20 Stands of Chaparral Vegetation (Ceanotho albiflori-Quercetum duratae) 2

Releve (altitude)

3

4

5

6

7

8

9

10

11

12

13

14*

15

16

17

18

19

20

S

(320m) (830m) (780m) (l70m) (640m) (340m) (380m) (780m) (270m) (690m) (320m) (71m) (690m) (350m) (730m) (700m) (450m) (380m) (720m) (450m) (540m)

Trees 1

+

1

1

1

1

1

1

2

1

+

1

1

3

1

1

2

1

+

2

V

Quercus durata

4

3

3

4

4

4

4

4

4

5

3

4

4

4

4

4

5

5

3

4

V

Ceanothus albiflorus

1

+

+

1

2

1

1

3

1

2

2

1

1

2

1

2

2

1

+

2

V

Arctostaphylos pulchella

+

+

1

+

1

+

1

+

2

+

+

1

+

1

1

+

1

+

+

1

V

Heteromeles arbutifolia

+

2

1

1

1

+

+

1

1

+

1

1

2

2

+

2

2

+

3

V

Melica torreyana

1

+

1

1

2

1

1

1

2

+

2

2

3

1

2

IV

1

+

1

1

2

+

+

+

+

+

1

1

1

+

Pinus sabiniana

Characteristics

1

Ceanothus cuneatus Garrya congdonii Adenostoma (asciculatum

+

1

Galium andrewsii

+

Lomatium purpureum

1

+

1

1 1

2

+

1

1

+

1

+

1

2

+

1

1

1

1

2

+

+

1

1

+

+

1

1

1

1

+

1

1

1

2

Fraxinus dipetala Pickeringia montana

+

Lomatium dasycarpll1n

+

Calystegia collina s. I.

+

1

+

+

1 1

+

+

+

1

1

1 1

Rhamnus ilici(olia

+

Quercus berberidi(olia

+

+ +

+

+ +

1

1 1

1

2

1

III

+

1

III

1

1

Il

+ +

1

+

III

1 1

Il

+

+

1

+

Il

+

1

Il

+

1

1

Styrax redivivus

+

+

II

1

2

+

2

1

2

Fremontodendron califomicum

1

+

Cupressus macnabiana

1

+

Perideridia kelloggii Quercus durata x Q. berberidi(olia Rhus trilobata

+

+

+

Il

1

Polygala cali(omica Adiantum jordanii

IV III

+

Others Eriophyllum lanatum s. I.

+

+

Toxicodendron diversilobum

+

Galium porrigens

+

Bromus anomalus

+

Eriodyction califomicum

+

+ + + +

+

+

+

+

+

+

+

2

3

2

IV

+

+

+

2

IV

+

III

+

III

+

+

+

+

+

+

+

+

+

+

+

III

+

III

+

Il

+

+

+

III III

+

+

+

+

+

+

+

+

+

Il

+

Il II

+

+

+

+

+

+

+

+

Pentagramma triangularis

Iris douglasiana

+

+

+

Elymus glaucus

Monardella villosa s. I.

+

+ +

Chlorogalum pomeridianum

Lonicera vacillans

IV

III

Triteleia laxa Eriogonum nudum s. I.

+

+

+

Melica califomica

+

+

2

+

Cercocarpus montanus s. I.

+

+

Elymus elymoides

Umbellularia califomica (52)

+

+

2

+ +

+

+

+

Frangula tomentella Festuca califomica

+

+

+

+

Il

+ +

+

Il

Il

2

+

Character taxa: Frangllla califomica + in 1,1 in 3; Clematis lasiantha + in 4 and 19; Castilleja foliolosa + in 5 and 13; Rhamnus crocea + in 15,1 in 20; Aescullls califomica + in 17,1 in 19; funiperus califomica 1 in 6; Quercus agrifolia (52) + in 19; Quercus wislizenii (52) + in 20; Salvia sonomensis + in 20. Others, not shown: Hypericum concinnum 1 in 3 and 13; Aspidotis densa + in 4 and 12; Phacelia corymbosa 1 in 5; Stipa lemmonii + in 6, 1 in 8; Diplacus aurantiacus + in 6; Cirsium occidentale s. I. + in 8 and 20; Ceanothus sorediatus + in 9; Silene califomica + in 9, 1 in 12; Dichelostemma capitatum 1 in 11; Calycanthus occidentalis 1 in 11, + in 19; Pseudotsuga lnenziesii (52) + in 11 and 19; Keckiella lemmonii + in 12,1 in 19; Calamagrostis ophitidis + in 14; Arceuthobium occidentale (P) 1 and Cercis occidentalis (52) + in 15 and 17; Monardella sheltonii + in 15 and 20; Alliunz falcifolium + and Calochortus amabilis 1 in 16; Cupressus sargentii (52) 1 in 13 and 15; Achillea califomica 1 in 18 and 20; Ribes amarum, Vitis californica, Quercus chrysolepis, and Dichelostemma volubile + in 19; Eschscholzia califomica + and Pedicularis densiflora 1 in 20.

NOTE: From Tehama (12, 19), Glenn (6), Lake (I, 8, IS, 16), Colusa (5, 7, 10, 14, 17, 18, 20), Napa (4, 9, 13), and Sonoma (2, 3, 11) Counties. Table format is the same as that described for Table 3.1. Holotypus association: Releve 14*, Colusa Co.: 15 miles West of Williams, Blue Ridge near Bear Creek, N39(04'-WI22(23', SE face, 80 % cover,S % slope, 200 m + 2 area [reg. DSM 59b/9-96, RM 1996b: 160/9-96] [RivasMartinez 1997: 41]. Accepted taxonomic rank for abbreviately mentioned taxa: Achillea califomica: Achillea millefolium var. califomica; Arctostaphylos pulchella: Arctostaphylos viscida subsp. pulchella; Ceanothus albiflonls: Ceanothus jepsonii var. albiflorus; Ceanothus sorediatus: Ceanothus oliganthus subsp. sorediatus; Frangula tomentella: Frangula califomica subsp. tomentella; Lomatium purpureum: Lomatium marginatum var. purpureum; Lonicera vacillans: Lonicera hispidula var. vacillans; Rhamnus ilicifolia: Rhamnus crocea subsp. ilicifolia; Styrax redivivus: Styrax officinalis var. redivivus. Abbreviations: P = parasitic species.

Jeffrey Pine Woodland

These open woodlands occur at the most xeric ultramafic sites (mainly growing on peridotite, gabbro, and serpentinite-derived soils), from low to higher elevations (300-2,000 m). Pinus jef{reyi and Calocedrus decurrens are the only common trees present in these open woodlands; Cupressus bakeri subsp. matthewsii, Abies lowiana, Pseudotsuga menziesii, Pinus

attenuata, Pinus monticola, Pinus lambertiana, Pinus ponderosa, and Pinus ponderosa x Pinus jef{reyi, co-occur as scattered individuals. Open spaces are covered by shrubs and small trees such as Arctostaphylos klamathensis, Arctostaphylos nevadensis,

Arctostaphylos viscida s. I., Arctostaphylos patula, Ceanothus cuneatus var. cuneatus, Frangula californica s. I., Garrya buxifolia, Garrya fremontii, Lithocarpus densiflorus var. echinoides, Quercus garryana var. breweri, Quercus vaccinifolia, Vaccinium parvifolium, and shrub forms of Umbellularia californica. Perennial grasses (Elymus elymoides, Elymus glaucus, Festuca califomica, Festuca idahoensis, Melica californica, Melica geyeri, and Stipa lemmonii) and forbs are often present. These woodlands are in the association Aspidoto densae-Pinetum jef{reyi (Rivas-Martinez and Sanchez-Mata 1997). Fir Forests

Fir forests with Picea breweriana occur in several localities in northwest California and southern Oregon on ultramafics at 1,200-1,800 m elevation. The geographic area of Brewer spruce includes Curry, ]osephine, and ]ackson Counties in Oregon and Del Norte, Siskiyou, Humboldt, Trinity, and Shasta Counties in California (Waring, Emmingham, and Running 1975). These forests are mainly structured by Abies lowiana at midelevations and by Shasta red fir (Abies shastensis) at higher elevations. Pinus contorta var. murrayana, Pinus jeffreyi, Pinus lambertiana, Pinus monticola, and Pseudotsuga menziesii are common associates. The understory is frequently dominated by shrubs such as Arctostaphylos canescens s. I., Arctostaphylos nevadensis,

Arctostaphylos patula, Ceanothus pumilus, Quercus sadleriana, Quercus vaccinifolia, Vaccinium membranaceum, Vaccinium parvifolium, and Xerophyllum tenax. Herbs include Achlys triphylla, Asarum hartwegii, Corallorhiza maculata, Disporum hookeri s. 1., Goodyera oblongifolia, Linnaea americana, among others. Noble fir and mountain hemlock forests grow at elevations > 1,800 m sometimes with Shasta red fir (Abies shastensis). These forests should be framed into the phytosociological class Tsugetea mertensiano-heterophyllae (Rivas-Martinez, Sanchez-Mata, and Costa 1999b). In a few localities, Abies lasiocarpa is an associate in this forest type (Sawyer, Thornburgh, and Bowman 1970).

Subalpine Woodland

On some high, rocky, exposed places that do not accumulate snow-packs, a subalpine woodland structure by Pinus albicaulis occurs. The understory contains Cercocarpus ledifolius s. I., Quercus vaccinifolia, and Purshia tridentata and it may be considered as a xero-edaphic climax. Somewhat more protected, mesic

102

sites su~port an open foxtail pine woodland (Pinus balfouriana subsp. balfouriana), vegetation more restricted to ultramafics than whitebark pine woodland (Fig. 3.14; Sawyer, Thornburgh, and Bowman 1970). Associated trees include Abies shastensis, Pinus albicaulis, Pinus contorta var. murrayana, and Pinus monticola, and common understory herbaceous perennials in openings are Aspidotis densa, Carex rossii, Eriogonum alpinum, Lewisia leana, and Polygonum davisiae. The subalpine woodland that dominates the most mesic subalpine locales, at the bottom of late-melting snowpacks, is structured by Tsuga mertensiana in the overstory and Phyllodoce empetriformis in the understory. This forest type is a hydro-edaphic climax.

Riparian Cedar Forest and Adjacent Wetlands

This ultramafic forest type, dominated by Chamaecyparis lawsoniana, grows in ravines, draws, along streams and on mesic, protected slopes where there is a strong maritime inflence (Zobel and Hawk 1980). Associated tree species in this hydro-edaphic forest include Acer macrophyllum, Alnus

rhombifolia, Alnus rubra, Chamaecyparis lawsoniana, Pinus monticola, Pseudotsuga menziesii and Taxus brevifolia. Subcanopy trees and shrubs are Cornus nuttallii, Quercus sadleriana, Rhododendron macrophyllum, Rhododendron occidentale s. I., Ribes bracteosum, Rosa californica, and Physocarpus capitatus. The perennial herbs and ferns Xerophyllum tenax, Adiantum pedatum, Darmera peltata, Polystichum munitum and Streptopus amplexifolius are common. Nearby seeps on ultramafics produce fen meadows structured by Darlingtonia californica (see Fig. 3.14). Caltha leptosepala var. biflora, Carex echinata, Castilleja miniata subsp.

elata, Cypripedium californicum, Darlingtonia californica, Epipactis gigantea, Eriophorum criniger, Gentiana setigera, Hastingsia alba, Lilium bolanderi, Lilium kelloggii, Lilium pardalinum subsp. vollmeri, Narthecium californicum, Platanthera sparsiflora, Raillardella pringlei, Rudbeckia californica var. glauca, Tofieldia occidentalis subsp. occidentalis, Trillium rivale, and Viola primulifolia subsp. occidentalis are common perennials. Less acidic springs support megaforbs, such as

Veratrum californicum, Veratrum insolitum, Veratrum viride, Aconitum columbianum, Angelica arguta, Aquilegia eximia, Heracleum lanatum, Lilium bolanderi, Lilium occidentale, Pteridum aquilinum var. pubescens, Platanthera leucostachys, Senecio triangularis, and Valeriana sitchensis. SOUTHERN CALIFORNIA

The ultramafic vegetation of southern California predominantly consists of chaparral in the most continental sites, pine woodlands at mid- and high elevations, Sargent cypress woodlands on uplands that experience some maritime influence, and pine-oak woodlands. Ultramafic chaparral occurs at elevations 200/0 of the study sites were converted to other land uses, primarily residential subdivisions, during this period

TABLE 12.8

Mean Basal Area (m 2 /ha) of Predominant Overstory Trees in the Interior Live Oak Subseries of the Montane Hardwood Forest

Dominant Tree Species Arbutus menziesii

Community

Subseries Coast Range and Sierra Nevada

Sierra Nevada

Pinus sabiniana 7

Interior Live Oak-Blue OakFoothill Pine

Quercus douglasii 7

9

7

Interior Live OakMadrone/Poison Oak

Quercus wislizenii

Quercus kelloggii

1

2

15

1

4

2

4

Interior Live Oak/Toyon

1

Interior Live Oak/Whiteleaf Manzanita

2

Interior Live Oak/Yerba Santa/Grass

2

2

2

Interior Live Oak-Foothill Pine/Manzanita

6

2

5

TABLE 12.9

The Area and Extent in Hectares of California Oak Woodland Habitats in Various Regions of the State

San Joaquin Valley/ Eastside

Habitat Type (CWHR)

Central Coasf

Blue oak woodland

443,951

436,299

Blue oak- foothill pine woodland

114,603

Sac. Valley/ North Interior

Central Sierra

382,510

148,088

134,397

185,604

93,295

22,097

6,827

712

Coastal oak woodlands

517,057

10,000

8,414

Montane Hardwood

256,127

313,824

440,277

1,353,835

901,348

1,017,517

Valley oak woodland

TOTAL

° ° 412,758 654,141

North Coast

So. California

TOTAL

30,717

13,760

1,455,325

°

527,899

°

902

405

30,943

88,083

161,475

785,029

218,141

34,804

1,675,931

337,844

210,444

4,475,128

NOTE: Adapted from Greenwood et al. 1993 aDescription of regions: Central Coast: Alameda, Contra Costa, Lake, Marin, Monterey, San Benito, San Luis Obispo, San Mateo, Santa Barbara, Santa Clara, Santa Cruz, Solano, Sonoma, Ventura Counties. San ]oaquin Valley/Eastside: Fresno, Kern, Kings, Madera, Merced, San ]oaquin, Stanislaus, Tulare Counties. Sacramento Valley/North Interior: Butte, Colusa, Glenn, Lassen, Modoc, Plumas, Sacramento, Shasta, Sierra, Siskiyou, Solano, Sutter, Tehama, Trinity, Yolo, Yuba Counties. Central Sierra: Amador, Calaveras, Eldorado, Mariposa, Nevada, Placer, Tuolumne Counties. North Coast: Del Norte, Humboldt, Mendocino Counties So. California: Imperial, Los Angeles, Orange, Riverside, San Diego, San Bernardino Counties.

(Holzman 1993). A similar study of changes in tree and total woody cover of foothill oak woodlands from 1940 to 1988 found these areas were relatively stable (Davis 1995). Pollen analysis studies document the dynamics of oak woodland composition over a very long-term period and highlight the changing influence of human populations (Byrne et al. 1991). Oak woodlands were relatively stable

during the long period of use by Native Americans. Following European settlement, approximately 200 years ago, livestock introduction and clearing for intensive agriculture caused significant declines in oak pollen. Exotic annuals first show up in the pollen record at this same time. Since this initial exploitation, oak pollen has increased dramatically. Current oak pollen deposition is at its highest level, probably due to

OAK WOODLANDS AND FORESTS

325

fire-exclusion policies of the last 80 years, and low-intensity management practices associated with ranching.

Oak Woodland Ecosystem Processes Pre-European Settlement: Herbaceous Flora and Historic Change

Most of the tree species of oaks occupied their current distributions by about 10 million years before the present (BP) in California (AlIen et a1. 1999). The California Mediterranean climate, with its dry summers, was probably well in place by 5 million years BP (Rundel 1987). Byrne et a1. (1991) suggested that the woodland oak species (e.g., blue and valley) moved to higher elevations in the Sierra between 10,000 and 5,000 years ago, based on pollen evidence, but by the mid1800s, woodland species had retreated to their present locations at lower elevations in the Sierra and Coast Ranges. The species composition of herbaceous vegetation in oak woodlands prior to European contact is unknown. Many believe native perennial grasses, particularly the bunchgrass Nassella pulchra, once enjoyed a more widespread distribution (Clements 1934; Heady 1977). Hamilton (1998) has rather convincingly argued against overuse of this paradigm, suggesting that native annual forbs were once dominant, especially in drier parts of the woodland. Holstein (2001) suggested the rhizomatous perennial grass, Leymus triticoides, dominated the pre-agricultural Central Valley floor. (However, his analysis relied on the relict method for which he and others criticize Clements.) Two studies provide physical evidence of pre-European settlement composition. Bartolome et a1. (1986) found greater abundance of distinctive opal phytoliths (silica bodies that are resistant to decay with shapes specific to certain taxonomic groups) at soil depths corresponding to > 150 years ago. The shapes these phytoliths took were specific to those found in perennial grasses, indicating their greater abundance in the past at that particular site (Jepson Prairie near Davis, CA). Mensing and Byrne (1999) offer some physical evidence of the pre-European flora. They examined pollen in sediment cores from the Santa Barbara channel and determined that the presence of the exotic annual, Erodium cicutarhim-now ubiquitous in much of California-pre-dated European settlement and livestock introduction. They show patterns suggesting it invaded from Baja California prior to the Mission Period. Beginning with the introduction of domestic livestock and exotic annuals by European settlers, oak woodland ecosystems have changed dramatically. Herb cover has shifted from perennial to annual (Holmes 1990). Fire interval and intensity have increased (McClaran and Bartolome 1989). Overstory cover has generally increased (Holzman and Allen-Diaz 1991). Soil moisture late in the growing season has decreased, and soil bulk density has increased due to compaction from large herbivore numbers (Gordon et a1. 1989). Riparian zones are now less dense and diverse (Tietje

326

OAK WOODLANDS AND FORESTS

et a1. 1991). A general summary of the changes in ecosystem inputs from presettlement conditions to the current time is shown in Table 12.10. These ecosystem changes are discussed below. The pre-European herbaceous oak woodland understory included native perennial bunchgrasses, annual grasses and annual and perennial forbs (Holmes 1990). Native species were reduced in cover with multiple introductions of alien annual species from Europe, Asia, Africa and South America (Burcham 1970; Heady et a1. 1992). Species extinctions were few (Solomeshch and Barbour 2006). Species composition differences between oak understory and open grassland have been demonstrated by several authors (Saenz and Sawyer 1986; Jackson et a1. 1990; Davis et a1. 1991; Maranon and Bartolome 1994). Some species appear to be strongly controlled by this dichotomy, but generalizations for California's oak woodlands are tenuous. A given species may be strongly associated with tree canopy cover at one site but with open grassland in another. For instance, Nassella pulchra is thought to be an open grassland species, but it has been observed scattered beneath relatively continuous oak canopy in the Sierran foothills Uackson and Bartolome 2002). Alternatively, Cynosurus echinatus is rare in open grassland but very common beneath oak canopy throughout the state. Rice and Nagy (2000) sought the mechanism for the spatial separation of Bromus diandrus (found under canopy) and Brolnus hordeaceus (found in open) and reported that interspecific competition was important in the high-resource soils beneath oak canopy, but only B. hordeaceus could tolerate the harsher physical conditions of open grassland (Le., there was little evidence of competition between these species in the open). Working (2002) found that aspect, measured at the 10- to 100-m 2 scale, was a more important determinant of species composition than canopy cover. Hence, species composition of the herbaceous understory is a result of a complicated mix of time, site, and abiotic and biotic interactions. Generalizations are tenuous. Plant species richness was shown to be highest at intermediate (35-57 g/m 2 ) herbaceous biomass levels (Heady et a1. 1992; Maranon and Bartolome 1994) following the model of Grime (1979) and discussed by Maranon and Garcia (1997) and Garcia et a1. (1993). They suggested that maintaining an oak overstory component provides for maximal landscape diversity, owing to different plant assemblages under canopy and in the open.

Grazing Processes and Forage Production

Livestock grazing has had a major impact on California's oak woodlands. By 1880, Spanish coastal missions had 4 million sheep and 1 million cattle (Holmes 1990), fostering a large demand for forage and oak browse. Currently, twothirds of all woodlands are grazed (Huntsinger 1997). In addition to domestic livestock grazing, feral hogs consume acorns, as do ground squirrels and pocket gophers.

TABLE 12.10

COlnparison of Oak Woodland Conditions Before European Settlement, During Extensive Ranching Period, and in Urban Interface Areas

Pn:-European Settlenzent Conditions

Extensive Ranching Period

Current Urban Inf7uence

Perennial herbaceous layer

Exotic annual invasion

Increasing annual invasion, especially noxious weeds

Regular fire interval

Continuation of regular fire interval

Fire suppression policies and long fire interval and increased intensity

More open overstory layer

Rangeland clearing and tree thinning

Increased multi-species overstory layer of unconverted stands

Soil lnoisture higher later into growing season

Soil moisture less late into the growing season due to exotic annuals

Decreased soil moisture late in growing season due to exotic annuals

Lower soil bulk density

Increased soil bulk density

Increased soil bulk density

Snags, large woody debris

Snags, woody debris cleaned up in typical management activities

Less attention to clean-up; increased snags and woody debris

Dense, diverse riparian zone

Riparian zones less dense and diverse

Higher human use of riparian zones, and increased storm runoff from urban areas

Lower herbivore densities

Higher herbivore density, primarily domestic livestock

Decrease in numbers of domestic livestock and wildlife

NtHL:

Adapted

frOln

Standiford 2001.

Grazing has both positive and negative effects on oak woodland sustainability. Positive grazing effects include reduced rnoisture competition between oaks and herbaceous material (Hall et a1. 1992); reduced leaf area in grasses, which may help conserve moisture late in the growing season (Welker and Menke 1990); reduced habitat for rodents that consume acorns and young seedlings; and elimination of fuel ladders, reducing the probability of crown fires. Some of the negative effects of livestock grazing include consumption of oak seedlings and acorns (Davis et a1. 1991; Adams et a1. 1992; Hall et a1. 1992; Swiecki et a1. 1997); increased soil compaction, Inaking root growth for developing oak seedlings more difficult (Cordon et a1. 1989); and reduced soil organic matter. The oak canopy has an effect on forage production, composition, and quality; the magnitude of the effect depending on precipitation, oak species, and amount of overstory cover (Table 12.11). Oaks compete with the understory for Inoisture, and they alter the nutrient status of the site because of their deep-rooting habit and litter quality. Oak renl0val was historically recommended as a means of increasing forage production on hardwood rangelands (Ceorge 1987). For the deciduous blue oak, most studies have dernonstrated increased forage production following tree renl0val on areas previously containing over 2591) canopy cover and receiving > 50 cm of rain (Kay 1987; Jansen 1987). Conversely, where there is 25(Y() canopy cover will have less forage growth than cleared areas. One study of drought years in the Southern Sierra Nevada foothills, showed that live oak shade helped conserve soil moisture, resulting in higher understory production than on open sites (Frost and McDougald 1989). Production increases beneath blue oak canopies (or in areas previously beneath blue oak canopies) is attributed, in part, to increased soil fertility due to leaf fall and decomposition Oackson et a1. 1990; Frost and Edinger 1991; Firestone 1995). Enhanced soil fertility also improved forage quality. Because the nutrient input from leaf litter ceases after tree removal, herbaceous production gradually declines to the levels of adjacent open areas (Kay 1987; Camping et a1. 2002).

OAK WOODLANDS AND FORESTS

327

TABLE 12.11

The Effect of Oak Canopy on Hardwood Rangeland Forage Production Canopy Cover

Winter Forage Production

Spring Forage Production

-/+

-/+

-/+

-/+

Scattered « 1O(y!> cover)

+

+

Sparse (10(Y!>-25(Y!> cover)

+

+

-/+

-/+

Live oaks Scattered « 1Oq1h cover) Sparse (10(Yh-25(J;() cover) Moderate (25(J;()

-60(~h

cover)

Dense (over 60(Y!> cover) Deciduous oaks

Moderate (25 (Y!)-60(J;() cover) Dense (over 60(Y!) cover)

NOTI: a "+ indicates that forage production is enhanced by oak canopy, and a "-" indicates that forage production is inhibited by oak canopy. Adapted from Allen-Diaz et al. 1999. 11

Oak canopies also have an effect on forage species composition. Studies have found that understories of both blue and live oak stands favor later-successional herbaceous species such as wild oats (Avena (atua), soft chess (Brolnus hordeacells), and ripgut brome (Bromus diandrlls). Clovers (Trifolium spp.), annual fescues (Vulpia spp.), filaree (Erodium spp.), and soft chess account for more of the total herbage biomass in open areas than under oak canopy (Holland 1980; Ratliff et a1. 1991). Current oak management guidelines for ranchers are (Standiford and Tinnin 1996): • There is little or no value in removing blue oaks in areas with 50 cm of annual rainfall, thinning oaks where the canopy cover is >50(J;() will have the greatest positive effect on herbaceous production. • In areas thinned for forage enhancement, residual tree canopy cover of 25(Y!)-35(Y!) is able to maintain soil fertility and wildlife habitat, and minimize erosion processes. • Tree removal should always consider all values of the trees, including wildlife habitat, soil stability, and so forth in addition to possible forage production benefits. Soil Processes and Nutrient Cycling

Frost and Edinger (1991) found higher organic carbon levels, greater cation-exchange capacity, lower bulk density, and greater concentrations of SOlne nutrients (at a soil depth of 0-5 cm) under blue oak canopies than in open grassland. Organic matter input from blue oak leaf litter primarily accounts for this finding, and leaching of nutrients fron1 rain-

328

OAK WOODLANDS AND FORESTS

water drip may also make a significant contribution. The soil conditions beneath interior live oak and blue oak are similar; more intensive shading from the evergreen canopy, therefore, is thought to primarily account for the reduced total annual herbage production under interior live oaks growing in moderate environmental conditions (Frost and Edinger 1991). Oak woodlands with perennial grasses retain soil moisture later in the growing season than woodlands with annual grasses (Gordon et a1. 1989). This difference in soil moisture may partially explain the observed lack of sapling recruitment in oak woodlands containing an annual grass understory. Evaluation of hardwood rangeland soil bulk density shows that areas with livestock grazing have a higher bulk density than ungrazed areas. Jackson et a1. (1990) found that soils under blue oak canopies have higher nitrogen turnover rates and inorganic nitrogen contents than surrounding open grassland soils, due primarily to the higher nitrogen content from mineralization of oak leaf litter. There was no difference in soil water potential between the understory and the open grassland. The increased fertility under blue oak canopy did not result in enhanced forage productivity. In general, grazing accelerates carbon and nutrient cycling by effectively bypassing the microbial decomposition pathway. Livestock mineralize plant organic material much more quickly than microbes and return it to the soil and atmosphere as feces, urine, and gas. In perennial grasslands of the Midwest, accelerated nutrient cycling is credited for stimulating net primary productivity (Frank and McNaughton 1993; Frank et a1. 1994; Frank and Evans 1997). However, similar grazing effects on nutrient dynamics in California annual grassland were not evident (Davidson et a1. 1993; Dahlgren et a1. 1997). Nitrogen quickly cycles within annual-dominated

ecosystems where plant species possess low nutrient-use efficiencies and high litter qualities irrespective of herbivory (Jackson et al. 1988; Schimel et al. 1989; Davidson et al. 1990). Dahlgren et al. (1997) describe soils beneath oak canopy as "islands of fertility" because of greater carbon, nitrogen, and phosphorous stocks, compared to adjacent open grasslands sites. The patchiness of oak woodland canopy may be enhanced by the ability of oaks to garner water and nutrients from beyond the canopy perimeter, from the open grassland spaces between them and their neighbors, and then preferentially returning leaf litter below the existing canopy, thereby redistributing ecosystem resources.

Conservation and Restoration Issues Oak Regeneration and Recruitment Processes

One of the key concerns that landowners, policymakers, and the public have about oak woodlands is whether oak regeneration is adequate to sustain current woodlands and savannas. Several surveys of oak regeneration (White 1966; Griffin 1973; Bartolome 1987; Bolsinger 1988; Bernhardt and Swiecki 1991; Danielsen and Halverson 1991; Standiford et al. 1991; Swiecki et al. 1997) have shown a shortage of saplings for certain species (especially blue oak, Engelmann oak, and valley oak) in certain regions of the state (sites at low elevation, on south- and west-facing slopes, on shallow soils, and with high populations of natural or domesticated herbivores). If this shortage of small trees continues over time, then the oak stands may gradually be lost as natural mortality or tree removal decrease the number of large, dominant trees, and woodlands convert to other vegetation types such as brushfields or grasslands. Deciduous oaks have reproduced poorly in the past 50+ years (Griffin 1977; Muick and Bartolome 1987). Although seedlings become established, few develop into saplings. Live oaks, whose seedlings may be more resistant to grazing and browsing, have produced saplings with more success than have deciduous oaks. Pocket gophers, a significant seedling predator, may prefer deciduous oak roots to those of live oaks. Where there has been a failure of deciduous oak seedling establishment, the cause may be attributed to damage to acorns and seedlings by insects, cattle, deer, and rodents (Griffin 1977; Borchert et al. 1989). Older individuals dominate most extant blue oak populations (Davis 1995; Swiecki et al. 1997). Recruitment of valley and blue oak saplings is not sufficient to maintain existing stands according to Muick and Bartolome (1987) but Tyler et al. (2006) disagree and others think stump sprouting may reduce the concern about sustainability (McCreary et al. 1991,2002; Standiford et al. 1996).

Riparian Ecosystem Processes

Although a small percentage of the state's water supply originates on hardwood rangelands, Virtually all of it flows

FIGURE 12.6 Typical spring-fed wetland in the blue oak-foothill pine type in the Sierra foothills near Marysville, CA. Spring-fed wetlands are a high quality source of forage and water for domestic and wild animals.

through oak woodland riparian zones (Ewing et al. 1988). Also, most of the state's major reservoirs are located within oak woodlands. Riparian zones proVide important habitat for wildlife and aquatic organisms. Management activities influence water quality and wildlife and fisheries habitat. Yet, removal of up to one-third of the oak canopy had little effect on water quality and yield in one regional study (Epifanio et al. 1991). New efforts have been started to develop rangeland management practices that will minimize erosion and improve water quality. In urban interface areas, riparian zones are often subject to very high levels of human use for recreational purposes. Scott and Pratini (1997) documented how urban development increases human use of riparian areas, lowering the habitat value for various wildlife species and decreasing overall biological diversity. Spring-fed wetlands and riparian areas are often the only sources of summer water in oak woodlands and they are heavily utilized by graZing animals (Fig. 12.6). However, light-to-moderate, autumn/winter graZing had little effect on Sierra Nevada foothill spring-fed vegetation, even after 10 years of treatment (Allen-Diaz et al. 2004; Allen-Diaz and ]ackson 2000). Continued monitoring of these systems under experimental treatments has shown that, by years 7 through 10, moderate grazing reduced herbaceous cover, light grazing had minimal effect, and grazing removal significantly increased cover (Jackson 2002). However, the increased cover brought with it an undesirable accumulation of plant litter that suppressed subsequent plant productivity. Studies examining grazing effects on vegetation in riparian systems other than spring-fed wetlands are few. Intensive graZing can negatively affect water quality, plant biodiversity, productivity, wildlife habitat, wildlife species biodiversity, and nutrient cycling in riparian areas in regions with continental-type climates (Kauffman et al. 1983; Kauffman and Krueger 1984; Fleischner 1994; Clary 1995; Belsky et al. 1999), but extrapolation to Mediterranean-type

OAK WOODLANDS AND FORESTS

329

regions should be made very cautiously (Larsen et al. 1998; Gasith and Resh 1999). Biodiversity in oak woodland spring ecosystems was maximized with light grazing (Allen-Diaz and ]ackson 2000). Nitrate release in spring waters, contrary to common belief, increased with removal of cattle grazing. Methane production from the springs was reduced by grazing removal (Allen-Diaz et al. 2004). Effects of livestock grazing and grazing removal on wetland ecosystems of these regions are varied (Allen-Diaz et al. 1998, 2004; Allen-Diaz and ]ackson 2000). Research shows that the timing of grazing and sampling of grazing affects must be considered to understand livestock-ecosystem relationships (Tate et al. 1999).

Fire Ecology

Fire is an important ecosystem process and management tool in oak woodland. Fire affects oak woodland stand structure, oak regeneration, wildlife habitat, nutrient cycling, and livestock grazing. The ecological effects of fire depend on fire frequency, timing, intensity, and complexity. Adjacent vegetation types, such as chaparral and montane forests, influence oak woodland fire regimes. Recent increases in the acreage of stand-destroying fires in oak woodlands point to the need to include fire in management plans in order to sustain the economic and ecological values of oak woodlands. Because of the long period of human habitation of oak woodlands, it is extremely difficult to define the "natural" fire regime. Lightning-caused fires originating from major storms coming northward from Mexico have helped shape oak woodlands. It is speculated that decades may pass between lightning-caused fire events in oak woodland (Griffin 1977). Mature oaks can survive regular low-intensity surface fires, and most woodland oak species have the capacity for young trees to resprout after being top-killed by fire. Native Americans used fire in their stewardship of oak woodlands (Holmes 1990). There are numerous accounts of burning by Native Americans in woodlands to enhance habitat for game species, to improve access for hunting and gathering of acorns, to reduce insect pest populations, and to maintain plant materials in an appropriate growth form for crafts (]epson 1910; Cooper 1922; Anderson 2005). However, it is difficult to document the frequency, intensity, and extent of burning by Native Americans. The first European settlers continued to use fire as a management practice to keep stands open for livestock production and to encourage forage production. Surveys indicated oak woodland burning intervals of 8-15 years by ranchers (Sampson 1944). Local prescribed burning associations were set up in various locations around the state, where neighbors came together annually to help conduct burns in the highest priority areas. The use of prescribed burning as a management tool, to mimic the effects of nature, ceased on the state's conifer forest lands in the early 1900s. However, ranchers continued the extensive use of prescribed burning on oak woodlands until the 1950s. Since then, fire use declined, driven by neg-

330

OAK WOODLANDS AND FORESTS

ative urban attitudes toward fire, increasing housing density in rural areas of the state, and concerns about liability and air quality. Fire suppression eventually became the standard management policy. McClaran and Bartolome (1989) evaluated fire frequency in Sierran foothill oak woodlands. Fire-return interval was around 25 years prior to settlement by Europeans in the mid-1800s. After settlement, the fire-return interval shortened to 7 years. No fires were observed from 1950 to the mid-1980s, when fire suppression was the dominant practice. Stephens (1997) observed similar fire-return intervals in the Sierra Nevada. Shorter fire-return intervals in the past may have created conditions more conducive for oak regeneration. McClaran and Bartolome (1989) compared oak stand age structure with fire history and showed that oak recruitment was associated with fire events. Most oak recruitment in their Sierran foothill study area occurred during periods of high fire frequency in the 1880s to 1940s. Oak recruitment has been rare since fire suppression. The factors leading to enhanced oak regeneration from higher fire frequencies are not entirely clear. Allen-Diaz and Bartolome (1992) evaluated blue oak seedling establishment and mortality with grazing and prescribed burning treatments in coastal oak woodlands dominated by blue oak. Neither of these treatments significantly affected oak seedling density nor the probability of mortality, when compared to unburned and ungrazed areas. Lawson (1993) evaluated prescribed fire effects on coast live oak and Engelmann oak in southern California and found higher seedling mortality in areas of prescribed fire. Perhaps the importance of fire for oak regeneration is explained by enhanced postfire oak sprout growth documented by Bartolome and McClaran (1989). Other factors affecting oak regeneration, which would be influenced by the timing of fire, include modifications to the seedbed, decreased competition for moisture from herbaceous species, and the size of wildlife populations that feed on acorns and seedlings. Fire also has an effect on oak woodland stand structure and composition. Lawson (1993) showed differential postfire effects on coast live and Engelmann oaks, coast live oak having a higher mortality and Engelmann oak having greater height growth following fire. The thicker bark of Engelmann oak provided more protection. Declines in Engelmann oak habitats in Southern California might be mitigated by reintroduction of fire. Fry (2002) found high survival of blue oaks following prescribed burning in central coastal California, and Horney et al. (2002) and Dagit (2002) documented high blue and coast live oak survival after wildfire. Fire also kills diseases and pests, such as the filbert weevil (Cllcllrlio occidentalis) and the filbert worm (Melissopus lati(erreanus), which can infest acorns (Lewis 1991). Fire also reduces fuel ladders under oak canopies, preventing highintensity crown fires.

160

'1 140

I

'"C G>

...., 120 u

:.c; ~

CU)

•

100

Amphibians

G>

'uG>

0-

U)

Birds

80

"l-

e

G>

..0

60

•

Mammals

g

Reptiles

E ::J

Z

40 20 0 Blue oak

FIG U RE 1 2.7

Blue oak/pine

Valley oak

Coastal Oak

Montane Hardwd.

Total

Number of vertebrate species by Wildlife Habitat Type (adapted from Mayer and Laudenslayer 1988).

Wildlife Habitat and Biodiversity Processes

California's oak woodlands provide habitat for over 300 vertebrate species, more than 2,000 plant species, and an estimated 5,000 species of insects. Figure 12.7 shows vertebrate wildlife diversity predicted by the California Wildlife Habitat Relationships (CWHR) model for the five major oak woodland communities (Mayer and Laudenslayer 1988). In a 3-year study of nongame wildlife populations in the Sierra Nevada, Block and Morrison (1990) found 113 bird species (at least 60 of which bred at the site), much of the bird species richness being directly related to plant richness. Hutton's vireos (Vireo huttoni), orange-crowned warblers (Vennivora celata), and Wilson's warblers (Wilsonia pusilia) are closely associated with interior live oak. White-breasted nuthatches (Sitta carolinensis) and western bluebirds (Sialia mexicana) are closely associated with blue oak (Block 1990). The specific habitats utilized by the birds change seasonally. For example, many resident woodland birds obtained insects from the foliage of blue and interior live oaks during the breeding season and shifted to live oaks when the blue oaks were leafless. Favorable oak woodland habitats supply food, water, and cover to sustain wildlife species. The absence of a particular element in a habitat may limit species diversity. Important habitat elements in oak woodlands include riparian zones, vernal pools, wetlands, dead and downed logs, and other woody debris, brush piles, snags, rock outcroppings, and cliffs. Riparian habitat elements are used by almost 900/0 of all oak woodland wildlife species, illustrating the importance of conserving this habitat. Over one-third of all woodland bird species use snags, suggesting that management strategies maintaining an appropriate snag density will result in greater wildlife species richness. Downed coarse woody debris provide valuable habitat for most reptiles and amphibians and

for many bird species. Oak woodland wildlife management must include trees in various stages of vigor (Block et al. 1990). Mid-elevation habitats, with several oak species, vertical diversity in vegetation structure, and diverse riparian zones, have the highest wildlife diversity (Motroni et al. 1991). Currently the threats to continued high biodiversity on oak woodlands include (1) fragmentation of large blocks of extensively managed oak woodlands; (2) reduction in important habitat elements such as snags, woody debris, and diverse riparian zones; and (3) increasing encroachment of urban areas, bringing household pets, humans, and fire suppression policies into contact with woodland habitats. These threats can be reduced by encouraging cluster housing development and maintaining connecting corridors between large oak woodland blocks (Merenlender et al. 1998). Exotic Pathogens

Beginning in 1995, tanoaks (Lithocarpus densiflorus) in coastal forests and woodlands showed widespread and unexplained mortality. By 1998, similar patterns were noted in coast live oak and California black oak in the area just north and south of San Francisco Bay (Svihra 1999). In 2000, the causal agent for this disease, popularly known as "sudden oak death" (SOD), was determined to be a pathogen new to science, Phytophthora ramorum (Rizzo and Garbelotto 2003). This same pathogen has since been isolated from ornamental rhododendrons in Europe and coastal California (Werres et al. 2001). Mortality is currently widespread on coast live and black oak species, as well as on tanoak throughout coastal California. This pathogen is apparently an introduced organism, based on preliminary evaluations of its genetic structure (Garbelotto et al. 2002). The mechanisms for the spread

OAK WOODLANDS AND FORESTS

331

FIGURE 12.8 Typical downed coast live oak with Phytophthora ramorum (Sudden Oak Death) infection.

and infection biology are currently unknown. It is known that P. ramorum is found on the leaves of a wide variety of plant species in coastal oak woodlands, and it has also been isolated in rain splash, watercourses, and soil that may contribute to its movement (Davidson et al. 2002) Sudden oak death represents a major threat to ecosystem health of California's oak woodlands (Fig. 12.8). Work on silvicultural and arboricultural management techniques are underway. The potential for genetic resistance is being evaluated, as well as the risk factors for various woodland stand structures and site conditions.

Economic Values and Utilization of Oak Woodlands Oak woodlands have been important to humans living in California for centuries. Recent trends in human use, however, are leading to conversion from ranching to agriculture, residential development, and industry. Some of the economic and utilization issues facing oak woodlands are discussed later. The original oak woodland human inhabitants were Native Americans. Acorns were the dietary staple and sustained the cultures of those that lived among the oak woodlands (Pavlik et al. 1991). Virtually all tribes west of the Sierra Nevada harvested acorns for food. Acorns are estimated to have been the primary diet for more than three fourths of all Native Americans in California (McCarthy 1993). Black oak was the preferred species in many regions. Each tribe had special mechanisms for acorn gathering, storing, hulling, drying, leaching, pounding, and cooking. The bark, roots, wood, small branches and galls of oaks were also utilized. Acorns were second to salt among the most frequently traded foods or condiments among Native Americans. The trade in acorns flowed from west to east. For example, Miwoks gathered black oak acorns from the western Sierra and trading with the Mono Lake Paiute for pinyon pine

332

OAK WOODLANDS AND FORESTS

nuts (Pavlik et al. 1991). Trading across elevational zones was also common (McCarthy 1993). Territorial claims of tribes, villages, families, and individuals were often based on the distribution of acorn-producing oak groves. The fact that many cultural traditions and celebrations focused on oaks attests to the central role oaks played. Oaks and acorns were also used for medicines and dyes. Fire was the most prevalent and effective management tool native Californians used to manage oaks and acorn crops (McCarthy 1993). Low-intensity fires were also used to promote oak growth, to reduce the probability of damaging high intensity fires, and to help keep prized oaks from being overtopped by conifer species. Many Village sites were found to be located near mature black oak stands. Grazing animals have been part of California grassland, savanna, and woodland ecosystems for thousands of years (Edwards 1992). Before European contact and the establishment of Widespread cattle and sheep grazing (Burcham 1957), large herds of pronghorn, tule elk, mule deer, and rodents grazed these grasslands, savannas, and wetlands (Edwards 1997). Many grassland species have habits (e.g., prostrate growth) and structures (e.g., basal meristems) that mitigate the damage from grazing or surface fire (Briske 1991). Since the 1800s, California's oak woodlands have been used mainly for livestock grazing. Today, two thirds of California's oak woodlands are grazed by livestock (Ewing et al. 1988; Huntsinger and Fortmann 1990). Dramatic annual fluctuations in livestock markets, coupled with risk from forage shortages due to high variability in annual rainfall, make many livestock operations marginal. Uncertainty about federal grazing policies (many rangeland operators lease summer forage on Federal land) also hinders economic viability of oak woodland livestock enterprises (Sulak and Huntsinger 2002). Low profitability and high risk have accelerated conversion of extensively managed private ranches to suburban developments and intensive agriculture (e.g., Vineyards). Traditional efforts to increase profitability of oak woodlands have focused almost exclusively on enhancing forage production through removal of oaks (George 1987). This simplification of the ranch ecosystems did pay short-term dividends in improved forage yields, but the same risk from fluctuating markets and weather continued to make ranching a low profitability enterprise. New markets have developed in the last 20 years for oaks for firewood, furniture (as new techniques have allowed for economic utilization of small-diameter logs), and as habitat for commercial hunting enterprises. This diversified economic portfolio has helped to enhance the economic sustainability of these areas by spreading risk out over several enterprises, increasing overall returns per acre, and proViding an economic incentive to conserve more diverse woodlands (Standiford and Howitt 1992,1993). Diversified markets have reduced tree harvesting and intensity of livestock use. Historically, the market value of oak woodlands for subdivisions near urban areas has exceeded their value for

alnenities and ecological functions. Recent human population increase in these areas, however, has raised the potential value of woodland amenities to a point where they may be a financially viable alternative to land development (Scott 1996). Woodlands provide a large component of the quality-of-life sought by many relocating industries, and the relatively low cost of industrial sites in these woodlands is equally appealing. Woodland owners along the wildland-urban interface often find that their management options track public demand for specific values. If woodland conversions trigger a public demand for amenity protection, the solutions typically must be found on private lands. Open space easelnents and other deed restrictions provide financial, tax, or development incentives for the voluntary maintenance of public amenity values on private lands. Mitigation banking provides another economic value for hardwood rangeland in urban interface areas.

Areas for Future Research

predominant land use. Where individual landowners have the ability to implement management activities that affect large acreage, education and research has contributed to decisions that favor conservation of oak woodlands. However, for much of California, conversion of oak woodland habitats to urban or suburban use is having the largest impact on sustainability of resource values. IncorporatiOjn of ecologically-based material into land use plans adopted by the county government is only beginning. Since conversion to residential and industrial uses is ultimately a land-use decision, it is a political process involving action by elected officials with input from different constituencies. The political and economic forces vary greatly in different parts of the state. Since" success" in this area involves multiple individuals agreeing on a political course of action, this issue will present a large challenge.

References Adams, T. E., P. B. Sands, W. H. Weitkamp, and N. K. McDougald.

Research must address political and economic forces shaping human conversion of oak habitats as well as biological issues related to long-term habitat sustainability under pressures from climate change, invasive species (e.g. SOD), and regeneration problems. Research focused at the wholeecosysteln scale, including development of viable economic alternatives for land use, will be most productive in ensuring the long-term preservation of California oak savannas. Much of the recent work in oak savanna has focused on reproduction, incorrectly assuming that it is synonymous with recruitment. Future research will have to focus on recruitment success, which appears to vary with species, past stand structure, and tree mortality, and seems to vary at multiple spatial and temporal scales. Management of savanna understory species composition and productivity has been assumed to fit the residual drymatter model, which has been successfully applied to grasslands (George et a1. 1985). This model assumes that if land managers leave recommended levels of mulch at the end of the dry season, the best combination of positive effects to the grassland ecosystem as a whole will result, such as optimum forage production, optimum biodiversity, optimum protection from soil erosion, and optimum wildlife habitat. However, there is little research to back up this assumption. Because of the highly variable nature of canopy effects on understory composition and productivity, many more sitespecific studies are needed to fully examine ecosystem response and management options. Protection of ecosystem services is of concern due to rapid population growth, and the resulting conversion and fraglnentation of woodland habitats. A large number of counties have started the process of adopting local conservation strategies to conserve oak woodlands. Education and research have played a major role in conservation. Major accomplishments have been made in rural areas of the state, where livestock and natural resource management are the

1992. Oak seedling establishlnent on California rangelands. Journal of Range Management. 45(1): 93-98. Allen, B., Evett, R., Holzn1an, B. and Martin, A. 1989. Rangeland Cover Type Descriptions for California Hardwood Rangelands. A report for California Department of Forestry and Fire Protection. Sacramento, CA. AlIen, B. H., Holzman, B. A., and Evett, R. R. 1991. A classification system for California's hardwood rangelands. Hilgardia 59(2): 1-45. Allen-Diaz, B. H., and Holzrnan, B. 1991. Blue oak comInunities in California. Madroiio 38: 80-95. Allen-Diaz, B. H., and Bartolome,]. W. 1992. Survival of Qllerclls dOllglasii (Fagaceae) seedlings under the influence of fire and grazing. Madroiio 39: 47-53. Allen-Diaz, B. H., and R. D. Jackson. 2000. Grazing effects on spring ecosystem vegetation of California's hardwood rangelands. ] Range Manage. 53: 215-220. Allen-Diaz, B., R. ]ackson, ]. Bartolome, et al. 2004. Grazing management of spring-fed wetlands in Californian oak woodlands: summary of results of a long tenn study. California Agriculture 58: 144-148. Allen-Diaz, B., Bartolome,]. W., and McClaran, M. P. 1999. California Oak Savanna. Chapter 20 in: R. C. Anderson, J. S. Fralish, and J. M. Baskin (eds.), Savannas, barrens, and rock outcrop plant COIlHnunities of North America. Cambridge University Press. Allen-Diaz, B. H. and]. W. Bartolome. 1992. Survival of QllerclIs dOllglasii (Fagaceae) seedlings under the influence of fire and graZing. Madroiio 39(1): 47-53. Anderson, M. K. 2005. The use of fire by Native Americans in California in: N. G. Sugihara,]. W. van Wagtendonk,]. Fites-Kaufrnan, et al., (eds.), Fire in California ecosystems. University of California Press, Berkeley. Barbour, M. G., and]. Major (Eds.) 1988. Terrestrial vegetation of California. Wiley, New York. 1002 pp. Barbour, M. G., and R. A. Minnich. 2000. Californian upland forests and woodlands. Pages 161-202 in: M. G. Barbour and W. D. Billings (eds.), North American terrestrial vegetation. 2nd edition, Cambridge University Press, New York. Barrett, R. H. 1980. Mammals of California oak habitats-Il1anageInent implications. Pages 275-291 in: Proceedings of the SyInposiuln on

OAK WOODLANDS AND FORESTS

333

the Ecology, Management, and Utilization of California Oaks, June 26-28,1979. USDA Forest Service General Technical Report PSW-44. Bartolome,]. W., and B. Gemnlill 1981. The ecological status of Stipa plllchra (Poaceae) in California. Madroflo 28: 172-184. Bartolome,]. W., S.E. Klukkert, and W.J. Barry. 1986. Opal phytoliths as evidence for displacement of native Californian grassland. Madroflo 33: 217-222. Bartolome. J. W., P. C. Muick, and M. P. McClaran. 1987. Natural regeneration of Californian hardwoods. Pages 26-31 in: Proceedings of the Symposium on Multiple-use Management of California's Hardwood Resources; 1986 San Luis Obispo, Calif. USDA Forest Service, Pacific Southwest Forest and Range Experiment Station, Berkeley, California. General Technical Report PSW-100. Bartolome, ]. W. 1987. California annual grassland and oak savannah. Rangelands 9(3): 122-125. Bartolome, J. W., B. H. Allen-Diaz, and W. Tietje. 1993. The effect of Qllerclls dOllglasii removal on understory yield and composition. Journal of Range Management 47(2): 151-154 Bartolome, ]. W., and M. P. McClaran. 1992. Composition and productivity of California oak savanna seasonally grazed by sheep. Journal of Range Managenlent 45: 103-107. Beetle, A. A. 1947. Distribution of the native grasses of California. Hilgardia 17: 309-357. Belsky, A.]., A. Matzke, and S. Uselman. 1999. Survey of livestock influences on stream and riparian ecosystems in the western United States. Journal of Soil and Water Conservation 54: 419-431. Bernhardt, E. A., and T.]. Swiecki, 1991. Minimum input techniques for valley oak restocking. USDA Forest Service General Technical Report PSW-126: 2-8. Bolsinger, C. L. 1988. The hardwoods of California's timberlands, woodlands, and savannas. USDA Forest Service Pacific Northwest Research Station Resource Bulletin PNW-RB-148: 149. Block, W. M. 1990. Geographic variation in foraging ecologies of breeding and nonbreeeding birds in oak woodlands. Studies in Avian Biology 13: 264-269. Block, W. M., and M. Morrison, 1990. Wildlife diversity of the central Sierra foothills. California Agriculture 44(3,4): 19-22. Block, W. M., M. L. Morrison, and]. Verner. 1990. Wildlife and oak woodland interdependence. Fremontia 18: 72-76. Borchert, M. 1., F. W. Davis,]. Michaelsen, and L. D. Oyler. 1989. Interactions of factors affecting seedling recruitment of blue oak in California. Ecology 70: 389-404. Briske, D. D. 1991. Developmental morphology and physiology of grasses. Pages 85-108 in: Grazing management: An ecological perspective. Heitschmidt, R. K. and ]. W. Stuth (eds.). Timber Press, Portland OR. Burcham, L. T 1970. Ecological significance of alien plants in California grasslands. Pages 36-39 in: Proceedings of the Association of American Geographers. 2. Byrne, R., E. Edlund, and S. Mensing. 1991. Holocene changes in the distribution and abundance of oaks. Pages 182-188 in: Proc. Symp. Oak Woodlands and Hardwood Rangeland Management. USFS General Technical Report PSW-126. Camping, TJ. R. A. Dahlgren, K. W. Tate, et al. 2002. Changes in soil quality due to grazing and oak tree removal in California blue oak woodlands. Pages 75-85 in: Proceedings of the Fifth Symposium on Oak Woodland: Oaks in California's Changing Landscape, October 22-25,2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184. Clary, W. P. 1995. Vegetation and soil responses to grazing simulation on riparian meadows. Journal of Range Management 48: 18-25.

334

OAK WOODLANDS AND FORESTS

Clements, F. E. 1934. The relict method in dynamic ecology. Journal of Ecology 22: 39-68. Cooper, W. S. 1922. The broad-sclerophyll vegetation of California. Carnegie Institute, Washington Publication 319. 124pp. Danielsen, K. C. and W. L. Halvorson, 1991. Valley oak seedling growth associated with selected grass species. USDA Forest Service General Technical Report PSW-126: 9-11. Dagit, R. 2002. Post-fire monitoring of coast live oak (Quercus agri-

folia) burned in the 1993 Old Topanga fire in: Proceedings of the Fifth Symposium on Oak Woodland: Oaks in California's Changing Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184: 243-249. Dahlgren, R. A., M.]. Singer, and X. Huang. 1997. Oak tree and grazing impacts on soil properties and nutrients in California oak woodlands. Biogeochemistry 39: 45-64. Davidson, E. A.,]. M. Stark, and M. K. Firestone. 1990. Microbial production and consumption of nitrate in an annual grassland. Ecology 71: 1968-1975. Davidson, E. A., D.]. Herman, A. Schuster, et al. 1993. Cattle grazing and oak trees as factors affecting soil emissions of nitric oxide from an annual grassland. Pages 100-119 in: Harper, L. A., G. Peterson, P. S. Baenziger, and R.J. Luxmoore (eds.), Agricultural ecosystem effects on trace gases and global climate change: Proc. of a symposium. ASA Special Pub I. No. SS. Madison, WI: American Soc. Agronomy, Crop Science Soc. America, Soil Science Soc. America. Davidson,]. M., D. M. Rizzo, M. Garbelotto, et al. 2002. Phytophthora ramorunl and sudden oak death in California: 11. Transmission and survival. Pages 741-749 in: Proceedings of the Fifth Symposium on Oak Woodland: Oaks in California's Changing Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184. Davis, F. W. 1995. Vegetation change in blue oak and blue oak/foothill pine woodland. Report to California Department of Forestry and Fire Protection. Contract 8CA06673. 34 P 8CA4215 1 (available from CDF-FRAP, 1920 20th St., Sacramento, CA 95814). Edwards, S. W. 1992. Observations on the prehistory and ecology of grazing in California. Fremontia 20: 3-11. Edwards, S. W. 1997. A Rancholabrean-Age, Latest-Pleistocene Bestiary for California Botanists. Four Seasons 10(2): 5-34. Epifanio, C. R., M.J. Singer, X. Huang. 1991. Hydrologic impacts of oak harvesting and evaluation of the modified universal soil loss equation. Pages 189-193 in: Proceedings of the Symposium on Oak Woodlands and Hardwood Rangeland Management. USFS General Technical Report PSW-126. Davis, F. W., M. Borchert, L. E. Harvey, et al. 1991. Factors affecting seedling survivorship of blue oak (Quercus dOllglasii) in Central California. Pages 81-86 in: Proc. Symp. on Oak Woodlands and Hardwood Rangeland Management. Tech. Coord., Standford, R. B. USDA Forest Service Gen. Tech. Rep. PSW-126. Ewing, R. A., N. Tosta, R. Tuazon, et al. 1988. California's forests and rangelands: growing conflict over changing uses. California Department of Forestry and Fire Protection, Forest and Rangeland Resources Assessment Program, Anchor Press, Sacramento. Eyre, F. H. 1980. Forest cover types of the United States and Canada. Society of American Foresters, Washington, D.C. Firestone, M. K. 1995. Nutrient cycling in managed oak woodlandgrass ecosystem. Final report to the Integrated Hardwood Range Management Program. 56p. Fleischner, T L. 1994. Ecological costs of livestock grazing in western North America. Conservation Biology 6(3): 629-644.

Frank, D. A., and S.]. McNaughton. 1993. Evidence for the I--)fomotion of aboveground grassland production by native large herbivores in Yellowstone National Park. Oecologia 96: 157-161.

Griffin, J. R. 1971. Oak regeneration in the upper Cannel Valley, California. Ecology 52(5): 863-868. Griffin, ]. R. 1973. Valley oaks- the end of an era? Fremontia 1: 5-9.

Frank, D. A., R. S. Inouye, N. Huntly, et a1. 1994. The biogeochelnistry of

Griffin, ]. R. 1976. Regeneration in Qllerclls lohata savannas, Santa

north-ten1perate grassland with native ungulates: nitrogen dynamics

Lucia mountains, California. American Midland Naturalist 95(2):

in Yellowstone National Park. Biogeochemistry 26: 163-188.

424-435.

Frank, D. A., and R. D. Evans. 1997. Effects of native grazers on grass-

Griffin, J.R. 1977. Oak woodland. Pages 382-415 in: M.G. Barbour

land N cycling in Yellowstone National Park. Ecology 78: 2238-2248.

and]. Major (eds.), Terrestrial vegetation of California. Wiley,

FRAP. 2003. Forest and Range Assessment. The Changing California. Assesslllent SUlnmary. October 2003. California Department of

New York. Griffin, ]. R. 1980. Sprouting in fire dan1aged valley oaks, Chews

+ Appendices.

Ridge, California. Pages 216-219 in: Proceedings of the Sylllposium

Frost, W. E., and N. K. McDougald. 1989. Tree canopy effects on

on the Ecology, Management and Utilization of California Oaks.

Forestry and Fire Protection. Sacramento, CA. 198 pp.

herbaceous production of annual rangeland during drought.

Pacific Southwest Forest and Range Experiment Station, Forest

Journal of Range Management 42: 281-283.

Service, USDA. General Technical Report PSW-44.

Frost, W. E., and S. B. Edinger. 1991. Effects of tree canopies on soil characteristics of annual rangeland. Journal of Range Management 44(3): 286-288. Frost, W. E.,

J. W. Bartolome, and]. M. Connor. 1997. Understory-

Griffin, ]. R., and W. B. Critchfield. 1972. The distribution of forest trees in California. USDA Forest Service Research Paper PSW-82. Grime, ]. P. 1979. Plant strategies and vegetation processes. Wiley, New York.

canopy relationships in oak woodlands and savannas. Pages

Hall, L. M., M. R. George, D. D. McCreary, et al. 1992. Effects of cattle

183-190 ill: Proceedings of a Symposium on Oak Woodlands:

grazing on blue oak seedling damage and survival. Journal of Range

Ecology, Management, and Urban Interface Issues. USDA Forest Service General Technical Report PSW-GTR-160. Fry, D. 2002. Effects of a prescribed fire on oak woodland stand structure. Pages 235-242 in: Proceedings of the Fifth SylnposiuTI1

Management 45(5): 503-506. Hamilton, ]. G. 1998. Changing perceptions of pre-European grasslands in California. Madrofio 44: 311-333. Harris, R. R., and S. D. Kocher. 2002. Oak management by county

on Oak Woodland: Oaks in California's Changing Landscape,

jurisdictions in the central Sierra Nevada, California. Pages 464-472

October 22-25,2001, San Diego, CA. USDA Forest Service General

in: Proceedings of the Fifth Syn1posium on Oak Woodland: Oaks in

Technical Report PSW-GTR-184. Garcia, L. v., T. Maranon, A. Moreno, et a1. 1993. Above-ground bioIllass and species richness in a Mediterranean salt marsh. Journal of Vegetatic)n Science 4: 417-424.

California's Changing Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184. Harris, R. W., A. T. Leiser, and R.E. Fissel1. 1980. Tolerance of oaks to flooding. Pages 238-241 in: Proceedings of the Symposium on the

Garbelotto, M., D. M. Rizzo, K. Hayden, et a1. 2002. Phytophthora

Ecology, Management and Utilization of California Oaks. Pacific

ranl0run1 and sudden oak death in California: Ill. Preliminary stud-

Southwest Forest and Range Experiment Station, Forest Service,

ies in pathogen genetics. Pages 765-774 in: Proceedings of the Fifth Sylnposium on Oak Woodland: Oaks in California's Changing

USDA. General Technical Report PSW-44. Heady, H.E 1977. Valley grassland. Pages 491-513 in: M.G. Barbour

Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Ser-

and ]. Major (eds.), Terrestrial Vegetation of California. Wiley,

vice General Technical Report PSW-GTR-184.

New York.

Garrison, B. 1996. Vertebrate Wildlife Species and Habitat Associa-

Heady, H. E,]. W. BartoloTI1e, M. D. Pitt, et al. 1992. California prairie.

tions in: Standiford, R. B. and P. Tinnin (eds.), Guidelines for man-

Pages 313-332 in: R.T. Coupland (ed.), Natural grasslands. Elsevier,

aging California's hardwood rangelands University of California

New York.

Division of Agriculture and Natural Resources Leatlet no. 3368.

Holland, V. L. 1980. Effect of blue on rangeland forage production in

Garrison, B., and R. B. Standiford. 1996. Chapter 2: Oaks and Habi-

central California. Pages 314-318 in: Proceedings of the Sympo-

tats of the Hardwood Rangeland in: R. B. Standiford and P. Tinnin

siUln on the Ecology, Managelnent and Utilization of California

(eds.), Guidelines for managing California's hardwood range-

Oaks. Pacific Southwest Forest and Range Experiment Station,

lands. University of California Division of Agriculture and Nat-

Forest Service, USDA. General Technical Report PSW-44.

ural Resources Leaflet no. 3368.

Holland, V. L. 1988. Coastal oak woodland in: K. E. Mayerand W.E

(iasith, A., and V. H. Resh. 1999. Strealns in Mediterranean climate

Laudenslayer, .If. (eds.), 1988. A guide to wildlife habitats of Califor-

regions: abiotic influences and biotic responses to predictable sea-

nia. California Dept. of Forestry and Fire Protection, Sacran1ento

sonal events. Ann Review Ecology and Systematics 30: 51-81.

(available from CDF-FRAP, 1920 20th St., Sacralnento, CA 95814).

George. M. 1987. Management of hardwood range: a historical

Holland, V. L., and]. Morton. 1980. Effect of blue oak on nutritional

review. Range Science Report 12, Agronomy and Range Science,

quality of rangeland forage in central California. Pages 319-322

u.e. Davis. 15pp.

in: T. R. Plumb (ed.), Proceedings of sylnposium on the ecology,

Cieorge, M. R., W.J. Clawson,]. W. Menke, and J. W. Bartolome. 1985.

management, and utilization of California oaks; 1979 June 26-28;

Annual grassland forage productivity. Rangelands 7(1): 17-19.

Claremont, CA. USDA Forest Service, Pacific Southwest Forest and

Gordon, D. R.,]. M. Welker,]. W. Menke, et a1. 1989. Competition for soil water between annual plants and blue oak seedlings. Oecologia 79(4): 533-541. (ireenwood, G. B., R. K. Marose, J. M. Stenback. 1993. Extent and ownership of California's hardwood rangelands. Prepared for Strategic Planning Program, California Dept. of Forestry and Fire Protection unpublished report. 8CA421 5 1 (available from CDFFRAP, 1920 20th St., Sacramento, CA 95814).

Range Experiment Station, Berkeley, California. General Technical Report PSW-44. Holmes, T. H. 1990. Botanical trends in Northern California Oak Woodland. Rangelands 12(1): 3-7. Holstein,G. 2001. Pre-agricultural grassland in Central California. Madrofio 48: 253-264. Holzman, B. A., and B. H. Allen-Diaz. 1991. Vegetation change in blue oak woodlands in California. Pages 189-193 in: Proceedings

OAK WOODLANDS AND FORESTS

335

of Symposiunl on Oak Woodlands and Hardwood Rangeland Managernent. USFS General Technical Report PSW-126. Holzman, B. A. 1993. Vegetation change in California's blue oak woodlands 1932-1992. PhD dissertation. University of California, Berkeley. Horney, Iv!., R. B. Standiford, D. McCreary, et al. 2002. Effects of wildfire on blue oak in the northern Sacralnento valley. Pages 261-267 ill: Proceedings of the Fifth Symposium on Oak Woodland: Oaks in California's Changing Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184. Huntsinger, L. and L. P. Fortnlann. 1990. California's privately owned oak woodlands: Owners, use, and management. Journal of Range Management 42(3): 147-152. Huntsinger, L. 1997. California's oak woodlands revisited: changes in owners, use, and managernent, 1985 to 1992. Pages 626-630 ill: Proceedings of the Sylnposium on Oak Woodlands: Ecology, Managenlent, and Urban Interface Issues ... Pacific Southwest Research Station, Forest Service, USDA. General Technical Report PSW-162. Huntsinger, L., and P. Hopkinson. 1996. Sustaining rangeland landscapes: a social and ecological process. Journal of Range Managelnent 49(2): 167-173. Jackson, R. D., and J. W. Bartolome. 2002. A state-transition approach to understanding nonequilibrium plant conlmunity dynamics in Californian grasslands. Plant Ecology 162: 49-65. Jackson, L. E. 1985. Ecological origins of California's Mediterranean grasses. Journal of Biogeography 12: 349-361. Jackson, L. E., R.B. Strauss, M.K. Firestone, et a1. 1988. Plant and soil nitrogen dynamics in California annual grassland. Plant and Soil 110: 9-17 . .I ackson, L. E., R. B. Strauss, M. K. Firestone, et a1. 1990. Influence of tree canopies on grassland productivity and nitrogen dynamics in deciduous oak savanna. Agriculture, Ecosystelns and Environment 32: 89-105. Jansen, H. C. 1987. The effect of blue oak renloval on herbaceous production on a foothill site in the northern Sierra Nevada. Pages 343-350 ill: Proceedings of the Symposiunl on Multiple-Use Management of California's Hardwood Resources. Pacific Southwest Forest and Range Experiment Station, Forest Service, USDA. General Technical Report PSW-I00. Jensen, D. B., M. Tom, and J. Harte. 1990. In Our Own Hands: A Strategy for Conserving Biological Diversity in California. California Policy Seminar Research Report. University of California Press, Berkeley. Jepson, W. L. 1910. The silva of California. University of California Mem., Vol. 2. 480pp. Kay, B. L. 1987. Long-tenn effects of blue oak removal on forage production, forage quality, soil and oak regeneration. Pages 351-357 in: Proceedings of the Symposium on Multiple-use Management of California's Hardwood Resources; 1986 San Luis Obispo, Calif. USDA Forest Service, Pacific Southwest Forest and Range Experiment Station, Berkeley, California. General Technical Report PSW-1 00.

Kuchler, A. W. 1988. The map of natural vegetation of California. Dept. Geography. University of Kansas in: M. G. Barbour and J. Major (eds.), 1988. Terrestrial Vegetation of California. California Native Plant Society, Special Publication, Number 9. Lawson, D. M. 1993. The effects of fire on stand structure of mixed Qllerclls agritc)/ia and Q. engehnannii woodlands. Unpublished MS Thesis, San Diego State University. 122pp. Lewis, V. 1991. The temporal and spatial distribution of filbert weevil infested acorns in an oak woodland in Marin County, California. Pages 156-160 in: Proceedings of Symposium on Oak Woodlands and Hardwood Rangeland Management. USFS General Technical Report PSW-126. Light, R. H., and L. E. Pedroni. 2002. When oak ordinances fail: unaddressed issues of oak conservation. Pages 483-500 in: Proceedings of the Fifth Symposiurn on Oak Woodland: Oaks in California's Changing Landscape, October 22-25,2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184. Maranon, T., and J. W. Bartolome. 1994. Coast live oak (Quercus agritc)/ia) effects on grassland biomass and diversity. Madrofio 41: 39-52. Mayer, K. E., and W. F. Laudenslayer, .If. eds. 1988. A Guide to Wildlife Habitats of California. California Dept. of Forestry and Fire Protection, Sacramento (available from CDF-FRAP, 1920 20th St., Sacramento, CA 95814). McBride, J. 1974. Plant succession in the Berkeley Hills, California. Madrofio 22: 317-380. McCarthy, H. 1993. Managing oaks and the acorn crop. Pages 213-228 in: Before the wilderness: environmental management by native Californians, Thomas C. Blackburn and Cat Anderson, (eds). A Ballena Press, Menlo Park. McClaran, M. P., and Bartolome, J. W. 1985. The importance of oak to ranchers in the California foothill woodland. Rangelands 7(4): 158-161. McClaran, 1v1. P., and J. W. Bartolome. 1989. Fire-related recruitment in stagnant Qllerclls dOllglasii populations. Canadian Journal of Forest Research 19: 580-585. McClaran, M. P., and J. W. Bartolorne. 1989a. Effect of Qllercus dOllglasii (Fagaceae) on herbaceous understory along a rainfall gradient. Madrofio 36: 141-153. McCreary, D. D. 1989. Emergence and growth of Blue and Valley Oak seedlings as affected by acorn sowing date. Pages 200-204 in: Proceedings of the 10th North American Forest Biology Workshop, Vancouver, B. C. Canada. McCreary, D. D. 1990. Acorn sowing date affects field performance of Blue and Valley Oaks. Tree Planters Notes 41(2): 6-9. McCreary, D. D. 2001. Regenerating rangeland oaks in California. University of California Div. of Agric. And Nat. Res. Publication 21601. 62p. McCreary, D. D. 1995. Augering and fertilization stimulate growth of blue oak seedlings planted from acorns, but not from containers.

Kauffman, J. B., W. C. Krueger, and M. Vavra. 1983. Impacts 6(cattle

Western Journal of Applied Forestry 10(4): 133-137. McCreary, D. D., W. D. Tietje, R.H. Schmidt, et a1. 1991. Stump

on streambanks in northeastern Oregon. Journal of Range Man-

sprouting of blue oaks. Page 64 in: Proceedings of Symposium on

agement 36: 685-691. Kauffrnan,]. B., and W. C. Krueger. 1984. Livestock impacts on riparian ecosystelns and streamside management implications: a review.

Oak Woodlands and Hardwood Rangeland Management. USFS

Journal of Range Management 47: 430-437. Knudson, M. D. 1987. Life history aspects of Qllerclls lobata in a riparian community. Pages 38-46 in: Proceedings of Symposium on Oak Woodlands and Hardwood Rangeland Management. USFS General Technical Report PSW-100.

336

OAK WOODLANDS AND FORESTS

General Technical Report PSW-126. McCreary, D. D., W. D. Tietje, and W. E. Frost. 2002. Stump sprouting of blue oaks ten years after harvest. Pages 573-580 in: Proceedings of the Fifth Symposium on Oak Woodland: Oaks in California's Changing Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Service General Technical Report PSWGTR-184.

Sacramento (available from CDF-FRAP, 1920 20th St., Sacra-

McDonald, P. M. 1988. Montane hardwood in: K. E. Mayer and W. F. Laudenslayer, Jr. (eds.), A guide to wildlife habitats of California. California Dept. of Forestry and Fire Protection, SacraInento (available from CDF-FRAP, 1920 20th St., Sacramento, CA 95814). McDonald, P. M. 1990. Blue oak in: Silvics of North America, Volume 2, Hardwoods. USDA Forest Service Agricultural Handbook 654:

mento, CA 95814). Rizzo, D. M., M. Garbelotto, ]. Davidson, et al. 2003. Sudden oak death: endangering California and Oregon forest ecosysten1s. Front Ecol Environ 1(5): 197-204. Saenz, L., and]. O. Sawyer. 1986. Grasslands as compared to adjacent Quercus garryana woodland understories exposed to different graz-

631-639. McDonald, P. M. 1990a. California black oak. Pages 661-671 in: Silvies of North America, Volume 2, Hardwoods. USDA Forest Service Agricultural Handbook 654. Medvitz, A. G., and A. D. Sokolow. 1995. Population growth threatens agriculture, open space. California Agriculture 49(6): 11-17. Merenlender, A. M. 2000. Mapping vineyard expansion provides information on agriculture and the environment. California Agriculture 54(3): 7-12. Merenlender, A. M., K. L. Heise, and C. Brooks. 1998. Effects of subdividing private property on biodiversity in California's North Coast oak woodlands. Transactions of the Wildlife Society 34: 9-20. Mensing, S., and R. Byrne. 1999. Invasion of Mediterranean weeds

ing regimes. Madrofio 33: 40-46. Sampson, A. W. 1944. Plant succession on burned chaparral lands in northern California. California Agricultural Experiment Station Bulletin No. 685. Sawyer, J., and T. Keeler-Wolf. 1995. A manual of California vegetation. California Native Plant Society. Sacramento, CA. 471pp. Saving, S. C., and G. B. Greenwood. 2002. The potential impacts of development on wildlands in El Dorado County, California. Pages 443-461 in: Proceedings of the Fifth Symposium on Oak Woodland: Oaks in California's Changing Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184. Schimel, ]. P., L. E. ]ackson, and M. E. Firestone. 1989. Spatial and

in California before 1769. Fremontia 27(3): 6-9. Motroni, R. S., D. A. Airola, R. K. Marose, et al. 1991. Using wildlife

temporal effects of plant-microbial competition for inorganic nitrogen in a California annual grassland. Soil Biology and Bio-

species richness to identify land protection priorities in California's hardwood rangelands. Pages 110-119 in: Proceedings of Sylnposium on Oak Woodlands and Hardwood Rangeland Management. USFS General Technical Report PSW-126. Muick, P. C. 1997. Effect of shade and clipping on coast live and blue oak seedling mortality and growth in California annual grasslands. Pages 135-146 in: Proceedings of a Symposium on oak woodlands: ecology, management, and urban interface issues. USDA Forest Service General Technical Report PSW-GTR-160. Muick, P. C., and]. R. Bartolorne. 1987. An assessment of natural

Scott, T. 1996. Chapter 7: Open Space and Private Land Opportunities for Hardwood Rangeland Conservation in: R. B. Standiford and P. Tinnin. 1996. Guidelines for managing California's hardwood rangelands. University of California Division of Agriculture and Natural Resources Leaflet no. 3368. Scott, T. A., and N. Pratini. 1997. Edge effects and recreational impacts on a population of woodland birds in: Proceedings of a Symposium on Oak woodlands: Ecology, Management, and Urban Interface Issues, March 19-22, 1996, San Luis Obispo, CA.

regeneration of oaks in California. Report submitted to the Forest and Rangeland Assessment Program, California Department of Forestry and Fires Protection, in partial fulfillment of contract #8CA42136. 101pp. (8CA42151 (available from CDF-FRAP, 1920 20th St., Sacramento, CA 95814). Munz, P. A., and D. D. Keck. 1973. A California with Supplement. University of California Press, Berkeley. Neal, D. L. 1980. Blue oak-digger pine. Pages 126-127 in: Forest cover types of the United States and Canada, F. H. Eyre (ed.) Society of American Foresters, Washington, D.C. Pavlik, n., P. Muick, S. Johnson, et al. 1991. Oaks of California. CachUlna Press. Plun1b, T. R., and A. P. Gomez. 1983. Five southern California oaks: identification and postfire management. Berkeley, CA: Pacific Southwest Forest and Range Experiment Station, Forest Service, USDA General Tech. Report PSW-71. 56p. Ratliff, R. D., D. Duncan, and S. E. Westfall. 1991. California oakwoodland overstory species affect herbage understory: manageInent implications. Journal of Range Management 44(4): 306-310. Rice, K.]., and E. S. Nagy. 2000. Oak canopy effects on the distribution patterns of two annual grasses: the role of competition and soil nutrients. American Journal of Botany 87: 1699-1706. Ritter, L. V. 1988. Valley oak woodland in: K. E. Mayer and W. F. Laudenslayer, ]r. (eds.), 1988. A guide to wildlife habitats of California. California Dept. of Forestry and Fire Protection, Sacramento (available from CDF-FRAP, 1920 20th St., Sacramento, CA 95814).

chemistry 21: 1059-1066.

USDA Forest Service Research Paper PSW-GTR-160. Shlisky, A.J. 2002. Hierarchical relationships between plant species communities and their ecological constraints at multiple scales in oak woodland/annual grassland systems of the Sierra Nevada foothills, California. PhD thesis, University of California, Berkeley. 200pp. Solomeshch, A.1., and M. G. Barbour. 2006. Reconstruction of California's precontact interior grassland based on the presence of native taxa. Ecology (in press). Standiford, R. B., and R. E. Howitt. 1992. Solving empirical bioeconomic models: a rangeland management application. American Journal of Agricultural Economics May, 1992: 421-433. Standiford, R. B., N. K. McDougald, R. Phillips, et al. 1991. South Sierra oak regeneration survey. California Agriculture 45(2): 12-14. Standiford, R. B., and R. E. Howitt. 1993. Multiple use management of California's hardwood rangelands. Journal of Range Management 46: 176-181. Standiford, R. B., D. McCreary, S. Gaertner, et al. 1996. Impact of firewood harvesting on hardwood rangelands varies with region. California Agriculture 50(2): 7-12. Standiford, R. B., and P. Tinnin. 1996. Guidelines for managing California's hardwood rangelands. University of California Division of Agriculture and Natural Resources Leaflet no. 3368. Standiford, R. B. 2001. California's Oak Woodlands. Pages 280-303

Ritter, L. V. 1988a. Blue oak woodland in: K. E. Mayer and W. F.

in: W.]. McShea and W. M Healy (eds.), Oak forest ecosystems: ecology and management for wildlife. Johns Hopkins University Press, Baltimore, MD.

Laudenslayer, Jr. (eds.), 1988. A guide to wildlife habitats of California. California Dept. of Forestry and Fire Protection,

Stephens, S. L. 1997. Fire history of a mixed oak-pine forest in the foothills of the Sierra Nevada, El Dorado County, California.

OAK WOODLANDS AND FORESTS

337

Pages 191-198 in: Proceedings of a Symposium on Oak Woodlands: Ecology, Management, and Urban Interface Issues. USDA Forest Service General Technical Report PSW-GTR-160. Stewart, W. 1991. Monitoring values and practices of oak woodland decision makers on the urban fringe. Pages 174-181 in: Proceedings of Symposium on Oak Woodlands and Hardwood Rangeland Management. USDA Forest Service General Technical Report PSW-126. Sulak, A., and L. Huntsinger. 2002. The importance of federal grazing allotments to central Sierran oak woodland permittees: a first approximation. Pages 43-51 in: Proceedings of the Fifth Symposium on Oak Woodland: Oaks in California's Changing Landscape, October 22-25, 2001, San Diego, CA. USDA Forest Service General Technical Report PSW-GTR-184. Svihra, P. 1999. Tanoak and coast live oak under attack. Univ. of California Integrated Hardwood Range Management Program, Oaks In Folks 14(2): 1-2. Swiecki, T.]., E. A. Bernhardt, and C. Drake. 1997. Factors affecting blue oak sapling recruitment. Pages 157-168 in: Proceedings of a Symposium on Oak Woodlands: Ecology, Management, and Urban Interface Issues. USDA Forest Service General Technical Report PSW-GTR-160. Tate, K. W., R.A. Dahlgren, M.]. Singer, et al. 1999. On California rangeland watersheds: timing, frequency of sampling affect accuracy of water quality monitoring. California Agriculture 53: 44-48. Tecklin, ].,]. M. Connor, and D. D. McCreary. 2002. Rehabilitation of an oak planting project on cleared rangeland using treeshelters and grazing: a ten-year saga. Fifth Symposium on Oak Woodlands: Oaks in California's Changing Landscape, October 22-25, 2001. San Diego, CA. USDA Forest Service Gen Techn Rep PSWGTR-184. Thompson, K. 1961. Riparian forests of the Sacramento Valley, California. Ann Assoc Amer Geog 51: 294-315.

338

OAK WOODLANDS AND FORESTS

Thornburgh, D. A. 1990. Canyon live oak. Pages 618-624 in: Silvics of North America, Volume 2, Hardwoods. USDA Forest Service Agricultural Handbook 654. Tietje, W. D., R. H. Barrett, E. B. Kleinfelder, et al. 1991.Wildlife diversity in valley-foothill riparian habitat: North central versus central coast California. Pages 120-125 in: Proceedings of Symposium on Oak Woodlands and Hardwood Rangeland Management. USFS General Technical Report PSW-I00. Tyler, C.M., B. Kuhn, and F. W. Davis. 2006. Demography and recruitment limitations of three oak species in California. Quarterly Review of Biology 81: 127-152. Verner, ]. 1980. Birds of California oak habitats-management implications. Pages 246-264 in: Proceedings of the Symposium on the Ecology, Management, and Utilization of California Oaks, June 26-28, 1979. USDA Forest Service General Technical Report PSW-44. Verner, ]. 1988. Blue oak-digger pine in: K. E. Mayer and W. F. Laudenslayer, Jr. (eds.), A Guide to Wildlife Habitats of California. California Dept. of Forestry and Fire Protection, Sacramento (available from CDF-FRAP, 1920 20th St., Sacramento, CA 95814). Welker, ]. M., and]. W. Menke. 1990. The influence of simulated browsing on tissue water relations, growth and survival of Quercus douglasii (Hook and Arn.) seedlings under slow and rapid rates of soil drought. Functional Ecology 1990(4): 807-817. Werres, S., R. Marwitz, W. A. Man in'T Veld, et al. 2001. Phytophthora ramorum sp. Nov., a new pathogen on Rhododendron and Virburnum. Mycological Research 105: 1155-1165. White, K. L. 1966. Structure and composition of foothill woodland in central coastal California. Ecology 47: 229-237.

THIRTEEN

Chaparral JON E. KEELEY AND FRANK W. DAVIS

INTRODUCTION

BIOGEOGRAPHICAL PATTERNS

Flora

the associated urban environments with which it is often juxtaposed. Previous reviews that include topics not covered here are in Cooper (1922), Hanes (1977), Keeley and Keeley (1988) and Keeley (2000).

Community Patterns Community Classification Local Diversity Patterns FIRE

Demographic Patterns of Woody Dominants Demographic Patterns of Ephemeral Flora Regional Variation in Fire Regime COMMUNITY AND ECOSYSTEM PROCESSES

Successional Changes Shrub Life History Syndromes Flowering and Dispersal in Herbs Seed Germination Allelopathy Alien Plants EVOLUTIONARY AND GEOLOGICAL HISTORY

Shrub Life History Syndromes Chaparral Origins AREAS FOR FUTU RE RESEARCH

Introduction Chaparral is the evergreen sclerophyllous shrubland that dominates the cismontane side of coastal mountain ranges from about San Francisco south to Ensenada in Baja, California, as well as the foothills of the Sierra Nevada (Fig. 13.1). The dense impenetrable nature of chaparral (Fig. 13.2) means that it is largely off limits to all but the more dedicated ecologists and other naturalists. It is primarily known for the spectacular crown fires that frequent chaparral and

Biogeographical Patterns Chaparral is distributed from northwestern Baja, California, to south-central Oregon, with disjunct stands on mountaintops farther south in Baja, California, and arid interior slopes as far north as Washington. It reaches its greatest extent in the Transverse and Peninsular ranges of central and southern California but is also an important part of the western foothills of the interior Sierra Nevada. It continues farther east in patches on "sky islands" of desert ranges and then in a broad mid-elevation band across central and southern Arizona. Small patches of chaparral also occur in an oak and pine mosaic landscape of the Sierra Madre accidental of mainland Mexico and in larger stands in the Sierra Madre Oriental south of Monterey, Mexico (Keeley 2000). At a regional scale chaparral tends to dominate at elevations from 300 to 1,500 m, but under favorable maritime conditions can reach down to sea level and on south-facing slopes and ridges that extend higher into the coniferous zone. Chaparral in California ranges over nearly 10 degrees of latitude, and its elevational distribution is considerably higher at the southern than the northern end of the state. When elevations are adjusted to a reference latitude of San Diego, 40Y/cJ of mapped chaparral occurs between 1,000 and 1,500 m, and 90(1) occurs below 2,000 m (Table 13.1). Within this elevational zone, landscapes comprise a mosaic of different vegetation types including chaparral, sage scrub, grassland, and oak woodland. Chaparral is naturally displaced by woodlands on very mesic slopes and by sage scrub on xeric slopes. Grasslands, the vast majority of which are disturbancedependent annual alien grasslands (Huenneke 1989), often

339

FIGURE 13.1 Distribution of chaparral in California.

N

I

+

I

o occupy former shrublands that were displaced by frequent burning, beginning with the Native Americans (Cooper 1922; Wells 1962; Keeley 2002). Chaparral is closely associated with the Mediterranean climate pattern of winter rain and summer drought. Within that regime it can be found under a wide range of rainfall and temperature conditions, but over 60% of the present distribution is in areas receiving between 250 and 750 mm of annual precipitation and where average January daily temperature falls between 5°C and 15°C (Table 13.2). These patterns are consistent with studies indicating that summer drought stress may limit chaparral shrub seedling establishment and that injury to adult shrubs from winter freezes may impose species-specific distributional limits (Langan, Ewars, and Davis 1997; Boorse, Ewers, and Davis 1998). Chaparral soils tend to be shallow and rocky except near the coast where it occurs on deep Aeolian sands of marine benches and terraces. Substrates include fractured sandstones and shales, coarse-grained decomposed granitic soils, fine-grained weathered volcanics, and mafic substrates such as serpentine and gabbros. Mafic substrates have a variety of effects on chaparral, the most obvious being the number of

340

CHAPARRAL

250 Kilometers

endemic plant species (Kruckeberg 1984). These substrates add to the landscape diversity and have substantial effects on plant species diversity (Harrison and Inouye 2002). The lowered productiVity of these soils slows the rate of fuel accumulation and reduces fire frequency, thus favoring slow growing serotinus conifer species, because of both reduced postfire competition and longer fire-free intervals (Keeley and Zedler 1998). Lower productivity maintains a more open habitat and higher plant diversity than chaparral on more fertile sites, although the postfire increase in diversity is lower (Safford and Harrison 2004).

Flora

California, like other Mediterranean climate regions, supports exceptionally high plant diversity (4,846 native vascular species) and endemism (1,693 species [35%] are confined entirely or nearly entirely to the State; Cowling et al. 1996; CalFlora 2005). These levels of species richness and endemism are lower than the Mediterranean climate regions of the Cape Region of South Africa or Southwest Australia, but on a per-area basis are still six to seven times higher than the

FIG URE 13.2 Mixed chaparral in the San Gabriel Mountains of southern California. White flowering shrub is CemlOthlls crass/fa/ills. Photograph by]. Keeley.

continental United States. According to the Calflora database, 1,177 (24'l'b) of the state's native vascular species occur in chaparral communities, and 497 (42%) of these are endemics. Most of these species are also associated with several other communities. The life-form spectrum of chaparral-associated species is similar to that of the flora as a whole but is proportionally higher in annual herbs, perennial bulbs, and shrubs and lower in other perennial herbs (Table 13.3). Only 110 species, half of them shrub species, are associated solely with chaparral vegetation in the CalFlora database. Notably, 76 of those chaparral-restricted species (69%) are endemic to California, and 62 (56%) are considered rare (CaIFlora). Like other Mediterranean-climate regions, much of the diversity (25%) in the state is contributed by rare and localized species (Cowling et al. 1996). Chaparralleads other communities in the number of rare plant taxa, haVing 18% more than expected based on areal extent (Keeley 2005). Within this shrubland vegetation, rarity is not randomly distributed across growth forms. Annuals are very underrepresented on the rare plant list, whereas there are three times more rare herbaceous perennials and double the number of shrub species than expected based on the total number of species.

Landscape relations to plant species richness have been broadly studied in the California flora, and climatic variables, in particular precipitation, are the strongest predictors of diversity (Richerson and Lum 1980). In this region the marked orographic gradient results in a strong relationship between elevation and precipitation, and not surprising, elevation is also a strong predictor of plant diversity (Qi and Yang 1999). Chaparral covers a large elevational gradient, and this factor, perhaps acting through effects on precipitation, is a major determinant of diversity (Keeley, Fotheringham, and Baer-Keeley 2005b). Ninety percent of chaparral's mapped distribution falls within 74 of 284 geographic subregions created by Harrison et al. (2000) for their analysis of serpentine species diversity patterns as encoded in the CalFlora database. Contrary to patterns for the flora as a whole, the number of chaparral-associated species in these 74 units is only moderately correlated with subregion area (r = -0.23) or climate factors such as annual precipitation (r = -0.27) and is better predicted by the log(area) of chaparral (r = 0.60) and mean (r = -0.57) or minimum (r = -0.68) latitudinally adjusted elevation in the subregion (Figs. 13.3a-c). A robust regression model accounting for chaparral area, subregion area, and minimum elevation explains 80% of the total variation in chaparral plant species richness and leaves little evidence of strong regional or latitudinal gradients in the diversity of the chaparral flora (Figs. 13.3b and 13.3c). The only systematic pattern at this scale appears to be the decreased richness in the chaparral flora with increasing elevation or decreasing winter temperatures. The composition of the chaparral flora by geographiC subregion indicates fairly steep species turnover with distance. For example, the list of species associated with chaparral in western San Diego County is 42% dissimilar to that listed for the San Bernardino Mountains and 73% dissimilar to the chaparral flora of southern Lake County (Fig. 13.4a), regions of nearly identical size and chaparral extent. Again using comparable subregions, the chaparral flora of western Calaveras County in the central Sierra Nevada is 54% dissimilar to that of central Siskiyou County to the north, and 88% dissimilar to that of the northern foot slopes of the San Gabriel and San Bernardino Mountains to the south (Fig. 13.4b). Steep geographic turnover is found not only among chaparral-associated species but also among chaparral-restricted species and community dominants (Table 13.4). The two most diverse woody genera-Ceanothus and Arctostaphylosboth contain many narrowly endemic taxa. Of the 50 Ceanothus taxa (species and varieties) listed in the CalFlora database, 43 (86%) are recorded in subregions totaling less than 40,000 km 2 (Fig. 13.5). Similarly, 71/81 Arctostaphylos taxa have recorded ranges less than 40,000 km 2 (Fig. 13.6). Cody (1986) estimated a local turnover rate of 50% of the species in these two genera within a distance of 100 to 300 km, depending on the steepness of local environmental gradients. These replacement patterns are illustrated in the distribution patterns of four closely related Ceanothus species in the subgenus Cerastes (Fig. 13.7). Similar elevational replacements have also been described by Zedler (1995a).

CHAPARRAL

341

TABLE 13.1

Distribution of Chaparral Vegetation by Elevation

Area in State (km 2)

Adjusted Elevation (m)

Chaparral Area (km 2)

Chaparral as % State Area

% of Chaparral Area

0-500

86,430

927

1.07

3.78

500-1,000

113,703

7,856

6.91

32.05

1,000-1500

75,397

9,961

13.21

40.64

1,500-2000

47,686

3,884

8.14

15.85

2,000-2500

49,042

1,348

2.75

5.50

2,500-3000

20,826

484

2.32

1.97

3,000-3500

7,516

44

0.59

0.18

3,500-4000

3,513

5

0.14

0.02

>4,000

655

0

0.00

0.00

404,768

24,509

6.06

100.00

Total

NOTE: To adjust for the effect of increasing latitude, 0.625 m is added to elevation for every km north from the southernmost point in the state. Statewide maps of chaparral and elevation were combined to produce the statistics.

TABLE 13.2

Percentage of Chaparral Area in Different Combinations of Mean January Temperature and Total Annual Precipitation

Mean Annual Precipitation (mm) 0-250

250-500

500-750

750-1,000

1,000-1,250

1,250-1,500

1,500-2,000

>2,000

< -5

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.1

-5-0

0.0

0.0

0.2

0.5

0.4

0.2

0.2

0.0

1.6

0-5

0.0

1.2

4.5

1.4

1.5

1.0

1.2

0.4

11.3

5-10

0.6

7.5

13-5

8.2

4.9

2.3

0.5

0.2

37.5

10-15

0.7

25.2

15.4

3.6

0.8

0.2

0.0

0.0

46.0

>15

0.1

3.3

0.0

0.0

0.0

0.0

0.0

0.0

3.4

Total

1.4

37.3

33.7

13-8

7.6

3.8

1.9

0.6

100.0

Mean Jan temp CC)

NOTE:

Based on overlay of statewide maps of chaparral and climate data.

Relatively few dominant chaparral shrub species are widespread, the exception being Adenostoma fasciculatum. Perhaps most striking are the chaparral landscapes where this species is absent (Fig. 13.8). Specifically, desert borders in southern California, much of the Tehachapi Range (Bauer 1930), above 1,000 m in interior drainages of the southern Sierra Nevada, and montane chaparral and eastside chaparral of the Sierra Nevada and Cascade ranges. On the desert border and interior drainages of the southern Sierra Nevada a couple of the dominants that replace A. fasciculatum are Ceanothus cuneatus (Fig. 13.7) and Cercocarpus betuloides. Based on climatic analysis of their geographic ranges Westman (1991) showed that the latter two species occupied a similar precipitation range (500-750 mm) as A. fasciculatum l

342

Total

CHAPARRAL

but were more tolerant of lower January temperatures than

A. fasciculatum. Indeed, Malanson, Westman, and Yan (1992) place the optimum growth position of A. fasciculatum at 9°C-12°C for the coldest month but at 0°C-3°C for C. cuneatus. The extreme winter temperatures on many of these interior and northern sites is likely to be the main factor limiting the distribution of A. fasciculatum l and this may be the primary reason it is absent from Arizona chaparral. A number of chaparral shrub species have distributions far outside the California Mediterranean climate region, yet seldom dominate chaparral in the state. These include Garrya spp., Rhamnus crocea/ilicifolia and Rhus ovatal all common elements in the Arizona chaparral. A few California chaparral species, Rhamnus californical Ceanothus greggiil and

TABLE 13.3

Life-form Spectra for All California Native Vascular Plant Species, Species Associated with Chaparral, and Species Limited to Chaparral

Chaparral Species

All Native Species

Chaparral Species

1,443

29.78

409

34.75

Annual herb (aquatic)

9

0.19

2

0.17

Annual herb (hemiparasitic)

8

0.17

2

0.17

All Native Species

(Yo

(Yr.)

Chaparral Only Species

(M) Chaparral Only Species

27

24.55

Life fornl Annual herb

Annual herb, Vine Annual herb, Vine (parasitic) Annual, Perennial herb Annual, Perennial herb (aquatic)

0.02 8

0.17

2

0.17

76

1.57

20

1.70

2

0.04

2

0.17

Annual, Perennial herb (rhizOIl1atous) Perennial herb

0.02

45.09

389

33.05

Perennial herb (aquatic)

40

0.83

4

0.34

Perennial herb (bulb)

68

1.40

29

2.46

3

0.06

11

0.23

Perennial herb (Il1osslike)

5

0.10

Perennial herb (parasitic)

28

0.58

2

0.17

162

3.34

25

2.12

6

0.12

Perennial herb (hemiparasitic)

Perennial herb (rhizolnatous) Perennial herb (saprophytic) Perennial herb (stem succulent)

17.27

0.91

0.91 2

1.82

0.02 10

0.21

2

0.17

Perennial herb, Vine

14

0.29

8

0.68

219

18.61

9

0.76

Perennial, Biennial herb

0.91

0.02 543

11.21

5

0.10

Shrub (stem succulent)

47

0.97

Shrub, Perennial herb

3

0.06

Shrub (parasitic)

19

0.08

Perennial herb, Shrub

Shrub

0.91

0.08

2,185

Perennial herb (carnivorous)

0.91

Shrub, Tree

54

49.09

0.02

Tree

84

1.73

22

1.87

2

1.82

Tree, Shrub

62

1.28

26

2.21

1

0.91

4

0.08

Vine, Shrub

11

0.23

3

0.25

Unclassified

4

0.08

Grand Total

4,846

100.00

110

100.00

Vine

SUlll{U::

CalFlora database, December 2004.

1,177

100

.62.118 .119.178

.53.139 .140.217 _ 218·278

.179.255 0256.334

0279.339

0335.414

0340.422

N

i I 70

I

i

"0

A

• .206 ..121

_.120.-41 _-40.7 '8·69

c:J 69 ·139

.;.~:!> ",I.',':

.';,;

:}< -•• :;'."

;;.;,

.

N

Species richness. Observed (A) and predicted (B) patterns of chaparral plant species richness in 74 subregions containing 90% of the total chaparral area within California. The prediction is based on the model: S = 123.S*log(chaparral area in km 2) -22.8*log(subregion area) - 0.04*(latitudinally adjusted elevation, m). Multiple r2 = 0.80 and model residuals are mapped in (C).

i

FIGURE 13.3

Arctostaphylos pungens are found in Arizona chaparrat and

also are important elements in the Sierra Madre Oriental chaparral of northeastern Mexico (Keeley 2000).

Community Patterns

In the absence of disturbance, communities are dominated by shrubs and subshrubs with a minor representation of other growth forms such as lianas, vines, geophytes, and annuals. Composition often changes markedly after fire

344

CHAPARRAL

I

70

1.0

I

I 200 Kilomelers

c with the sprouting of dormant bulb and seed banks. With rare exceptions the prefire dominants persist after fire either as resprouts from basal buds or from dormant seed banks. On complex landscapes there is a diversity of floristically different plant associations not easily explained by any single factor (Keeley 2000). Although Adenostoma fasciculatum sometimes forms nearly pure stands, more often it occurs in mixed stands with species of Ceanothus and Arctostaphylos. For example, maritime chaparral, associated with sandy substrates in

.'-20

.0

.0

.21-40

.21-40

041-60

041-60

ml61-S0

ml6'-SO

b,;::l81-'00

b:;JS'-100

.'-20

N

N

1 A

I

I 0

80

I 160

1 I

I 320 KJlomelers

B

I

I 0

80

I 160

I

I 320 KJlometers

FIGURE 13.4 Sorensen percentage dissimilarity between the chaparral floras of western San Diego County (A) or western Calaveras County (B) and 47 other geographic subregion containing 7S% of the current distribution of California chaparral. See Veirs et al. (2006) for description of the database used in the analysis.

level or rolling terrain within 10 to 20 km (6-12 miles) of the coast, occurs in Torrey Pines State Reserve in San Diego and is scattered along the coast from northern Santa Barbara County to Sonoma County. Although Adenostoma fasciculatum co-dominates maritime chaparral throughout the range of the maritime chaparral type, there is geographical replacement of rare endemic ceanothus and manzanita species from south to north; Ceanothus verrucosus/A. glandulosa ssp. crassifolius in San Diego, C. impressus/A. rudis/A. purissima in Lompoc, C. rigidus/A. morroensis near Morro Bay, and C. cuneatus var. rigidus/A.hookeri ssp. hookeri/A. pajaroensis near Monterey. On highly dissected landscapes community composition is rather fine grained with communities varying between arid, usually south-facing slopes and ridges, and mesic, north-facing exposures. One of the strongest determinants of community composition is a soil moisture pattern; thus associations can be recognized as characteristic of the arid or mesic end of this gradient. Chamise (Adenostoma fasciculatum) is the nearly ubiquitous dominant on most arid chaparral sites. Also, well developed on arid south-facing slopes and ridges are nonsprouting species of ceanothus (Ceanothus spp.) (Fig. 13.2) or manzanita (Arctostaphylos spp). More mesic north-facing slopes often favor associations comprising broader-leaved evergreen shrubs, including scrub oak (largely Quercus berberidifolia but occasionally Q. wizlizenii), coffeeberry (Rhamnus californica), redberry (R. crocea), silk tassel (Garrya spp.), holly leaf cherry (Prunus ilicifolia), and chaparral holly (Heteromeles arbutifolia). Even more fine-grained

distribution patterns on individual slope faces can be related to cold air drainage patterns and freeZing tolerance of different species (Davis, Pratt, and Bowen 2004).

Community Classification

Different workers have classified chaparral communities by a combination of environmental factors (e.g., maritime chaparral, serpentine chaparral, montane chaparral, semidesert chaparral; mesic north-slope chaparral, xeric southslope chaparral) and dominant genera (e.g., ceanothus chaparral, manzanita chaparral; e.g., Horton 1960; Hanes 1977; Mayer and Laudenslayer 1988). Holland's (1986) widely used system identified 28 chaparral communities defined by environment and dominant species. Most recently, Sawyer and Keeler-Wolf (1995) have adopted the national hierarchical vegetation classification system to California and described chaparral alliances defined by one or 2 dominant species as well as a number of associations within alliances based on quantitative plots measurements. Their list of chaparral types includes 9 "undifferentiated" chaparral scrub types and more than 60 alliances including 9 chamise alliances, 15 ceanothus alliances, 10 manzanita alliances, 18 scrub oak alliances, and 4 redshank (Adenostoma sparsifolium) alliances. Although this floristically based system has advantages as a naming system for some resource management purposes, it is of more limited value to ecologists interested in community structure and function (Zedler 1997), the primary limitation being the independent nature of species

CHAPARRAL

345

TABLE 13.4

Common Shrubs in Chaparral and their Regeneration Characteristics in the First Postfire Year

Common Name

Seedling Recruitmentll

Basal Sproutint

Malosma laurina

laurel sumac

postfire

yes

Rhus ovata

sugar bush

postfire

yes

Arctostaphylos glauca

big-berry manzanita

postfire

no

A. glandulosa

Eastwood manzanita

postfire

yes, burF

A. hooked

Hooker manzanita

postfire

no

A. manzanita ssp. roofii

Common manzanita

postfire

yes, burl d

A. parryana ssp. parryana

Parry's manzanita

postfire

no

A. parryana ssp. tuberescens

Parry's manzanita

postfire

yes, burl d

A. rainbowensis

Rainbow manzanita

postfire

yes, burl

A. rudis

sand mesa m.

postfire

no/yes, burl e

A. stanfordiana

Stanford's manzanita

postfire

no

A. tomentosus

woolly-leaf manzanita

postfire

yes, burl

A. viscida

white-leaf manzanita

postfire

no

Xylococcus bicolor

mission manzanita

unknown (rare)

yes, burl

Pickeringia montana

chaparral pea

unknown

yes

Prunus ilicifolia

holly-leafed cherry

older stands

yes

Quercus berberidifolia

scrub oak

older stands

yes

Q. durata

leather oak

older stands

yes

Q. garryana var. brewed

Brewer's oak

older stands

yes

Q. wizlizenii

shrub live oak

older stands

yes

silk tassel

postfire

yes, burl

Ceanothus crassifolius

hoaryleaf ceanothus

postfire

no

C. cuneatus

buck brush

postfire

no

C. greggii

cupleaf ceanothus

postfire

no

C. impressus

Santa Barbara ceanothus

postfire

no

C. leucodermis

chaparral whitethorn

postfire

yes, burl

C. megacarpus

bigpod ceanothus

postfire

no

C. oliganthus

hairy ceanothus

postfire

no

C. spinosus

greenbark ceanothus

postfire

yes, burl

C. tomentosus

woolly-leaf ceanothus

postfire

Scientific Family/Species Name Anacardiaceae

Ericaceae

Fabaceae

Garryaceae

Garrya spp. Rhamnaceae

TABLE 13.4

Scientific Falnily/Species Name

Common Nalne

(continued)

Seedling Recruitmentl

Basal Sprouting)

var. tOlnentosus no

var. olivaceous C. verrucosus

wart-stemmed ceanothus

postfire

no

Rhalnnus californica

coffeeberry

older stands

yes

R. crocea

redberry

older stands

yes

Adenostolna fasciculatum

chamise

postfire

yes, burl

A. sparsifolium

red shank

unknown

yes, burl

Cercocarpus betuloides

mountain mahogany

various disturbances

yes, base and rhizomes

Hetero1rzeles arbutifolia

chaparral holly

woodland gaps

yes

flannel bush

postfire

yes

Rosaceae

Sterculiaceae

Fremontodendron spp.

NOTE: Postfire obligate seeders are non-sprouting species with postfire seedling recruitment,. Postfire facultative seeders are sprouters with postfire seedling recruitment. Postfire obligate resprouters resprout but do not recruit seedlings in the postfire environment. a Seedling recruitment is largely unknown in a few species. h Not all species that resprout will form burls or lignotubers as a normal ontogenetic developmental stage. In some cases, such as Querclls spp. swollen burl-like structures will form as a type of "coppicing" effect. l Although a vigorous resprouter distributed from Baja California to Oregon, a single non-burl forming population, unable to resprout after fire, is known frOln northern Baja California O. Keeley, Vasey, and Parker in press). d Arctostaphylos manzanita and A. parrayana have non-burl forming subspecies in open woodland habitats and burl forming subspecies in chaparral (Keeley, Boykin, and Massihi 1997). l' Arctostaphylos rudis has burl-forming and non-burl-forming plants in the same population such as Nipoma Mesa. I Ceanothus tomentosus populations from the central Sierra Nevada are resprouters and the southern California populations are non-sprouters.

distribution on most chaparrallandscapes (Ackerly 2003). For example, overlaying maps of 19 chaparral shrub species in a 60,000-ha area of the Santa Monica Mountains in southern California produced 220 unique combinations of locally dominant species that could be construed as alliances (Syphard, Franklin, and Keeley 2006). Additional quantitative phytosociological studies are needed to establish general chaparral types that are both useful for inventory and management and ecologically meaningful.

Local Diversity Patterns

Community level plant diversity in California shrublands has long been known to follow a marked temporal pattern with the greatest richness concentrated in the early postfire years (Sampson 1944; Horton and Kraebel 1955; Keeley et al. 1981; Davis, Hickson, and Odion 1988; Guo 2001). In early successional stands plant richness may range from 0 to 30 species per m 2 (average ~ 10) and 25 to 80 or more per 0.1 ha (average ~ 50; Keeley and Fotheringham 2003a). These values are quite similar to those reported for the very species rich South African fynbos or Western Australian heathlands. Thus,

although these latter mediterranean climate shrublands may have substantially higher regional plant species diversity, they are quite comparable at the community level to postfire Californian chaparral. The primary difference between California and these Southern Hemisphere shrublands is that in chaparral, high species richness is a transient postfire condition, and as shrub canopies close in, diversity drops. Diversity in chaparral is made up of a large number of relatively minor species and this is illustrated by dominance-diversity curves (Fig. 13.9) that tend to fit a geometric series in early succession and even more accentuated in later succession (Keeley and Fotheringham 2003a). Communities fitting this geometric curve are thought to be driven by the niche-preemption model, which describes communities where a single species dominates a substantial fraction of resources and subordinate species in sequence occupy a similar fraction of the remaining resources (Whittaker 1972). Dominance in chaparral is driven by the fact that a substantial portion of postfire resources are immediately occupied by resprouts from large root crowns and lignotubers, even though these resprouters comprise only about a quarter of the flora (Keeley 1998). Additionally, some of these same species recruit

CHAPARRAL

347

200000 180000 160000 140000

E .Y

120000

0-

~ 100000 eel

Q)

~

80000 60000 40000 20000

5

7

9

11

13

15

17 19

21

23

25

27

29

31

33

35

37

39

41

43

45

47

49

Rank FIG U RE 1 3.5 Rank ordering of range sizes (km 2) for 50

Ceanothus taxa based on documented occurrence in the CalFlora database within 284 geographic subregions in California.

from fire-stimulated seed banks and thus are poised for expanding dominance with age. The annual life history results in unstable competitive boundaries that are continually readjusted each year, eventually being all but completely crowded out as the shrub canopy closes.

Fire Chaparral fires are nearly always active crown fires in which living and dead fuels in the canopy carry fire, and surface fuels play little or no role. Fire behavior is most strongly controlled by wind and in the absence of significant winds is controlled by the proportion of living and dead canopy biomass and topography (Keeley and Fotheringham 2003b). When strong winds are present, fires often burn large

swaths of landscape on the order of thousands of hectares and occasionally much larger areas. Recovery is rapid (Fig. 13.10) because it largely involves residual species regenerating by resprouts from vegetative structures or germinating from dormant soil-stored seed banks. Colonization is of relatively limited importance in terms of biomass but may significantly affect diversity. In a southern California postfire study it was found that of the species present in tenth hectare plots during the first 5 years, only about 50°/6 to 60(310 were present in the first year (Keeley, Fotheringham, and Baer-Keeley 2005a). However, in Year 5 those species present since the first year comprised about 90(Yo of the cover. Very few species are long-distance dispersers, and colonization is largely from localized populations in adjacent burned areas that expand follOWing fire

120000.,..........----------------·---------------·

100000

80000

E .:::£

rr

.5!!.-

60000

ctl

Q)

~

40000

20000

o

5

9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 Rank

FIG U RE 13.6 Rank ordering of range sizes (km 2 ) for 81 Arctostaphylos taxa based on documented occurrence in the CalFlora database within 284 geographic subregions in California.

348

CHAPARRAL

FIGURE 13.7 General distribution of four Ceanothus species in southern California. Data source: California Gap Analysis Project. http://www. biogeog.ucsb. edu/projects/gap/gap_home. html.

) "

••

--

0 I 0

I

40

C. megacarpus C. crassifolius

N C. cuneatus

c. greggii

I 80

I

1

160 Kilometers

Chamise

Chaparral

I

I

o

60

120

I 240 Kilometers

FIGURE 13.8 General distribution of chamise (Adenostoma (asciculatum) and chaparral in California. Data sources: California Gap Analysis Project and California Department of Forestry and Fire Protection.

(a) Coastal Chaparral

(b) Interior Chaparral

100.000 10.000 FIGU RE 13.9 Dominance-diversity curves for coastal (a) and interior (b) chaparral in southern California for the first postfire year (circles) and the fourth postfire year (triangles). Data taken from Keeley and Fotheringham (2003a).

G U) c.:l

1.000

. There was no trend of increasing or decreasing success between the beginning and ending years. Failure in meeting vegetation standards was far more common than failure in meeting hydrology standards. The usual reasons for failing to meet vegetation standards were inappropriately steep slopes, failure to apply sufficient inoculants to the pools (topsoil with plant propagules taken from pools to be demolished and spread onto the surface of new pools), and lack of continuous weed control. The most important locational reason for failure was creating pools at locations without natural horizons that restricted drainage and where natural pools did not exist. Longer monitoring, elsewhere, has shown variability in success. Ferren et al. (1998) followed 16 pools that had been enhanced or created in the Santa Barbara area for 10 years and concluded that" ... we found that vegetation zonation, cover, and floristic composition ... approached those of natural pools of the region ... " Black and Zedler (1998) summarized monitoring observations of up to 15 years for created pools at several locations in San Diego Country. Total species number, percentage of cover, biomass of vernal pool species, and the population size of rare target species Pogogyne abrarnsii were typically lower in created pools than in undisturbed pools. However, they still concluded:

"... the evidence suggested a gradual convergence of the function and characteristics of artificial basins toward those of natural pools ... land thatj artificial basins can support populations of native species for considerable periods ... " Collinge (2003) studied the establishment of the federally listed plant Lasthenia conjugens in constructed pools at the Travis Air Force Base in Solano County. In the winter, L. COlljllgells seeds were planted in permanently marked plots in constructed pools (100 seeds per 0.25 m 2 plot). In the two successive winters, 100 seeds were added to half of the treatIllent plots. Her team found that hydrologic function (average water depth and duration of inundation) of constructed pools was comparable to reference pools. L. conjllsells abundance increased in seeded plots over the 4-year saIllpling period. In the plots seeded three times, L. conjusells abundance increased from an average of 12 individuals in 2000 to an average of 240 individuals in 2002 and 317 individuals in 2003. In the plots seeded once, average abundance of L. conjllgc11s increased to 87 individuals in 2002 and 175 individuals in 2003. Collinge's results suggest that it Inay be possible to establish sufficient numbers of L. C011jllge11s in constructed pools. Another evaluation, conducted by Starr (2004) examined 27 pools (14 constructed and 13 natural) 8 years after constructed pools had been created at Wurlitzer Preserve near Chico in the Sacramento Valley. She concluded that floristic differences between pairs of created and natural pools were not significant. Ordination, for example, did not clearly separate created from natural pools, whether burned or unburned, grazed or ungrazed, and neither did an agglomerative cluster analysis. Perhaps the special conditions of this site led to such a level of success: (1) there was in existence already a continuous hardpan throughout the site, and new created pools were located aInong the natural pools; (2) the specialized pollinators were already present on the site; (3) all created pools were inoculated with topsoil froIn pools that were destroyed on the developed site; (4) the created pools were carefully matched in morphology to nearby natural pools; and (5) disturbance to the site during the creation phase was kept to an absolute minimuIn. Despite evidences of success, vernal pool mitigation is a highly controversial process that has been frequently criticized for its inability to adequately replace the original intact ecosysteIns (Ferren and Gevirtz 1990; Leidy and White 1998). Wacker and Kelly (2004) applied GIS analysis to determine how Initigation process affects the structure of vernal pool c()Jnplexes on a landscape level. They found that the more COInIll0n edaphic settings, such as Northern Hardpan and Low Terrace pools, are increasing, but the less common types, such as Northern Claypan and Volcanic Mudflow pools, are decreasing. They also showed that I)rainageway pools, a less specialized pool type with generally lower species richness, are becoIning more COInmon through mitigation. Probably the safest conclusion to reach about the value of created pools as a Illitigation strategy is that statistically based Il1eaSUres of success can be demonstrated, but it remains

to be determined what those measures/criteria should be and for how long monitoring should be required before determining whether those criteria have been met. There is a general agreement among plant ecologists that the length of monitoring should be longer than 5 years. Currently used success criteria ("performance standards," see DeWeese 1998) include the depth and duration of inundation and the cover, number, vigor, and identity of vernal pool endemics present in constructed pools. These criteria certainly correctly address pool hydrology and flora as the most important targets. We suggest, however, some modifications that will lnake some criteria stronger and others more realistic. 1. Depth and duration of created pools should not differ statistically significantly (p < 0.05) from those in nearby natural pools. The current criterion specifies that the created pools should not hold water deeper and longer than 125 :tQ) ..., c::

E

::] ..., Q) a:

• • .. •

Black Oak Ponderosa Pine - Mixed Conifer White Fir - Mixed Conifer Giant Sequoia - Mixed Conifer Jeffrey Pine

o

Red Fir Montane Chaparral Meadow Lodgepole Pine Foxtail Pine

o

•

100

Foothills Grassland & Hardwood Foothills Chaparral

V

~. T

Q)

ii:

50

c::

o

CU

Q)

~

500

much of the landscape in Sequoia and Kings Canyon National Parks the Illedian date of last fire was 1875 (Caprio et al. 2002), followed by an unprecedented fire-free interval. Fire exclusion led to several notable changes in forests, particularly in the ponderosa pine-, white fir-, and giant sequoia-mixed conifer forests (Vankat and Major 1978; Parsons and DeBenedetti 1979; Stephenson 1996; Taylor 2004; van Wagtendonk and Fites-Kaufman 2006). For example, giant sequoia reproduction, which in the past depended on frequent fires to expose mineral soil and open gaps in the forest canopy, effectively ceased, and reproduction of other shade-intolerant species such as ponderosa pine was reduced. Forests became denser in many areas, with increased dominance of shade-tolerant species such as white fir and incense cedar (Parsons and DeBenedetti 1979; Bouldin 1999). More area came to be dominated by dense intermediate-age forest patches (Bonnicksen and Stone 1982; Stephenson 1987). Shrubs and herbaceous plants became less abundant than in the past (Kilgore and Biswell 1971; Harvey et al. 1980). Perhaps most importantly, dead illaterial accumulated, causing an unprecedented buildup of surface fuels (Agee, Wakinl0to, and Biswell 1978; van Wagtendonk 1985). Additionally, "ladder fuels" capable of conducting fire into the crowns of mature trees increased (Kilgore and Sando 1975; Parsons and DeBenedetti 1979). One of the most immediate consequences of these changes was an increase in severity of wildfires that was rarely encountered in pre-Euroamerican times (Kilgore and Sando 1975; Stephens 1995, 1998; van Wagtendonk and Fites-Kaufman 2006). INSECTS AND PATHOGENS

Native bark beetles are the proximate cause of the majority of natural tree deaths in the Sierra Nevada (Ferrell 1996). Most often, beetles kill trees that were already stressed by drought, root disease, dwarf mistletoe infestation, or air pollution. A potential additional source of predisposing stress is increased forest density, brought about by fire exclusion

460

MONTANE AND SUBALPINE VEGETATION

1000

1500 2000 2500 Elevation (meters)

3000

3500

(Slnith, Rizzo, and North 2005) and some logging practices (Sherman and Warren 1988; Ferrell; Barbour et al. 2002). AIll0ng the native fungal root diseases, the most important are annosus root disease, armillaria root disease, and black-stain root disease (Ferrell 1996). Annosus root disease in particular is thought to be spreading more easily in overly dense forests resulting from fire exclusion (Slaughter and Rizzo 1999; Rizzo and Slaughter 2001). Additionally, cut stumps often prove to be the establishment points for the disease. Of particular concern is white pine blister rust, an epidemic disease introduced from Asia. Blister rust attacks fiveneedled pines, and has now been found in all five-needled pine species in the Sierra Nevada. However, the first and most severely affected species in the Sierra Nevada is sugar pine, often the second or third most abundant tree species in Illixed-conifer forests. Combined with other stresses, blister rust has been contributing to a long, steady decline in sugar pine populations, which is expected to continue into the future (van Mantgem et a1. 2004). WI N D AN D AVALANCH ES

Compared to much of the rest of North America, wind is a less important force shaping forests of the Sierra Nevada. However, forest edges or substantially thinned forests (whether created by logging, fire, avalanche, or other cause) tend to be vulnerable to windthrow (Fosberg 1986). Wind can affect upper montane forests significantly, although it is not clear how much of the current patterns are confounded by thinning. Avalanches are rare in lower elevation forests (such as the mixed-conifer forests; but see Fry 1933), instead tending to be most comlnon in red fir, lodgepole pine, and subalpine forests (Kattelmann 1996). Avalanches are most common on steep, north-facing slopes at these elevations, and often recur in given locations (avalanche paths). A moderate recurrence interval (every few decades) tends to result in dominance by aspen rather than by conifers; more frequent

recurrences can lead to avalanche paths that are chronically free of trees. LOGGING

The nUInber of large trees in the Sierra Nevada has declined significantly since the arrival of Euroalnerican settlers (USIJA Forest Service 2001). For example, in white fir-Inixed conifer forests of the central and northern Sierra Nevada during the 57-year period from about 1935 to 1992, numbers of trees greater than 24 inches (61 C111) in diameter declined by 60(Y36 inches (91 cm) in dianleter declined by >80(Y30

57

70

1935

Bouldin 1999

>30

64

30

1992

Bouldin 1999

>9.4 (>24)

721 (219)

75 (67)a

Plumas National Forest

1996

Ansleyand Battles 1998

157

58

Plumas National Forest

1957

Ansleyand Battles 1998

DF, WF, SP, IC, PP, Ba, MD

>24

DF, WF, IC, SP, MD

>2.5

80 (34-150)

Mesic sites on Lassen, Plumas, Tahoe, Eldorado National Forests

1980's

Fites 1993

DF, PP, IC, WF, SP, CL

>2.5

68 (23-104)

Xeric sites on forests listed previously

1980's

Fites 1993

Mixed Ponderosa Pine PP, IC, WF, Ba, SP

>2.5

1977

133

Sequoia-Kings Canyon National Park

1969

Royand Vankat 1999

IC, PP, WF, Ba

>2.5

1165

107

Sequoia-Kings Canyon National Park

1996

Royand Vankat 1999

PP, IC, WF, SP, Ba

>2.5

61 (17-128)

Lassen, Plumas, Tahoe, Eldorado National Forests

1980's

Fites 1993

NOTE: Species abbreviations are as follows: Ba, black oak; CL, canyon live oak; OF, Douglas fir; IC, Incense cedar; JP, Jeffrey pine; MD, mountain dogwood; PP, ponderosa pine; RF, red fir; SP, sugar pine; WF, white fir. aStandard deviation.

FIGURE 17.5 Old-growth mixed-conifer forest, Placer County Big Trees, Tahoe National Forest, approximately 1,500 m elevation. Dominant tree species (in declining importantce) are Pinus ponderosa, P lambertiana, Pseudotsuga menziesii, Abies concolor, and Calocedrus decurrens. Photo courtesy of M. G. Barbour.

Acer glabrum is found. Salix scouleriana also occurs on sites with a past history of large fires. Herbaceous cover is mostly sparse in white fir forests (seldom >5%), except in occasional moist swales or drainage bottoms, where it may approach 100%. It may be completely absent over large areas with thick litter accumulation. Common species in denser, shady forests include Hieracium albiflorum, Kelloggia galiodes, Viola lobata, Festuca rubra, and Carex rossii; and Smilacina racemosa, Pedicularis semibarbata, and Phacelia hydrophylloides occur in denser, shady forests. In more open forests, the herb layer is often diverse and can be rich (Mellmann-Brown and Barbour 1995; North et al. 2005b), including Anaphalis margaritacea, Brodiaea elegans, Calyptridium umbellatum, Collinsia parviflora, Eriogonum nudum, Eriogonum wrightii, Eriogonum umbellatum, Erysimum capitatum, Lupinus spp., and Achnatherum spp. In many sites in white fir forests where a heavy carpet of litter coats the soil surface, herbaceous vegetation is restricted to scattered individuals of Pyrola picta and Chimaphila menziesii. The diversity of these and other mycorrhizal epiparasites (including Allotropa virgata, Pyrola asarifolia, P. minor,

468

MONTANE AND SUBALPINE VEGETATION

P. secunda, Hemitomes congestum, Pityopus californicus, Sarcodes sanguinea, Corallorhiza striata, C. macula ta, and Cepahlanthera austiniae) is remarkable in the Sierra Nevada. The heavy snow cover and moderate winter temperature extremes may proVide conditions allowing critical fungal activity in the root zone throughout the year. Changes in white fir-mixed conifer forests have occurred since European settlement as in the ponderosa pine and Douglas-fir forests. Both increases in density, particularly of smaller trees (Bouldin 1999; Barbour et al. 2002; Taylor 2004; North et al. 2004) and decreases in large trees have occurred. The spatial pattern of the increase in density has not always been uniform, with at least some of the in-growth associated with aggregates of existing trees (North et al.), although overall there appears to have been an increase in uniform spatial structure compared to a more heterogeneous structure at the time of European settlement (Taylor). These shifts in composition have been attributed to fire suppression (Taylor) but may also be confounded with shifts in climate favorable for white fir (MiIIar and Woolfenden 1999). In xeric sites, with a higher proportion of ]effrey pine, as in some areas such as the Lake Tahoe basin, preferential logging of pine over fir for mining contributed to the shift in composition (Lindstrom 2000). A contrasting trend was observed by Roy and Vankat (1999) over a 27-year period since 1969 in Sequoia-Kings Canyon National Parks. Total stand density actually declined, including stems of white fir, which was attributed to prescribed burning (and in unburned stands, self-thinning) during this period. The effect of prescribed burning on reduction in stem density, particularly in smaller diameter white fir, has been documented by monitoring in the parks (Keifer and Stanzler 1995). Other less readily quantified changes in white fir-mixed conifer forests are on the distribution, size, and dynamics of large montane chaparral patches. Early explorers such as Leiburg (1902) documented the presence of large patches of chaparral amongst the white fir-mixed conifer and high-elevation red fir forests in the Sierra Nevada. Although the presence of some of these shrub patches have been attributed to shallow soils (Potter 1998), others are a result of fire (Russell et al. 1998; Nagel and Taylor 2005). Nagel and Taylor found lower historic fire frequencies in chaparral than in nearby forests in the Lake Tahoe basin and suggested that fire suppression has caused a reduction in chaparral area of 62%. It is unknown what role early settler fires played in creation or maintenance of large chaparral patches in the white-fir mixed conifer forests compared to historic fires from lightening or Native Americans. The shade-tolerant nature of white fir allows it to regenerate below chaparral (Conard and Radosevich 1982), with little effect on height growth (Oliver and Uzoh 2002). As a result, recruitment of white fir in chaparral can be nearly continuous over time (Conard and Radosevitch; Nagel and Taylor). Most of the shrubs sprout follOWing fire, resulting in mortality of white fir regeneration and perpetuation of shrubs with fire. Nagel and Taylor found that it took nearly 30 years on average for

TABLE

17.6

Composition of White Fir and Mixed White Fir-Jeffrey Pine Forests

Relative Composition (% of 5te m s!ha) Cohort and Species

A

B

Overstory trees

Abies concolor

D

C

E

>30 cm

>40 cm

>2.5 cm

68

66

H

30

17

53

40

88

Pinus lambertiana

3

2

1

19

54

20

Calocedrus decurrens

2

1

2.5

533

91

Sequoia-Kings Canyon National Park

1969

Roy and Vankat 1999

JP, IC, WF, LP

>2.5

844

105

Sequoia-Kings Canyon National Park

1996

Roy and Vankat 1999

JP, WF, SP, IC, RF

>1

278

45

Lake Tahoe basin

2000

Barbour et al. 2002

WF, JP, RF, WP, PP, SP

>4

1196

80

Lassen National Park

1998

Taylor 2000

NOTE:

GS

giant sequoia, WF = white fir, SP = sugar pine, RF = red fir,

le = incense cedar.

FIGURE 17.7 (a) Abies magnifica forests occupy a climatic zone that experiences a deeper and longer lasting snowpack than anywhere else in

California. (b) An old-growth red fir forest, with a single-canopy layer and modest regeneration, 2,200 m elevation, Tahoe National Forest. Overstory trees average 0.8 m dbh and an age of about 250-350 years. Photograph (a) courtesy of M. G. Barbour, (b) courtesy of J. Franklin.

slow. Shade-tolerant seedlings may grow only 3 cm a year for 60 to 80 years or more under these conditions (Gordon 1973). In years of heavy seed crops, favorable microsites such as forest openings and gaps are filled with large numbers of seedlings, which slowly thin to saplings and finally mature to become a new group of mature canopy trees. Large gaps for colonization may be produced by wind damage, fire, logging, or outbreaks of insects and fungal pathogens. It is these processes that produce the observed mosaics of relatively even-aged stands. The dense shade and thick litter accumulation beneath mature red fir canopies restricts the growth of most understory plant species. The most common understory shrubs present in the central and northern Sierra Nevada are Ribes roezleii and Symphoricarpos vaccinioides (Oosting and Billings 1943). Herbaceous cover seldom exceeds 5%, with the most common species being Hieracium albiflorum, Poa bolanderi, Pedicularis semibarbata, and Aster breweri. Epiparasitic herbs such as Sarcodes sanguinea, Pterospora andromeda, Corallorhiza maculata, and the related Pyrola picta, Orthila secunda, and Chimaphila umbellata are common. Openings with gravelly granitic soils often have greater herb cover, with

Arabis platysperma, Viola purpurea, Eriogonum nudum, Gayophytum nutta/lii, Monardella odoratissima, Calyptridium umbellatum, Elymus elymoides, and Wyethia mollis as typical species (Rundel et al. 1977; Potter 1998). Broader open areas with rocky soils or where fires have opened gaps support montane chaparral species such as Ceanothus cordulatus, C. velutinus, Arctostaphylos patula, ChrysoLepis sempervirens, and Ribes roezlii, with Prunus emarginata and Salix scouleriana on wetter sites and Quercus vaccinifolia on rocky outcrops (Potter 1998). Small lightning-caused fires within red fir stands may kill local groups of trees and open up a gap with appropriate soil conditions for colonization by lodgepole pine seedlings. Such local populations of lodgepole pine are relatively transitory, however, as fir seedlings regenerate well beneath such open canopies. In a time frame of 1 or 2 centuries, red firs overtop and shade out the lodgepole pines.

JEFFREY PINE

Jeffrey pine (Pinus jeffreyi) replaces the lower yellow pine forests of ponderosa pine on drier sites in the upper montane

MONTANE AND SUBALPINE VEGETATION

473

FIGURE 17.8 Open lodgepole pine stand near the edge of a wet meadow, upper montane zone. Photography courtesy of]. Franklin.

zone of the Sierra Nevada. Its belt of primary occurrence lies at 1,520 to 1,830 m in the northern areas of the range, and at 1,600 to 2,600 m in the southern Sierra Nevada. On the western slopes of the range, jeffrey pine most commonly occurs in mixed stands with white fir and incense cedar at lower elevations (and Sierra juniper in the north), and with red fir and lodgepole pine at higher elevations (Potter 1998). These stands are generally open, with canopy cover in the range of 40% to 65%, or less on rocky sites. Stand basal area for all tree species in jeffrey pine stands in Sequoia National Park have been reported over a range of 31 to 62 m 2 ha -I (Vankat 1970). The largest trees reach heights up to 61 m, with diameters up to 2.5 m (Van Pelt 2001). The relatively long needles of jeffrey pine and relatively open structure of these stands make for dry surface and ground fuels that burn readily. Thus, fires in jeffrey pine stands burn more frequently than those in adjacent red fir forests (Stephens 2001). Relatively pure stands of jeffrey pine occur on the east slope of the Sierra Nevada south of the Tahoe Basin (Fig. 17.8), mixing only at higher elevations with lodgepole pine and red fir. This dominance in areas that might otherwise seem appropriate for ponderosa pine may relate to colder temperatures. jeffrey pine commonly shares dominance with white fir and Sierra juniper in the Tahoe and Toiyabe National Forests (Vasek 1976). On the eastern slopes of the central and southern Sierra Nevada, jeffrey pines typically form relatively pure stands at 2,590 to 2,740 m elevation in a belt between red fir forests above and pinyon-juniper woodlands below. Typical understory shrubs beneath the open canopy jeffrey pine stands on this neast slope include Arctostyphylos patula, A. nevadensis, Ceanothus prostrata, C. cordulatus, Artemisia tridentata and Purshia tridentata (Vasek; Stephens 2001). Washoe pine (Pinus washoensis), a high elevation cousin of ponderosa pine, is locally important in the northern Sierra Nevada and southern Cascade Range (Hailer 1961; Smith 1994). Although seldom a stand dominant, it is the

474

MONTANE AND SUBALPINE VEGETATION

FIGURE 17.9 Stands with Pin LIS je(freyi can occur on the west slope of

the Sierra Nevada in the upper montane zone, but pure stands can also be found at 2,000 m on the eastside, as shown here. Mono County. Photograph courtesy of M. G. Barbour.

most important species above 2,400 m in the Bald Mountain Range and at about 2,290 m on dry slopes in the Warner Mountains (Hailer in Rundel et al. 1977). It reaches co-dominant status with jeffrey pine on Mt. Rose at elevations of 1,950 to 2,560 m. Red and white fir are less abundant in these stands. The open understory of jeffrey pine forests represents a mosaic of microsites with varying litter accumulation, shading, and moisture availability. Grasses are common on drier sites, including Elymus elymoides, Deschampsia elongata, and Achnatherum occidentalis, whereas Wyethia mollis and Monardella odoratissima are also frequently present.

LODGEPOLE PINE

Open stands of lodgepole pine forests make up a Widespread upper montane forest/woodland over much of the Sierra Nevada, tolerating both shallow rocky soils and semisaturated meadow edges, in an elevational belt within and above the red fir zone (Fig. 17.9) (Potter 1998). These forests, strongly dominated by Pinus contorta subsp. murrayana, generally occur at elevations of about 1,830 to 2,400 m in the northern Sierra Nevada, and rise to 2,440 to 3,350 m in the south. Stands of lodgepole pine may reach much lower; however, with cold air drainage down glacial canyons. The lodgepole zone is characterized by a short growing season of as little as 2 to 3 months, and the great majority of the 800 to 1,500 mm of annual precipitation falls as winter snow. The generally low stature and open stand structure of lodgepole pine forests is a function of these severe climate conditions and the thin, nutrient-poor soils that characterize this zone. Commonly there are few understory shrubs and little soil litter accumulation in these stands. Unlike the Rocky Mountain subspecies of lodgepole pine (P. contorta subsp. latifolia), which has a life history often tied closely to stand-replacing crown fires and often occurs in large even-aged cohorts, Sierran lodgepole pine does not

require fire for seedling establishment and typically occur in Inultiaged stands (Parker 1986). Large stand-replacing crown fires are rare in Sierran lodgepole pine stands (van Wagtendonk and Fites-KaufInan 1996). ()ver Inuch of its range in the Sierra Nevada, lodgepole pine attains heights of no more than 15 to 20 Ill, although individual trees Inay reach heights of 30 m and diameters of 60 to 130 CIll under better conditions with good soil resources (Rundel et a1. 1977). Stands of lodgepole pine forest salnpled in Sequoia National Park at elevations of 2,600 to 3,100 In had a Inean density of 3,390 trees ha -I, including saplings, and a basal area of 58 m 2 ha 1 (Vankat 1970). More than any other Sierra conifer, lodgepole pine is relatively tolerant of poor soil aeration, and thus grows well around the Inargins of wet 11leadows and other 1110ist areas (Rundel and Yoder 1998). Many upper montane and subalpine Ineadows in the Sierra Nevada exhibit invasion of young lodgepole pines Inoving inward from their drier Inargins. It is not clear how much this process has been intluenced by changes in fire frequency or grazing over the last 150 years. COIn parative photographic studies have shown changes in lodgepole pine stand densities in Yosemite (Vale 1987) and Sequoia National Parks (Vankat 1970). Attempts at restoration of plant cOllllnunities in disturbed areas of lodgepole pine forest have shown low to poor success (Moritsch and Muir 1993). Lodgepole pine is the overwhelming dOlninant within its forest cOllllnunity, nlixing occasionally with red fir, and with scattered Jeffrey pine and western white pine, and 1110untain helnlock at higher elevations (Potter 1998). It Inay occasionally fonn a treeline krulnmholz, growing with whitebark, lilllber, or foxtail pine. Scattered shrubs such as Arctostaphylos ncvadensis, Ribes rnontigenzl1ll, and Qllerclls vaccinit()lill are present within stands of lodgepole pine forest, but understory vegetation is generally sparse. Lake edges and wet Ineadows in the lodgepole zone support fringes of low ericaceous shrubs, most notably Cassiope 1rlertensiana, Vaccinizl1n cacspitoSll1rl, ]Jhyllodocc breweri, and Kalrnia fJolifolia. Fast-moving strealns are generally characterized by relatively dense populations of willows (Salix spp.).

Subalpine Vegetation

The subalpine landscape is comprised of a mosaic of subal pine forests/woodlands, 11leadows, rock outcrops, and scrub vegetation types. Although forests often comprise less than half or a third of the landscape, they have been the rnost studied vegetation types within the subalpine zone and are therefore emphasized in this section. Subalpine forests are open stands of conifers occurring on generally sandy soils or rocky slopes at elevations above the upper rnontane forest stands of red fir and lodgepole pine. Stand densities are low and trees rarely exceed 25 m in height (Fig. 17.10). Many, but not all, species form shrubby krulnlnholz fonns of growth near their upper elevational limits. The elevational distribution of subalpine forest COlllmunities varies with latitude. In the Lake Tahoe basin, such

stands begin around 2,450 In and extend up to treeline at 2,750 to 3,100 m (Graf 1999), whereas in the southern Sierra Nevada the range is more typically 2,900 to 3,660 m (Rundel et a1. 1977). Both upper and lower limits of subalpine species distributions are driven by a variety of factors, including soil resources, water availability, and climatic lirniting factors. Subalpine forests are characterized by a relatively short growing season with cool temperatures. Frequently the season of vegetative growth lasts no longer than 6 to 9 weeks, and frosts can occur at any time of the year. Annual levels of precipitation are 750 to 1,250 mm. With the exception of occasional summer thunderstorms, almost all of this precipitation falls as winter snow. Wet years with abundant snowfall can limit growth as these may produce late-lying snowfields that reduce the length of the growing season (Armstrong et a1. 1988). Winds are often severe, particularly around exposed ridges. Such wind conditions may produce snow-free winter areas that lower soil telnperatures and increase plant water stress. Most of the subalpine areas of the Sierra Nevada were subjected to repeated glaciation during the Pleistocene, and thus have thin and poorly developed soils with little organic matter. The small amounts of litter accumulation and open stand structure of subalpine forests means that fire is rare. Because of the solid granite parent material, areas with deeper soil accumulation can become waterlogged for much of the year. For these reasons, the length of the growing season is a function of early season limitation due to low temperatures and snowfields, and late season limitations due to drought. Studies of the dynamics of alterations of treeline elevation over the past several millennia using tree ring chronologies have reinforced the significance of complex interactions of both temperature and water availability in determining such changes (Lloyd and Graumlich 1997; Bunn et a1. 2(05). Responses of high-elevation forests to future global warming may depend strongly on water supply. Treeline growth of multistemmed trees and shrubby krulnlnholz growth of conifers varies with latitude in the Sierra Nevada. Treeline in the northern Sierra Nevada is dominated by whitebark pine (Pinus albicalllis), which frequently occurs with a krummholz form of growth near its upper limit (Fig. 17.11). Several other species may also form krulllmholz growth forms, including Sierra juniper, mountain hemlock, lodgepole pine, and rarely Jeffrey pine. Further south in the central Sierra Nevada, limber pine (Pinlls t1exilis) is the dominant treeline species, particularly on the east slope of the range, often occurring with lodgepole pine. Finally, treeline in the southern Sierra Nevada is characterized by open stands of foxtail pine (Pinus ba1t()uriana), a species that does not typically develop krulnmholz growth forms. Despite silnilarities in adaptations to their high elevation environment, these three subalpine pines COlne froln very different lineages and are not closely related. The genetic structure of many subalpine pines, most notably P. albicalllis, P. tlexilis, and P balfollriana, is strongly

MONTANE AND SUBALPINE VEGETATION

475

FIGU RE 17.10 (a) Mixed subalpine woodland dominated by whitebark pine, with a prominent understory of pinemat manzanita. Slide Mountain, 2,700 m, east of Lake Tahoe. (b) Subalpine stands are typically fragmented by intervening meadows and barrens, as here in Yosemite National Park above Tuolumne Meadows. Photograph (a) courtesy of M. G. Barbour, (b) from]. Fites-Kaufman.

influenced by the seed-caching behavior of Clark's nutcracker (Nucifraga columbiana). These birds actively gather huge quantities of pine seeds and disperse them, often into caches where multiple seeds may germinate (Tombach and Kramer 1980; Tombach 1982, 1986). Tree-ring analyses of upper treeline species in the Sierra Nevada have been compared to instrumental records to understand growth responses to temperature and moisture availability. Extrapolating these data back before instrumental records has given a long-term record or estimated temperature conditions. Rates of annual tree growth in foxtail pine, lodgepole pine, and western white pine are influenced by positive non linear interactions between summer temperature and winter precipitation (Graumlich 1993). Although maximum growth rates occur under conditions of high winter precipitation and warm summers for all three species, substantial species-to-species variation occurs in the response to these two variables (Graumlich 1994). L10yd and Graumlich (1997) reconstructed a 3,500-year history of fluctuations in treeline elevation and tree abundance in the southern Sierra Nevada using tree-ring records. Treeline elevation was higher than at present throughout most of this period. Declines in the abundance of live trees and treeline elevation occurred twice during the last 1,000 years. An elevational decline from 950 to 550 years BP coin-

476

MONTANE AND SUBALPINE VEGETATION

cided with a period of warm temperatures (relative to present) in which at least two severe, multidecadal droughts occurred. A second decline from 450 to 50 years BP was apparently triggered by an increase in the rate of adult mortality in treeline forests. This latter decline occurred during a period of low temperatures lasting for up to 400 years and was apparently caused by a sustained failure of regeneration in combination with an increased rate of adult mortality. Pollen and microfossil records from Sierran lakes have provided additional useful data in interpreting late Holocene changes in subalpine species occurrences. The upper altitudinal limits of many subalpine conifers began to decline about 2,500 years BP, coincident with the beginning of Neoglacial cooling (Anderson 1990). Tree-ring studies of recent growth trends of whitebark pine and lodegpole pine in a subalpine forest of the southern Sierra Nevada have shown a dominant trend of increasing basal area increment over time in all age classes of both species (Peterson et a!. 1990). This increased growth rate in subalpine trees is widespread and relates to the complex interactions of many factors including ambient CO 2 levels and their effect on water use efficiency (Graumlich 1991; Bunn et a!. 2005). Climatic variables account for a relatively small portion of the variance in short-term tree growth, and there is no clear relationship with long-term growth.

FIGURE 17.11 Krummholz of Pinus albicaulis, above Sonora Pass, 3,200 m elevation. The "trees" are crowded together into a continuous hedge about 60---70 cm tall (note blue back-pack for scale). Photograph courtesy of M. G. Barbour.

MOUNTAIN HEMLOCK

Mountain hemlock (Tsuga mertensiana) has a broad range extending from the coastal ranges of Alaska south through British Columbia, and the Pacific Northwest into the Sierra Nevada. In the northern Sierra Nevada it may be found in upper montane forests of red fir and lodgepole pine (Potter 1998), but it is more characteristic at higher elevations near treeline where it is often the most common tree species but often mixed with Sierra juniper and whitebark pine. Mountain hemlock thrives in moist but well-drained mountain soils, often showing a preference for north-facing slopes. This is in contrast to stands in the Cascade Range where greater summer precipitation and cooler temperatures provide broader topographic conditions for growth (Parker 1994, 1995). In the central Sierra of Yosemite National Park it often grows up to 30 m in height in extensive groves with a virtually closed canopy. Seedlings are relatively shade tolerant compared to other subalpine conifers and do well under this type of canopy. At higher elevations, however, mountain hemlock is more scattered and often assumes a lower, shrubby growth form. South of Yosemite, mountain hemlock becomes increasingly restricted to small stands groWing on favorable microsites in cold moist valleys and sheltered ravines where snow banks remain late into the summer. Unlike pure stands of the central and northern Sierra Nevada, these scattered trees in the southern portions of the range are commonly found mixed with lodgepole pine, foxtail pine, western white pine, and red fir. The southernmost occurrence of mountain hemlock is below Silliman Lake in northern Tulare County where there is a small grove of about 60 trees with heights up to 24 m, diameters to nearly 90 cm, and healthy reproduction (Parsons 1972).

but massive trunk appearing to grow out of seemingly solid granite substrate. This Sierran subspecies of western juniper is differentiated from the subspecies occiden ta lis, which grows in the Cascade Range with ponderosa pine and Great Basin sagebrush. Sierra juniper occurs on shallow soils from 2,100 to more than 3,000 m elevation, often growing with Jeffrey pine, red fir, whitebark pine, mountain hemlock, and/or lodgepole pine. More than any other subalpine tree, Sierra juniper shows a remarkable ability to colonize and grow successfully out of small fractures in granite domes that would not support other species. At the lower margins of lodgepole pine forest in the Tahoe Basin, there are mixed stands of Sierra juniper with red fir and Jeffrey pine, but these associated tree species are replaced by western white pine and mountain hemlock with increasing elevation. The largest Sierra juniper is reported to occur farther south in the Stanislaus National Forest, where there is a tree 26 m in height and 4 m in diameter (Lanner 1999). Sierra juniper may occur mixed in lodgepole pine stands up to treeline, where it increasingly takes on a krummholz form of growth. Some of these junipers are reported to reach ages of over 1,000 years (Graf 1999). WESTERN WHITE PINE

Western white pine (Pin us monticola) is locally abundant in subalpine habitats along the west slope of the Sierra Nevada, where it may occur in small pure stands but more commonly is found mixed with lodgepole pine, Jeffrey pine, mountain hemlock, and red fir (Potter 1998). Although Sierran trees of this species may reach 40 m in height and 2.5 m in diameter, these sizes are smaller than those reached by the same species in the northern Rocky Mountains and Pacific Northwest (Van Pelt 2001). Western white pine generally maintains a tree form of growth up nearly to treeline, where it is commonly replaced by whitebark pine or foxtail pine on rocky ridges. Seedlings are reported to be relatively few compared top other subalpine conifers (Parker 1988). A detailed analysis of tree distributions in a subalpine watershed dominated by western white pine has been made for Emerald Lake in Sequoia National Park. This watershed extends from 2,804 m at Emerald Lake up to 3,415 m at Alta Peak. A complete census of 1,206 trees in the watershed showed western white pine as the strong dominant with 71% of the individual trees and 83% of the basal area (Table 17.9). Lodgepole pine, largely restricted to mesic benches, was second in abundance with 17% of the trees, but formed only 3.5% of the basal area. Foxtail pine comprised 9.5% of the trees and 13.7% of the basal area, with these largely occurring on high north-facing ridges in the watershed. Small numbers of Jeffrey pine and red fir were largely restricted to the lowest elevations in the watershed in mesic bench habitats (Table 17.9).

SIERRA JUNIPER

WHITEBARK PINE

Sierra juniper (Juniperus occidentalis subsp. australis) is one of the most striking trees of the Sierra Nevada with its short

The most widespread treeline conifer in the Sierra Nevada is whitebark pine (Pinus albicaulis), which occurs abundilntly

MONTANE AND SUBALPINE VEGETATION

477

TABLE 17.9

Total Census Data for Conifers in the Emerald Lake Watershed (2804-3415 m) of Sequoia National Park

Relative Density ((M})

Species

Relative Basal Area

Mesic Bench

SW-facing Ridge

N-facing Ridge

Wet Meadow

Steep Talus Slope

Pinus monticola

71.2

48.7

27.7

9.0

8.4

6.2

Pinus contorta subsp. murrayana

16.7

68.3

17.8

1.5

11.4

1.0

Pinus balfouriana

9.5

16.8

77.0

4.4

1.8

Pinus jeffreyi

0.9

100.0

Abies magnifica

1.7

65.0

Total

35.0

100

Relative Basal Area «}9091) of these stems. Krummholz individuals of both pine species formed 39(}h of the total number of stems. The basal area of pines varied from 1.6 to 21.0 m 2 ha- 1 in subunits of the watershed, with lodgepole pine accounting for more than half of this basal area. The mean leaf area index of canopies of lodgepole pine was calculated to be 4.1 m 2 m- 2 , compared to 4.6 m 2 m- 2 for whitebark pine. The protein-rich and fat-laden pine nuts of whitebark pine form a staple food supply of Clark's nutcrackers (Tomback 1982, 1986). These birds collect massive numbers of these seeds and cache them in the soil. Studies in

the Sierra Nevada have reported that single adults Clark's nutcrackers can cache as many as 89,000 pine nuts in a season, far in excess of their short-term nutritional needs. These caches are the primary means of reproduction for whitebark pines, and thus have strong genetic consequences. Genetic variation is highly structured in within the natural groupings of krummholz thickets and upright tree clumps. Genetic studies have shown that multiple individuals are present within krummholz thickets, and genetic relationships often resemble half- to full-sibling family structure (Rogers et al. 1999). At lower elevations, most (72%) of the tree clumps contained more than one genotype; the remaining clumps appeared to be multistemmed trees. LIMBER PINE

Limber pine (Pinus flexiiis) is widespread in scattered stands over an area of the east slope of the Sierra Nevada from Mono Pass south to the Inyo National Forest east of Sequoia National Park. In this belt it appears to fill the niche occupied by whitebark pine to the north. Although also present in the higher Transverse and Peninsular Ranges of Southern California, limber pine is more typical of the White Mountains in California and eastward across the Basin and Range Province to the Rocky Mountains. Within its Sierra range, limber pine occupies steep, eroded, and/or nutrient-poor sites at or near treeline. It may occasionally be found on more mesic sites but under such conditions has been considered to be a pioneer species. Like whitebark pine, treeline individuals of limber pine frequently form low cushions of krummholz growth. Studies of limber pine across sites in the Sierra Nevada and southern California concluded that its distribution and abundance are limited by competition from other species. It is most dominant and persists longest on steep, dry slopes at higher elevations. Trees reaching 2,000 years of age have been reported, but such great ages may be atypical. Although similar in general appearance, limber pine is not closely related to whitebark pine. However, limber pine in its development as a subalpine species has convergently evolved large pine nuts that rely on Clark's nutcrackers rather than on wind for seed dispersal (Tombach and Kramer 1980; Carsey and Tomback 1994). FOXTAll PI N E

The dominant subalpine and treeline pine of the southern Sierra Nevada is foxtail pine (Pinus balfouriana). The disjunct distribution of this species-the southern Sierra Nevada and the Klamath Mountains of northern California-is unusual, but in some respects mirrors the disjunction of Abies magnifica var. shastensis between the southern Sierra Nevada and the Cascade and Klamath Ranges. These two subspecies of foxtail pine are well differentiated, with the southern Sierra Nevada taxon (subsp. austrina) morphologically distinct in many characteristics of the foliage, bark, cones, and seeds from populations from the subspecies balfouriana

restricted to the Klamath Mountains (Mastrogiuseppe and Mastrogiuseppe 1980). Despite this morphological differentiation and nonexistent gene flow between the subspecies, there are much higher levels of differentiation among populations within the Klamath Mountains subspecies than between the two subspecies (Oline et al. 2000). This high genetic differentiation is attributed to genetic divergence among small isolated populations in the northern subspecies. In many respects the subsp. austrina shows closer links to its cousin, Great Basin bristlecone pine (Pinus longaeva) in the White Mountains. Foxtail pine is closely related to both Great Basin and Rocky Mountain bristlecone pine, with the three species forming a close evolutionary lineage. In the Sierra Nevada, foxtail pine is restricted to the higher elevations of 2,600 to 3,660 m south of the Middle Fork of the Kings River. At its lower elevational limits it may occur in open stands with lodgepole pine, Jeffrey pine, western white pine, red fir. At higher elevations, however, it forms relatively pure but low density stands. Vankat (1970) sampled foxtail pine stands at 3,170 to 3,290 m elevation in Sequoia National park and reported a mean canopy cover of 26%, with a basal area of 31 m 2 ha- I and a density of 418 tree ha-I. Not surprisingly, the mean basal area of adult foxtail pine and density of seedlings declined with increasing elevation from forest to treeline stands (Lloyd 1997). These declines were also associated with lower nutrient inputs from aboveground litter and lower litter C:N ratio. However, neither nitrogen concentration of seedling needles or nitrogen relative accumulation rate differed significantly across elevations. Models of site moisture availability and irradiance coupled with field measurements of stand characteristics and tree ring records suggest that there are strong correlations of microsite conditions with age-class structure and ringwidth patterns (Bunn et al. 2005). Foxtail pine is shade intolerant at all stages of growth and prefers shallow, well-drained soils on exposed sites. Like other subalpine pines, it has deep, spreading root systems that tap snowmelt in fractures of the rocky granite soil. Unlike whitebark and limber pine, which typically form krummholz at treeline, foxtail pine retains an upright form of growth throughout its elevational range. Despite its high elevation of occurrence, foxtail pines can reach large sizes. A tree at 3,250 elevation on Alta Peak was reported to be 24 m in height and 2 m in diameter. Mastroguiseppe and Mastroguiseppe (1972) suggested that foxtail pine may well reach ages of 2,500 to 3,000 years.

Nonconiferous Vegetation DECIDUOUS FOREST

A number of broad-leaved tree species are also found in the Sierra Nevada. Those associated or intermixed with

MONTANE AND SUBALPINE VEGETATION

479

coniferous forests, such as black oak, big-leaf maple, and mountain dogwood have already been discussed in previous sections. Here the primary focus is on the more widespread species that occur widely, in particular aspen (Populus tremuloides). But first, a brief synopsis of other common deciduous species and their distribution is included. The majority of the deciduous forest types and species discussed are associated with riparian or wetland sites. Due to space limitation, there is not a comprehensive treatment of riparian vegetation but rather a broad overview of some of the dominant trees. Most of the deciduous trees in the Sierra and northeastern California are associated with riparian areas or subsurface water but may also occur on particularly rocky soils. Buckeye (Aesculus californica) occurs in small patches in the lowest margins of yellow pine forests on the western slopes of the Sierra. White alder (Alnus rhombi folia) is common along perennial streams and seeps in the montane forests on the western slopes of the Sierra, generally between 750 and 2,400 m. It often occurs intermixed with conifers but also by itself in small patches where the root zone is perpetually saturated. Cornus sessilis, a low growing tree or tall shrub 2600 N = 23

Elevation (m)

c

160 140 120 100 80 60 40 20 0

0-10 N = 40

10-20 N = 13

20-30 N = 28

30-40 N = 45

40-50 N 40

=

50·60 N 29

=

60-70 N 16

=

>70 N = 37

Time-since-fire (years) Fire-scar dendrochronology (FSD) studies in SeA reveal that trees were scarred by fires at rates of several times per century, but few trees were scarred since 1900 (McBride and Laven 1976; Everett 2003). Extensive forests have reached canopy closure, with dominance shifting from sun-tolerant ponderosa

and Jeffrey pine to shade-tolerant white fir and incense cedar in the subcanopy. Replication of VTM plots in SBM show that stem densities (> 10 cm dbh) had doubled since 1929 to 1932, ranging from 260 ha- 1 in white fir to 160-180 ha- 1 in other assemblages (Fig. 18.6; Minnich et al. 1995; cf. Vogl and Miller

SOUTHERN CALIFORNIA CONIFER FORESTS

519

400 Stem diameter class (dbh inches)

0 350

24-36

~

300

>36

•

12-24 4-12

250 ~

Cu

~

FIGURE 18.8 Tree-stem diameter distri-

butions for various montane forest stands in the San Bernardino Mountains (B) and in the Sierra San Pedro Martir (M) in 1932 and 60 years later in 1992.

(/)

E Q}

200

.-

Cl)

150

100

50

Location 0 B B B B M B B Year 32 92 32 92 92 32 92 Vegetation Ponderosa Ponderosa Monotypic Pine Pine Jeffrey (logged) Pine

1968). Densification rates are proportional to gradients in average annual precipitation (AAP), with increases of 300% in forests with AAPs of 100 cm, and no increase in forests with AAPs of 40 cm. The highest densities (380 ha-I) occur in logged ponderosa pine forest where harvest disturbance resulted in abundant recruitment beginning in 1850 to 1870, whereas densification began after 1900 in unharvested forests (Minnich 1988; Minnich et al. 1995). The density of stems 60 cm since 1932 may be related to competition from younger cohorts that weaken large trees, making them susceptible to infestations of bark beetles, pathogens and air pollution (Minnich et al. 1995; cf. Stone et al. 1999).

520

SOUTHERN CALIFORNIA CONIFER FORESTS

M B B

M B B

B B

92 32 92 Jeffrey Pine

92 32 92 White Fir

32 92 Singleleaf Pinyon

Elsewhere in SCA, modern forest structure varies with topography and climatic gradients (Fig. 2.17 in Stephenson and Calecarone 1999). Open forests are widespread in the rugged SGM and most leeward slopes of the Transverse Ranges. Dense forests grow on moist windward slopes such as Pine Mountain Ridge, Idyllwild, Mt. Palomar, and Mt. Cuyamaca. Few studies have quantified presuppression fire regimes in SCA. A FSD study in SBM, based on 15 samples of single trees at 1.0-km intervals, estimates that fire intervals were 15 to 29 years (McBride and Lavin 1976). Everett (2003) working on two 280-ha grids in SJM and SBM found that fires scarred at least one tree in each grid every 5 years. Spatially reconstructed rotation periods-the interval between scars of each tree-was 32 years on Black Mountain and 49 years at Big Pine Flat. Aerial photographs in the late 1930s show presuppression stand-replacement burns on steep slopes and ridges with subcanopy of brush and canyon live oak as well as dense harvested forests near Lake Arrowhead

(Minnich 1988). Stand-replacement burns were mostly 10(yb of canopy layer trees in 2291) of forests, but only 3q;{) of forests were denuded in stand-replacement burns. Forest stand-replacement gaps were mostly 30(1) of forests within fire perimeters have sustained stand-replacement burns, or an order of magnitude greater than in BCA (Minnich 1999). Postfire succession following surface burns was characterized by a slight increase in cover of shrub species and establishment of forest tree seedlings, especially ponderosa and ]effrey pines (McBride and Laven 1999). In stand-replacement fires, postfire succession is dominated by a shrub phase of species establishing from seedbanks (Ceanothus integerrimus, C. palrneri, C. cordulatus, Arctostaphylos patula) or recruit continuously from seed cached by fauna (Rhanznus californica, Arctostaphylos pringlei, A. pungens). Rhalnnus californica and Ceanothus cordulatus resprout and Chrysolepis sempervirens establishes new stems from rhyzomitous roots. Canyon live oak and black oak respond by epicormic sprouting after surface fires, and basal sprouting after severe burns (Plumb 1979). Black oak recruitment and resprouting promote dense thickets in forest openings (cL Kauffman and Martin 1990, 1991). Stand-replacement burns as early as 1919 and 1922 have persisted as canyon live oak and black oak woodlands (Albright 1998). On leeward slopes Cercocarpus ledifolius, a nonsprouter, is replaced by Chrysothamnus nauseosus and Artemisia tridentata, both species establishing from seed dispersed by wind. C. ledifolius recruitment gradually establishes a subcanopy in 30 to SO years. Conifer recruitment occurs continuously from

SOUTHERN CALIFORNIA CONIFER FORESTS

521

seed dispersed by wind or fauna. Studies of long-term conifer succession have not been undertaken. PATCH MOSAIC FOREST MODEL-SUBALPINE FOREST AND PINYON-JUNIPER WOODLAND

Stand-replacement fires periodically denude pinyon-juniper woodland and subalpine forests, but only at intervals of centuries due to low primary productivity rates (Minnich 1988; Wangler and Minnich 1996; Minnich and Chou 1997; Minnich et al. 2000; see Table 18.1). Stand-replacement burns result in stepwise increases in fire intervals to an order of centuries because the removal of the tree layer discourages short-term fire recurrences. Postfire shrub successions yield low fuel accumulation rates compared to the chaparral on coastal slopes. Even partly burned trees perish because the thin bark permits fatal cambium damage. Fires leave discrete patch structures that fade after ca. 100 years when trees develop mature stature and subcontinuous canopy closure. Stand-replacement fires eliminate carryover biomass that potentially contribute to short-interval fires, and low annual precipitation limits fuel buildup rates in postfire shrub successions compared to coastal chaparral. Using FSD analysis, Sheppard and Lassoie (1998) found that fires in lodgepole-limber pine forest on Mt. San Jacinto summit were single tree events. Ancient limber pine forests at this site, Mt. Baden-Powell (Thorne 1988) and Mt. San Gorgonio, have robust twisted boles and bark striping reminiscent of bristlecone pines (Pinus longaeva) in the Great Basin ranges due largely to single tree burns. Lightning detection densities predict that virtually every tree is struck by lightning by the time it reaches 1,000 years age. Aerial photographs in 1938 record linear treeless gaps in dense timberiand chaparral on south-facing slopes, the gaps apparently an outcome of terrain-channeled fire runs (Minnich 1988). Dense forests on north-facing slopes exhibit discrete even-aged patches from stand-replacement fires that collectively account for 10% of forests. It is unknown whether presuppression burns were crown fires or surfaces fires causing fatal cambium injury. Lodgepole pine forests undergo variable fire patterns from low intensity ground fires to canopy fires in extreme weather (Kilgore and Briggs 1972; Kilgore 1981; Parker 1986; Agee 1993). The trend for stand-replacement fires in pinyon-juniper of SBM and SJZ (Wangler and Minnich 1996; Minnich and Chou 1997; Minnich et al. 2000) supports fire-regime studies in dense forests of the Southwest (Erdman 1970; Koniak 1985; Floyd, Romme, and Hanna 2000). Fire sequences show evidence of nonrandom patch turnover as cumulative fuel buildup with gradual densification of canopy appears to be more significant in fire occurrence than short-term fluctuation in fuels due to climatic variability. Stand-replacement burns in subalpine forests are replaced by timberland chaparral of Ceanothus cordlllatus, Arctostaphylos patllla, and Chrysolepis sempervirens (Minnich 1978). Lodgepole pine recruitment reaches preburn densities in ca. 20 years, the trees establishing from wind-dispersed seed

S22

SOUTHERN CALIFORNIA CONIFER FORESTS

FIGURE 18.9 A nineteenth-century stand-replacement fire in subalpine forest on CharIeton Peak near Mount San Gorgonio, in 188S. Photograph courtesy of the SmiIey Library, Redlands, CA.

from unburned stands. Forest gaps recorded in photographs taken a century ago (Fig. 18.9) now host young trees of mature stature (Minnich 1988). Both lodgepole and limber pine are shade intolerant, limiting recruitment in stands with canopy closure. Limber pine has thicker bark and is more tolerant of surface fires than lodgepole pine, possibly explaining its greater longevity in harsh habitats. Early pinyon-juniper successions are dominated by a shrub layer of species establishing by long-range seed dispersal (Artemisia tridentata, Chrysothamnus spp.) and soil seedbanks (Ceanothlls greggii, Fremontodendron californicllm), as well as from resprouting (Purshia tridentata, Quercus johntuckeri, canyon live oak; Wangler and Minnich 1996). Although pinyon seed and juniper berries are Widely dispersed by birds and mammals (Van der Wall and Balda 1977; Van der Wall 1997), recruitment appears to be delayed 20 to 30 years until the establishment of the shrub layer which acts as nurse plants that protect seedlings from high soil temperatures, soil heaving, and predation by rodents (Wangler and Minnich; cf. Chambers 2001). The first pinyon recruits establish within shrub canopies, often within 2 cm of root axes. The development tree canopy reduces freeze-thaw processes after ca. 75 years, followed by spatially random recruitment throughout burns. The development of tree canopy after 100 to 150 years is accompanied by a decline in the shrub layer. Mature stands are mixed-aged due to continuous recruitment typical of white pines, with densities of 150 to 250 stems ha-I. In semiarid woodlands of Joshua Tree National Park (Minnich 2003), conifers perish largely from flames lines generated by a understory shrub layer. Woodlands are replaced by recruitment of Hymenociea salsola, Salazaria mexicana, Viguiera parishii, Eriogonum fasciculatum, and Ericameria cooperi. Obligate sprouters (Lyeium spp., Ephedra spp.) persist at low densities. Fires correlate with wet episodes of climate due to abundant fuels provided by bunch grasses (Pleuraphis rigida, Achnatherum speeiosa) and subshrubs, as well as annual wildflowers, all of which increase canopy in wet episodes and die back in drought. After unprecedented rains in 2004-2005,

cured wildflowers and native grasses (Aristida, Bouteloua) were ilnportant fuels in a 36,000 ha outbreak in the eastern San Bernardino Mountains and Joshua Tree National Park. The role of invasive annuals BrornllS rllbens and B. tectOrllll1 that have proliferated the past three decades is unclear. In the 6,OOO-ha fire complex of 1999, bromes had little effect in old-growth woodlands that burned at comparable intensities as chaparral. In reburn zones, bromes supported creeping, low-intensity flame lines that left shrubs unburned or scorched; that is, postfire successions were not breached. Mediterranean exotics do not represent novel herbaceous fuel hazard in deserts. Suppression appears to have little impact in pinyon-juniper and subalpine forests because fire-free periods are longer than the suppression era. Increases in fire intensity may not change fire mortality rates because stand-replacement fires characterize these ecosystelns. Forest densities and diameter frequency distributions of pinyon-juniper woodlands have not significantly changed since the VTM survey (Wangler and Minnich 1996).

Fire and Biogeography of Southern California Conifer Forests

The "real world" of forest fires, of course, does not play out in discrete fire regimes in lock step with forest distributions. Instead, it is best to view fire relationships as gradients in properties in a continuum of environmental change that exert intense selective pressure on the species composition and spatial pattern on the landscape. In southern California, forests distribute along a corresponding "pecking order" of resilience. High-resilience assemblages (i.e., species with life traits that include rapid establishment, short generation tilnes, and short life spans) grow in productive chaparral on windward slopes, are recycled by frequent severe fires, and grow on steep, undissected concave slopes subject to intense fire behavior. Low-resilience forests (slow recolonization, long generation times, long-lived) survive surface fires in low chaparral fuel loads in mesic productive environments on gentle surfaces or steep convex canyons, or occur on unproductive leeward slopes and the highest sumnlits. Successive fires, of course, leave a unique overprint of range expansions and contractions resulting from stochastic variability of fire regilne properties. It should be expected that forest distributions and stand properties over multiple fire cycles would oscillate around steady-state distributions as a function of random fire occurrence with weather, ignitions, and patch mosaic status. However, we can only dream of century-scale time-series empirical data necessary to assess forest equilibria. The following statements-based on space-for-time substitutions of disturbance and succession patterns-can be treated as hypotheses for future research. The linlit of mixed-conifer forest on moist windward slopes is a conspicuous ecotone in which tree cover with sparse chaparral understory is replaced by dense cover of chaparral (Minnich 2001). Intense chaparral fire cycles selec-

tively eliminate mixed conifers in favor of treeless zones, or colonies of serotinous conifers with vigorous establishment traits. Competitive exclusion processes are suggested by the observation that mixed-conifer forest grows as low as 1,300 m in shrub free basins, and are absent from steep, undissected chaparral slopes to 2,400 m, across large gradients in climate. Trees must also compete with chaparral for soil water. Intense fires selectively eliminate mixed conifers colonizing bigcone Douglas fir forest, which persist by resprouting (Minnich 1988). Above 2,000 m, mixed conifers coexist in open montane chaparral and arboreal canyon live oak because ground fires leave the tree layer. Intense fires convert canyon live oak into shrublands, elilninate mixed conifers and selectively favor Coulter pine over bigcone Douglas fir. This hypothesis is consistent with the distribution of bigcone Douglas fir in fire-resistant canyons and Coulter pine on concave slopes and ridgelines. The sprouting habit of bigcone Douglas fir may be a "last resort" life trait that maintains small groves in a hostile regime of intense chaparral fires below ca. 1,200 m. Low-intensity fires select for downslope expansion of lnixed-conifer forest, the permanent tree canopy selectively displacing shade-intolerant chaparral and serotinous conifers. Bigcone Douglas fir stands expand into arboreal cover of canyon live oak, with selective elilnination of serotinous conifers. It appears that bigcone Douglas fir does not invade mixed-conifer forests except on rapidly eroding slopes vital to its establishment. Suppression policies have already selected for the displacement of ponderosa and Jeffrey pine by shadetolerant white fir and incense cedar. The scarcity of Coulter pine-oak woodlands suggests that this asselnblage may be unstable over long time scales because the pine is short-lived and shade intolerant. A succession of nonfatal surface fires may encourage its replacement by long-lived mixed-conifer forests. Emerging stand replacement fire cycles may reverse this trend by selecting for replacement of mixed conifers by short-lived serontinous conifers. The mixed-conifer/subalpine forest ecotone coincides with a discontinuous shift froln frequent surface burns to infrequent stand-replacement burns. The pennanent canopy of mixed-conifer forest discourages the establishment of lodgepole pine, which is selectively eliminated by recurrent surface fires fueled by cumulative litter buildup. High-intensity fires select for the downslope displacement of subalpine forest into mixed-conifer forest. Mineral seedbed conditions favor lodgepole pine establishment, resembling successions in the Rocky Mountain subalpine forest. The destruction of the tree layer increases the chance for long fire intervals due sparse litter in mature subalpine forests. A discontinuity in fire intervals also occurs at the lnixed-conifer forest-pinyon-juniper woodland ecotone (Minnich and Chou 1997; Minnich et a1. 2000). Recurrent surface fires sustained by continuous litterfall from surviving tree canopy selectively elinlinate pinyons in mixed-conifer forest. The removal of canopy in pinyonjuniper woodland precludes short interval fires. Drought-tolerant mixed-conifers, especially Jeffrey pine, readily establish

SOUTHERN CALIFORNIA CONIFER FORESTS

523

in pinyon-juniper woodlands. However, mixed-conifers are preferentially extirpated in extreme drought (see below). The distribution of pinyon-juniper woodlands overlaps with extensive areas of open desert chaparral assemblages but rarely extends into contiguous chaparral on coastal slopes. Desert chaparral comprises long-lived species including Quercus turbinella, Q. comelius-mulleri, Arctostaphylos glauca, A. glandulosa, Cercocarpus beruloides, Rhus ovata, Prunus ilicifolia, and Rhamnus crocea. The longevity of shrubs diminishes the importance of seed reservoirs in stand maintenance. Although members of Ceanothus subg. Cerastes and Fremontodendron californicum establish local thickets in early postfire succession, fire intervals are longer than the life span of these shrubs. Hence, they do not contribute as fuels in fire sequences. The low productivity and open-stand structure of desert chaparral assure long fire intervals compatible with pinyon-juniper colonization. Short fire intervals select against pinyon-juniper, enhance recruitment of obligate seeding shrubs in Adenostoma and Ceanothus, and increase chaparral cover. Pinyon-juniper is most extensive on leeward slopes > 1,500 m, with cool climate, low annual precipitation and productivity. It is absent from low passes or mountain crests with high lee slope precipitation and productivity.

Establishing a Presuppression Baseline

A fundamental goal in fire ecology is to establish a presuppression baseline to assess how natural fire is integrated into forest systems and how suppression has changed them. There is evidence that several forest types have been little affected by twentieth-century fire control. In subalpine forest and pinyon-juniper woodland, fire-free periods are far longer than the suppression era. The earliest aerial photographs record virtually the same forest of stems as at present and suggest limited population turnover, even as a scale of a century. They also give a clear record of presuppression standreplacement burns in long-lived pinyon-juniper and subalpine forests, but provide only one generation of fires. In the chaparral belt, cohort regeneration after stand-replacing fires appears to characterize closed-cone conifer forest before and during suppression (Minnich 1988). Early aerial photographs reveal that presuppression patch structure in chaparral and closed-cone conifer forest was already emplaced by suppression era burns. However, enormous change has occurred in mixed-conifer forest. Dense" dog-hair" forests are now pervasive in southern California. Unfortunately, the fire process shaping dispersed forest structure of course was not rigorously documented at the time. Ideally, the reconstruction of baseline structure and dynamics of mixed-conifer forest requires spatially explicit data over long time scales. Spatially explicit data include fire perimeter histories, but the data begin in 1910 with the establishment of the National Forests, Le, records began only when suppression was initiated (Minnich 1988). Perimeters accurately depict the removal of biomass in stand-replacing events, but they prOVide no information on

524

SOUTHERN CALIFORNIA CONIFER FORESTS

fire severity and tree mortality in surface fires. Time-series aerial photographs have the advantage of providing raw records of forest status over large regions, including the limits of burns, removal and survival of canopy, population dynamics, and intraburn successions. Mixed-conifer forests were already in incipient densification (Minnich et al. 1995). In BCA, time-series photographs record uncontrolled fire regimes only to the early twentieth century. Studies have utilized space for time substitutions, such as synoptic analyses of time-series aerial photographs. This approach assumes that spatial and temporal variation are equivalent, but this assumption cannot account for transient effects such as shifting climate (Pickett 1985). The most widely used method for mixed-conifer forest is fire scar dendrochronology (FSD), which has the primary advantage of capturing fire records over many centuries, but it is a sitespecific methodology limited to forest systems with surface fire regimes. Likewise, time for space substitutions, such as FSD studies, assume the same equivalency but cannot account for spatial effects (see also Baker and Ehle 2001). A centerpiece of the FSD method is the estimation of mean fire return intervals (FRI) from ring counts between successive fire scars on tree bole catfaces. Studies have sampled scar records of single trees and multiple tree samples as small aggregates of point samples (Agee 1993; Taylor 2000; Grissino-Mayer 2001; Morgan et al. 2001). Recently FSD methods employed spatial sampling protocols to reconstruct fire histories at scales of thousands of ha (e.g., Brown, Kaufmann, and Shepperd 1999). A key issue in FRI reconstructions is how fire scars represent other fire regime properties in broad spatial scales. The assumption that FRIs can be substituted for spatial pattern must take into account scarring efficiency, relationships between fire size and intensity, and how scarring correlates with fire size frequency distributions. How does scar data extrapolate to the landscape when fire size frequency distributions are skewed? (See ]ohnson and Gutsell 1994.) Time-for-space substitutions are valid only if a sample represents a discrete event. But fire scars represent multiple events of unknown distribution except specifically to the sample sites, and spatial extrapolation is invalid because the skewed fire size frequency distributon does not permit it. Single tree samples are considered to conservatively underestimate past fire occurrence for a given point on the assumption that many fires do not scar trees, Le, intervals are shorter than measured (Skinner and Chang 1996). Alternatively, trees may scar easily once they form catfaces because heartwood ignites more readily than bark. An important step in the interpretation of FSD would be inductive experimentation. There are no studies in which scarring efficiency has not been independently tested against field burn experiments, where fire extent and severity are known. Fire intensity has not been directly evaluated from fire scars. Hence scarring efficiency cannot be directly estimated using FSD methods. Moreover, intensity estimates from stand age frequency distributions are not possible because forests

represent survivorship; that is, there is no direct evidence of fire processes that selectively remove canopy. However, stand-replacement burns can be inferred from the spatial pattern of the oldest trees ("age caps"; e.g, Brown, Kaufmann, and Shepperd 1999). Fire spread models predict that small fires have low intensity because energy release is below thresholds required to maintain flame lines (Rothermel 1972; Scott and Burgan 2005). Above this threshold there is uncertainty as to whether flame front intensities continue to increase with fire size (fractal; Minnich et al. 2000). Fire size cannot be estimated from FSD methods because studies are site specific. Spatially based sampling studies are smaller than the fires that account for most regional burn area (e.g., Kilgore and Taylor 1979; Beaty and Taylor 2001). Area estimates from stand structure (e.g., Kilgore and Taylor) cannot differentiate actual burn size and local fire behavior within burns. A tree may survive because it lies beyond the limit of a burn, or because of a shift in intensity within a burn. If fire intensity was proportional to fire size, the combination of long-tailed fire size frequency distributions (Minnich and Chou 1997; Malamud et al. 1998) and low scarring efficiency would yield a record of intense major burns exclusively. Alternatively, with perfect scarring, fire scars would reflect both mass burns and local microburns that collectively add to small fire area; that is, any randomly selected FSD sampling site may record the full range of fire sizes and intensities. In such case, regional burning rates may not covary with scarring rates. Large transient fluctuations may arise from the interannual variation in lightning discharge rates, anthropogenic starts, and fuel moisture, rather than from the regional rate of burning at the landscape scale.

SIERRA SAN PEDRO MARTIR: GRAZING, CLIMATE CHANGE, OR MICROBURNS?

The SSPM is a rare example where fire-interval estimates of FSD studies can be empirically tested against a spatially explicit fire history from aerial photographs. Estimates of fire intervals from perimeters (Minnich et al. 2000) are twice that estimated from FSD methods (Stephens, Skinner, and Gill 2003). Minnich et al. found that large fires are intense and self-organizing. Forests outside perimeters show fixed spatial arrangement of trees and shrubs over scales of decades; that is, no fires removed woody biomass. Microburns are abundant, but leave only ash beds of consumed needle litter, with little effect on forests. One important unanswered question is whether microburns leave scars on tree bole catfaces. The interpretation of site-specific FRls to the landscape assumes covariance between scarring and landscape burning rates. However, scarring may also reflect transient variation in local ignition rates. For example, to explain the decline in scarring rates after 1790, Stephens, Skinner, and Gill (2003) proposes that the introduction of cattle reduced dry herbaceous cover, thereby increasing fire intervals (a landscape interpretation). This view requires that herbaceous cover

contributed significant fuel. However, annual and perennial forbs form < 10% cover and remain green in summer. Exclosures show no significant effect of cattle grazing on herbaceous cover (Minnich et al. 1997). Alternatively, the decrease in scarring may have been caused by the decimation and dislocation of Kiliwa Indians with the establishment Mission San Pedro Martir in 1794, thereby eliminating a source of abandoned campfires (a site-based interpretation). Stephens, Skinner, and Gill assert that there is no historical record for fire use in the SSPM, but indigenous cultures there are poorly studied. Native Americans doubtless set fires for cooking and warmth, and for ceremonial rituals (Kroeber 1925). For example, the extraordinary 2- to 4-year fire intervals in Sierra Nevada Sequoiadendron groves (Swetnam 1993) is very likely an outcome of Native American cooking fires. In summary, FSD methods provide little direct evidence of fire regime properties other than FRls. Hence, the deduction of fire regime properties (intensity, selective removal of biomass, population dynamics) is dependent on the accuracy of FRI estimates at broad spatial scales. The grass extirpation hypothesis is circular because an assumption (former herbaceous cover) is used to affirm a conclusion. No paleobotanical studies have been undertaken in the area. Similarly, in the southwestern United States the assertion that variable scarring rates correlates with interannual precipitation variability due to the El Niflo cycle (Swetnam and Betancourt 1990) could also be reasonably explained by variable establishment of microburns due to changing fuel moisture of the litter layer rather than changes in landscape scale burning. The hypothesis that fires decreased because of reduced El Nino frequency and herbaceous cover between 1790 and 1840 (e.g., Stephens, Skinner, and Gill 2003) could also be explained by the decimation of Native Americans (Cook 1937, 1940, 1943). Given poor spatial sampling protocols to present, the effects of microburns from drought, and Native American burning practices cannot be differentiated. The inconsistencies in FSD chronologies and spatially explicit reconstructions of fire regimes from time-series aerial photography leads to a number of important questions: What is the long-term ecological role of small fires across landscapes, given that lightning strike rates are 1-3 strikes ha -1 yr- 1 and any hectare of forest contains several stems with bark stripped off by lightning? As first discussed in Minnich et al. (2000), do microburns leave scars in cat faces, thereby influencing fire-scar dendrochronologies, or is scarring exclusively an artifact of landscape-scale fires? At present, there is a lack of synoptic studies that incorporate plot data and spatial fire records. Indeed, to date there is not one FSD study where fire scars are correlated with independent spatially explicit evidence. Perhaps most significant is whether FRls underestimate spatial fire intervals because the method cannot differentiate mass burns from a cloud of microburns with little collective impact on forests (Minnich et al. 2000; Baker and Ehre 2001). Even at scales of a few hectares, fire scars may be created both by mass burns and microburns within a

SOUTHERN CALIFORNIA CONIFER FORESTS

525

TABLE 18.2

Major Forest Pests in Southern California

Coulter Pine

Ponderosa Pine

Western pine beetle (Dendroctonlls breviconnus)

X

X

Mountain pine beetle (Dendroctonus ponderosae)

X

(X)

Tree Species/Pest Species

Jeffrey pine beetle (Dendroctonus jeffreyi)

/effrey Pine

Sugar Pine

White Fir

Pinyon Pine

X

(X)

x

X

Red turpentine beetle (Dendroctonus valens)

x

x

x

x

Pine engraver beetles (Ips spp.)

x

x

x

x

x

X

X

x

Pinyon pine engraver (Ips confusus) California flatheaded borer (Melanophyla californica) Fir engraver (SCOlytllS ventralis)

x

Flatheaded fir borer (Melanophyla drumrnondi)

x

x

Dwarf mistletoe (Arcellthobizan spp.)

X

X

X X

X

Leafy mistletoe (Phoradendron pallciflorllm) Annosus root disease (Heterobisidion annosllln)

X

x x

Blackstain root disease (Leptographizan wagneri)

x

x

x

x X

NOTE: X high ability to kill vigorously growing trees, x = less mortality, (X or x) only an occasional host. Note that ]effrey pine bark beetle does not occur in the San ]acinto Mountains nor in the peninsular range.

single mass burn fire cycle. Other fire regime parameters deduced from FSD studies are not constrained. Studies that correlate fire occurrence with climatic variability in annual time scales cannot be applied to ecosystems where fires occur in response to cumulative fuel buildup over long time scales. Time-series climate correlations with fire should be scaled as a running average of the process tied to the cumulative build-up of vegetation architecture and fuels, rather than annual time-series statistics (Lovell, Mandondo, and Moriarty. 2002). Fire lnodels based exclusively on fireinterval estimates, a number, nothing else has empirical basis. Any error in fire-interval estimates affects all other fire-regime properties reconstructed by deduction. The entire model can fall like a house of cards. The bottom line is that the empirical basis of FSD methods is poor at best. The human mind is capable of extraordinary concepts, but needs a "reality check" from nature, the best teacher, which will always provide the unexpected.

Disturbance from Insects, Pathogens, and Air Pollutants Native insect herbivores and pathogens, including bark beetles, mistletoes, and root diseases, perform an important

526

SOUTHERN CALIFORNIA CONIFER FORESTS

function in conifer forests including killing decadent trees and nutrient cycling (Pronos et a1. 1999). Pine beetles are either host specific or generalists (Table 18.2). Among those that kill trees in southern California, Dendroctonus jeffreyi breeds almost exclusively in Jeffrey pine, Dendroctonus pselldotsllgae is exclusive to Douglas firs, Dendroctonus brevicomis attacks ponderosa and Coulter pine, and Scolytus ventralis attacks white fir. Dendroctonus ponderosae breeds in many pine species (Wood et a1. 2003). Beetles and engravers are opportunists that attack trees in a weakened or dying state due to root erosion, snow breakage, advanced age, lightning strikes, drought, and other factors that compromise tree vigor. With only a few rare exceptions, either the host tree is killed by the colonizing bark beetles or the host resistance of the tree kills the attacking adults. To kill a tree, large numbers must successfully colonize in a relatively short period of time (Paine, Stephen, and Taha 1984; Paine, Raffa, and Harrington 1997). Adults emerge from the larval host tree and search for susceptible hosts. Healthy pines and firs respond by exuding pitch, which either "pitches out" the adults or blocks its progress. In weak trees with reduced resin pressure, the adults are able to initiate colonization and produce aggregation pheromones that attract other colonizing adults. Pheromone production ceases when the host tree

ceases resin flow (Raffa and Berryman 1983) signaling the death of the tree. Eggs are laid in the inner bark and the larvae excavate galleries in directions generally perpendicular to the adult galleries. Pupation occurs either in the inner bark or in the outer bark, depending on the species of beetle. Western pine beetle can produce up to 4 generations in a year in southern California due to the mild thermoclimate permitting the populations to expand rapidly when there is an abundance of susceptible host material available for colonization. There is a threshold of attacking beetles required to kill trees, and that threshold is a function of the vigor of the host; a slnaller number of beetles is necessary to kill stressed trees than vigorously growing trees (Paine, Stephen, and Taha 1984). Under normal conditions when background populations of beetles are low, tree mortality is usually at low levels when stands have a few weak trees. However, under drought conditions when there are large numbers of highly stressed trees, a relatively small population of beetles can kill many trees. As a result, the population of beetles will increase in those drought stressed trees, and many more trees will be at high risk because more vigorous trees can be killed by the larger number of adults that respond to the aggregation pheromones (Paine, Stephen, and Taha). Root diseases and dwarf mistletoes account for most of pathogen-caused growth loss and tree mortality in the western United States (Pronos, Merrill, and Dahlsten 1999). Dwarf mistletoes (ArceuthobiLl1n spp.) place more stress on host trees than true Inistletoes. Lacking chlorophyll, dwarf mistletoes reduce the anlount of water and photosynthates available for tree survival and growth. True mistletoes take water from their host and produce their own photosynthates. Western dwarf mistletoe (ArcellthobiLul1 call1pylopodlun) attacks Coulter, Jeffrey, and ponderosa pine. A true mistletoe (Phoradendron paucitlorllnz) attacks white fir. Annosus root disease (Heterobisidion anlloslun), black-stain root disease (Leptographiurn wagneri), and armillaria root diseases (Arrnillaria spp.) are somewhat host specific and do not generally kill trees directly. They do predispose trees to subsequent attack by bark beetles. In annosus root disease, microsopic airborne spores land on recently cut sturnp surfaces or bole wounds, grow into root system, reducing water absorption (annosus root rot). Annosus may persist and spread among root systems for years. Mistletoes also tend to weaken trees reducing survival due to beetle infestations. Black stain root disease fungus, which attacks primarily single-leaf pinyon pine, colonizes the water-conducting vascular tissues and interferes with water movement, but does not destroy living tissues as with annosus. Sugar pine forests in northern and central California are extremely susceptible to the introduced white pine blister rust (Cronartiurn ribicola), but the pathogen has yet to spread into southern California (Kliejunas 1985). Bishop pine has recently sustained heavy mortality from needle blight caused by [)othistroma septospora (Ades et a1. 1992). The exposure of coniferous trees to oxidant air pollution in SCA results in lower photosynthetic rates and production of carbohydrates, as well as changes in plant priorities in

resource acquisition, allocation, and partitioning (Grulke 2003). Nitrogen deposition also modifies the effects the oxidant exposure. Although the physiological response of trees to oxidant pollution is well documented, the long-term impact of air pollution is complex in relation to other factors such as stand-densification, drought, and insect attack. Summaries of air pollution impacts in forests of SCA and the Sierra Nevada are given in McBride and Miller (1999) and Bytnerowicz et a1. (2003). The amount of air pollution transported to SCA forests depends on the distribution of pollution sources-primarily oxidant and nitrate from automobiles and power-generation plants as well as ammonia and nitrous oxides from agricultural areas-and transport due to onshore flows and anabatic winds during the summer pollution season (Edinger et a1. 1972; Lu and Turco 1995; Padgett et a1. 1999). The plume of highest oxidant and nitrate concentration moves from Los Angeles to SGM and SBM. The OrangeRiverside County plume moves to the eastern SBM and northern SJM, whereas the San Diego County plume moves into the Cuyamaca and Laguna Mountains and southward into far northern SJZ of BCA. Lacking upwind sources, low pollution concentrations occur in the lTIountains of Ventura and Santa Barbara Counties, the southern SJM, and Palomar and Hot Springs Mountains. The air is still pristine south of ca. lat. 32° in BCA. Air pollution is stably stratified in the marine layer during transport across the SCA coastal plains with onshore flows, but anabatic winds move pollutants upslope through the marine inversion along coastal front of the mountains. Pollutants in anabatic flows also stratify into the overlying inversion layer where they have longer residence times than in the marine layer (Edinger et a1. 1972). The highest forest pollution levels have been recorded in the western SBM where trees grow as low as 1,300 m and are exposed day and night to oxidant transported by anabatic winds and flows in the inversion layer (Watson et a1. 1999). The marine inversion normally dissipates on the desert side of the mountains, with oxidant and nitrates mixing and diluting into upper air layers. Hence, pollution levels normally decrease from windward to leeward slopes, although ozone transports more readily than nitrogen (Bytnerowicz et a1. 1999). Because oxidant exposure is also dependent on wind speed (advection), trees in exposed ridges and summits sustain greater flux of ambient pollution than in sheltered canyons and basins. The physiological impact of ozone is the reduction of net photosynthesis, stomatal conductance, production of carbohydrates, and nutritional content of tissues (McBride and Miller 1999; Grulke 2003). Foliage develops chlorotic mottle on older needles of each whorl, which accelerates necrosis leaf senescence. Crown injury includes the reduction of needle whorls, needle length, and percentage of live crown. Reduced photosynthetic capacity limits the carbohydrates available for the growth of the tree. Ozone injury assessments show that ponderosa and Jeffrey pine are most sensitive;

SOUTHERN CALIFORNIA CONIFER FORESTS

527

bigcone Douglas fir and Coulter pine are moderately sensitive, whereas lodgepole pine and sugar pine are moderately tolerant. White fir and incense cedar are visually the least affected (Miller et al. 1983). A model simulation of white fir response to elevated ozone and drought stress reduced branch and bole growth similar to field responses found in the S]M (moderate pollution) and SSPM (Retzlaff et al. 2000). Time-series data show that ring width and cross-sectional growth are reduced where ponderosa pine, ] effrey pine, and bigcone Douglas fir are exposed to chronic levels of oxidant air pollution (Arbaugh, Peterson, and Miller 1999). Oxidant exposure alters within plant priorities for resources as less carbon is allocated to roots and less foliar biomass is retained (Grulke 2003). Compared to herbs, conifers have lower stomatal conductance, and lower 0 3 uptake. However, the total 0 3 exposure and uptake is larger over the lifetime of a leaf because foliage is longer lived and active over a longer period of the year than herbs. Cumulative 0 3 exposure, high radiation, and drying of the upper soil horizons (insufficient uptake of N to replenish damaged pigments) may all contribute to chlorotic mottle. The period of the highest gas exchange rates and presumably the greatest air pollutant uptake takes place during the growth flush in early summer (Grulke et al. 2002). Stomatal conductance decreases with increasing pollution levels, with normal stomatal function being lost at the most polluted sites. Environmental factors that decrease stomatal conductance also decrease ozone uptake. The seasonal course of stomatal conductance, as regulated by water availability, is a key process controlling injury development each summer. Ozone injury is low in early summer, when fine root mass is wet. Chlorotic mottle increases by midsummer when drying trees rely on deep roots for water (Hubbert et al. 2001a). If significant rains occur in autumn, fine roots and mycorrhizae grow in the near surface soils, and chlorotic mottle can significantly decrease. Polesize and large trees can mitigate reductions in carbon acquisition with carbon assimilation in unpolluted days in winter. Oxidant pollution levels peaked in the 1970s and have decreased since that time due to air pollution control measures. The crown condition of pines improved slightly in the 1980s but no significant change has occurred in the 1990s (Miller and Rechel 1999). The effect of bark beetles from pollution injury is smaller than nonpolluted trees due to reduced cambial thickness; that is, reduced growth also reduces food reserves for bark beetles. However, it takes smaller number of beetles to kill weakened trees (Pronos et al. 2001). Ambient nitrogen deposition, mostly dryfall (Byrtneowicz et al. 1999), decreases carbon allocation to roots (Grulke, Andersen, and Hogsett 2001), further exacerbating the effects of oxidant exposure to roots (Grulke et al. 1998). Ambient nitrogen deposition accelerates needle loss by increasing foliar nitrogen content, producing redundant nitrogen with too high carbon coast for maintenance (Grulke 2003). Drought stress reduces both the number of

528

SOUTHERN CALIFORNIA CONIFER FORESTS

needle age classes and needles within each age class, which acts to synergistically with 0 3 exposure to promote premature senescence of foliage. This can occur as excess nitrogen saturation (Fenn, Poth, and ]ohnson 1996). Increased nitrogen availability also exacerbates 0 3 exposure effects on foliage turnover: fewer needle age classes are retained (Gower et al. 1993). Nitrogen deposition counteracts the effect of oxidant exposure on photosynthesis by increasing nitrogen available for photosynthetic pigments and enzymes, increasing stomatal conductance, but deleteriously resulting in increased 0 3 uptake (Grulke 2003). Nitrogen deposition may also mitigate the degree of foliar injury from oxidant pollution by increasing nitrogen available for reparation of photosynthetic pigments. Ozone induces premature foliar senescence and abscission, and N stimulates the production of foliar biomass, the net result of which is greater litterfall (Arbaugh, Peterson, and Miller 1999; Fenn et al. 2003). Carbon allocation to fine roots is significantly lower with increased pollution (Fenn and Poth 1999). Thus, the combination of drought stress inhibition of aboveground litter decomposer microbes, sustained litter production in trees capable of accessing deep water, greater C allocation above ground, and N stimulation of litter production (reduced decomposition rates) are factors that likely contribute to long-term C storage in above-ground detritus, as well as increasing fuel buildup in stands already at fire risk due to long-term fire suppression (Grulke et al. 1998, 2001; Arbaugh 1999; Grulke and Balduman 1999; Fenn et al. 2003). The greater C storage in above ground woody tissue (Grulke et al. 1998,2001) and reduction of fine root biomass and carbohydrate allocation belowground increase the risk of conifers to drought stress (Grulke et al. 1998, 2001). Over long time scales, ambient pollution may favor stand composition toward ozone-tolerant and fire-sensitive species like white fir and incense cedar (McBride and Miller 1999). Selection pressures from stand densification will also increase the abundance of these shade-tolerant conifers. In areas of high ambient pollution, multiyear drought can reduce 0 3 uptake, but can also reduce carbon and nutrient acquisition, reducing resource allocation to defenses (antioxidants, resins against insect infestation).

Catastrophic Dieback in Record Drought Drought reduces the photosynthetic capacity of trees and the levels of carbohydrates used for growth and tissue repair. Although the stress incurred by the dry conditions may kill trees directly, the reduction in resin pressure also predisposes these trees to attack by bark beetles and pathogens, and increases fire hazard (McBride and Miller 1999). Treering analysis for last 1,000 year shows that every decade has several years of pronounced reduction in tree growth due apparently to drought (Graumlich 1993). The correlation between beetle attacks and climate can be diffuse because bark beetles may delay or prolong the exact time of tree mortality. However, mortality tends to increase in multiyear

70 60 50 40 30 20 10

°

1850/51

1860/61

1870/71

1880/81

1890/91

1900/01

1910/11

1920/21

1930/31

1940/41

1950/51

1960/61

1970/71

1980/81

1990/91

2000/01

FI G U R E 18.10 Annual precipitation (in centimeters) for San Diego during 1850-2002. Notice that then driest year of record (6 cm,) was

2001-2002. Data are from the National Weather Service.

FIGURE 18.11 Whole-stand mortality of bigcone Douglas fir forest at Skinner Creek. (a) 1980; (b) 2003.

droughts (Taylor 1973), particularly in highly resource competitive dense stands, or stands with preexisting damage or stresses (Pronos et al. 1999). Elevated mortality occurred in SCA conifer forests in 1975 to 1977 and 1988 to 1991 (Savage 1994, 1997). Most recently, the 2002 to 2004 conifer dieback is unprecedented. The winter of 2001 to 2002 was the driest year in SCA since instrumental records began in 1849 (Fig. 18.10). Total precipitation varied from 10 to 25 cm, and 17% to 30% of normal. From SBM to SSPM this record year followed 3 years of subnormal precipitation. By the summer of 2002, conifers and broadleaf trees exhibited failure in leader growth, premature leaf shed, and aborted cones and fruits. Evergreen canyon live oak and chaparral taxa such as Quercus berberidifolia and Cercocarpus betuloides became virtually deciduous by late summer. Leaf shed and crown dieback was also widespread in Arctostaphylos spp., Ceanothus spp., and Adenostoma fasciculatum; and widespread light mortality occurred among species in Ceanothus subg. Cerastes. Historically unprecedented conifer mortality took place in 2002 to 2003. Time-series photographs have been taken at ca. 100 localities since the onset of the mortality outbreak in 2002 (Fig. 18.11). Heavy mortality first appeared in Coulter pine by May 2002, and then expanded to incense cedar and Jeffrey pine by early July. Light mortality began in all other mixed-conifer forest species by late summer. Major insect

attack of ponderosa pine developed on the north side of Lake Arrowhead and locally at Idyllwild. Defoliation of bigcone Douglas fir was first observed in September 2002. In spite of wet ground from normal precipitation in the winter of 2002 to 2003, mortality continued in Jeffrey and Coulter pine until the early summer of 2003, then declined only after trees began uptaking soil water in warm temperatures. Although Coulter pine mortality in 2002 was zonally distributed, mortality in this species the folloWing summer was patchier in association with expanding populations of bark beetles. Mortality virtually ceased in incense cedar. Widespread mortality from bark-beetle attack developed in ponderosa pine by July 2003, especially dense forests at Lake Arrowhead, Idyllwild, and Cuyamaca (see Fig. 18.6 and Fig. 18.11). Drought-compromised trees were apparently unable to rehydrate before terminal attack by bark-beetles. A low frequency of defoliated bigcone Douglas fir stems began resprouting from epicormic buds along the bole and largest branches. In the eastern SBM, populations as large as 100 ha sustained whole-stand mortality. In 2004, mortality had decreased in nearly all species, after 65% of normal precipitation but several soaking rains the previous winter. Surviving trees had improved growth flushes and resistance to insect predators compared to 2003 due apparently to resaturation of soils in combination with tree rehydration. There were also fewer surviving trees competing

SOUTHERN CALIFORNIA CONIFER FORESTS

529

for water. The decline in mortality shows that precipitation departures inclusively are a poor predictor of conifer health statis. Subnormal precipitation years resaturate soils because the mean annual precipitation exceeds soil field capacities, with ca. 30% of annual precipitation contributing to runoff (Franco et al. 2002). In 2004, ponderosa pine bark-beetle infestations expanded to more open stands beyond the urban forests of Lake Arrowhead. There was little mortality at Idyllwild. The highest levels of white fir mortality occurred in the summer of 2003 and continued into the winter of 2003-2004. At broad scales, forest mortality is manifested in predictable spatial pattern in complex terrain. Dead trees tend to concentrate on well-drained convex surfaces and southfacing exposures. Rates of survival were greatest in deep canyons, north-facing slopes, old erosion surfaces and valley floors subject to soil water convergent flow. Mortality was highest in short-lived closed-cone conifer forests, especially Coulter pine, and lowest in long-lived pinyon-juniper woodland and subalpine forest. This outcome should be expected because long-lived species have a greater probability of experiencing drought as extreme as 2002 than short-lived species. In addition, the longevity of trees translates into greater environmental space from which to obtain resources; that is, stems integrated into the habitat over long time scales have a large proportion of resources are carried over long time scales and buffer them from environmental variability (Grulke 1999, 2003). At the short-lived end of the continuum, Coulter pine sustained the highest mortality among coniferous species in the drought of 1988 to 1991 (as high as 50 ha-I) in SJM, and lower rates among mixedconifer species (Savage 1994). In 2002-2003, Coulter pine sustained whole-stand mortality in the order of hundreds of ha, and the order of 1,000 ha in the San Bernardino and Hot Springs Mountains. Apparently, selection for rapid growth results in the production of soft wood that is vulnerable to insect attack and disease. At the long-lived end of the continuum, subalpine forests sustained virtually no impact from the 2002-2003 drought. Soils were replenished by snowmelt as late as May and cool summer temperatures may have limited drought stress. In pinyon-juniper woodland, light mortality is associated with low AAP and absolute water deficits. In all forest types, mortality increased with time-since-fire due apparently to increasing leaf area, transpiration demand, and tree stress in successions. The majority of Coulter pine and bigcone Douglas fir perished in watersheds last burned >60 years ago but few trees perished in burns ----

I

I

I

=--'

Upper Mojavean blackbush, Joshua tree, and Juniper scrub (Achnatherum speciosum, Juniperus californica Coleog)me ramossisima, Yucca brevifolia)

~

'=J--

g

Mid-elevation upper bajadas and alluvial deposits (Pleuraphis rigida, Ericameria cooperi, Acamptopappus spherocephalus)

Upper elevation woodland and scrub (Pinus monophlla, Quercus Cornelius mulleri, Ericameria linearifolia, Arctostaphylos glauca)

>----

~

I

Low elevation (Sonoran) rocky uplands (Hyptis, Bebbia, Pleurocoronis)

Upper elevation washes (Prunus fasciculata, Acacia greggii, Artemisia dracunculus)

--

l r

~

Sonoran arborescent washes (Cercidium, Chi/opsis, Brandegea)

"---

l

I

Low elevation active washes (Psorothamnus spinosus, Hymenoclea, Encelia fructescens)

Mesquite hummocks (Prosopis glandulosa) Pennanent Seeps and Springs Populus, Salix, Baccharis

I

r~=~ ~~ r_r

Sandy (Brassica tournfourtii, Psorothamnus emoryi)

~

r

i i>-----I

l

i ;

i

I~

I f----

~

- r

I

~

-

r

FIG U RE 22.4 Sorensen's Flexible beta cluster analysis diagram of enVironmentally stratified vegetation samples from 300 plots collected for Joshua Tree National Park mapping and classification project (Keeler-Wolf, San, and Hickson 2005). Fifteen main cluster groupings identified in each with representative significant indicators (Dufrene and Legendre 1997) displayed for each group. Solid lines on left indicate upper level group divisions.

assumption most of his sites were below their theoretical carrying capacity, especially those along the lower bajada. Randall believed that the reduced cover reflected a catastrophic loss of cover in the recent past (perhaps drought). Cluster analysis of sample plots stratified to capture a wide array of environmental gradients in the MDEI and a separate Mojave Desert data set from Joshua Tree National Park (Keeler-Wolf, San, and Hickson 2005) indicated similar major groupings of samples based on significant (p 20% in 26 Releves Sampled on 9 Different Dune Systems in the Mojave Desert

Dune Species

FIGU RE 22.9 High cover of psammophytic annuals following a

good rainfall year, including Abronia villosa, Oenothera deltoides, Plagiobothrys sp., and the non native annual grass Schismus barbatus with scattered shrubs of Atriplex canescens and Prosopis glandulosa (in distance). Cronese Lakes Dunes, San Bernardino Co., April 1998.

americanus (American bulrush) often occupy the center immersed in standing water for at least a portion of the growing season. Rings of vegetation dominated or co-dominated by Juncus cooperi, Anemopsis californica, Sporobolis airoides, Distichlis spicata, [va acerosa, and so forth surround these central wet spots. The intensity of quantitative sampling has been relatively low in these situations so far, and the detailed relationships among the stands have not been well worked out (Fig. 22.8). The drier, more seasonally saturated wetlands in this moisture gradient outward from permanent moisture correspond to the alkaline meadows of Thorne (1976), Holland (1986), and other authors. Here, grasses such as SporoboIus airoldes and Distichlis spicata may dominate larger areas, particularly notable north of Furnace Creek in Death Valley and near Tecopa in the Amagosa River valley. PSAMMOPHYTlC VEGETATION OF DUNES AND SAND SHEETS

The vegetation of Mojavean sand dunes, sand sheets, and flats has been of interest for some years. Rowlands et al. (1982) discuss vegetation of eight dune systems of the

626

MOJAVE DESERT SCRUB VEGETATION

Frequency

Dicoria canescens

78%

Astragalus lentiginosus

56%

Coldenia plicata

56%

Pentalonyx thurberi

56%

Atriplex polycarpa

44%

Cammisonia claviformis

44%

Cilia latifolia latifolia

44%

Salsola tragus

44%

Abronia villosa

33%

Cryptantha angustifolia

33%

Larrea tridentata

33%

Palifoxia linearis

33%

Argemone corymbosa

22%

Atriplex confertifolia

22%

Cleome sparsifolia

22%

Eriogonum sp.

22%

Lupinus shockleyi

22%

Oenothera deltoides

22%

Oryzopsis hymenoides

22%

Prosopis glandulosa

22%

Suaeda moquinii

22%

Tidestromia oligosperma

22%

NOTE:

Unpublished data from Pavlik collected in 1980.

Mojave. Pavlik (1985) discusses the phytogeography of desert dunes in California and adjacent Nevada and Oregon. During this study Pavlik also collected releve data (averaging three per dune system) from nine dunes within the Mojave Desert (Pavlik personal communication). Average total vegetation cover values are low (particularly on the deeper active dunes), between 1% and 5%. However, in favorable rainfall years, annual vegetation cover on stabilized dunes and sand sheets may be over 30% (Fig. 22.9). Table 22.4 summarizes the frequency of occurrence of species in the MOjave dune systems sampled by Pavlik. The variance in species composition is substantial in the samples, with only one species, the summer annual Dicorea canescens, attaining frequencies of higher than 75%. Only three additional species occurred on more than half of the dune samples (although several others attained frequencies of >50% based on full species lists for each dune system proVided by Pavlik, personal communication). Many

dunes in the Mojave contain unique or near unique stands of sparse vegetation. The Eureka Dunes are well known for their stands of the endemic grass Swallenia alexandrae and the endemic evening primrose Oenothera caJifornica ssp. eurekensis, whereas other dune systems such as the Kelso Dunes appear to have stands of vegetation at least unique to Mojave dune systems including a sparse stand of hundreds of acres of the shrub, Penstemon thurberi. In the MDEI study, several plots were taken in dune areas. Additional alliances have been identified for the dune systems including: Pleuraphis rigida, Larrea-Ambrosia, Abronia villosa, and (Achnatherum) Oryzopsis hymenoides, Sarcobatus

vermiClllatlls, Atriplex canescens, Atriplex lentiformis, Prosopis glandulosa, and Pluchea sericea. All of these alliances are known from additional stands outside of the Mojave in either the Sonoran or the Great Basin desert. Most of the latter alliances are also found off of dune systems on other well-drained or water-bearing substrates and are not typical of the deeper shifting sands of the larger systems. The morphological traits most characteristic of plants of the deeper shifting sands include annuallifeforms (e.g., Abronia villosa, Dicorea canescens, or Palafoxia arida) and stoloniferous grasses (Panicum urvillianum or Swallenia alexandrae).

Vegetation of Bajadas, Hills, and Washes

This zone forms the core of the Mojave Desert. Cumulatively, vegetation within this category covers over 70% of the mapped portion of the central Mojave (Thomas, KeelerWolf, and Franklin 2002). Vegetation in this zone ranges from simple sparsely vegetated stands in extremely rocky, dry, and hot lower bajadas and hills to diverse stands with over 30 species of perennials on the cooler and better watered upper bajadas, at the interface with the upper bajadas and slope vegetation (see next section). The largest category by far in this group is that characterized by Larrea tridentata and Ambrosia dumosa. It includes most of the vegetation of Bajadas, Hills, and Washes category in Table 22.2. However, other types also exist including Atriplex hymenelytra (desert holly) scrub, and a number of desert wash communities dominated by such species as Ambrosia salsola (cheesebush), Acacia greggii (catclaw acacia), Ericameria paniculata (black-band rabbitbush), Ephedra caJifornica (California ephedra), and Psorothamnus spinosus (smoketree). Some of the excessively dry and hot lower bajadas and slopes with desert varnish appear unvegetated, but in good rainfall years may be covered with ephemeral annuals.

THE LARREA TRIDENTATA-AMBROSIA DUMOSA (CREOSOTEBUSH-BURROBUSH) ALLIANCE

Stretching from the Antelope Valley in the west to the Eureka Valley in the north and ranging south into the northwestern portion of the Sonoran Desert, this single alliance strongly dominates the main Mojave Desert. It covers vast areas of sandy alluvial fans and bajadas, as well as rocky

FI G URE 22.10 Spring aspect of Larrea-Ambrosia dllmosa alliance stand on a dissected alluvial fan southeast of Kelso Peak, San Bernardino Co. April 2005. Larrea shrubs average 1.8 m tall, Ambrosia dllmosa are shorter shrubs and are about 0.5 m. The annuals are mostly Malacotl1rix glabrata and Cl1aenactis fremontii.

uplands and slopes (Fig. 22.10). Although L. tridentata is perhaps the single most characteristic shrub of all North American deserts (Bender 1982; MacMahon 2000), L. tridentata in combination with Ambrosia dumosa are characteristic of the Mojave and the adjacent northwestern portions of the Sonoran Desert (Shreve 1942; Turner, Bowers, and Burgess 1995). Because both of these shrubs are widespread in the adjacent Colorado Desert, discussion of their ecological characteristics will be taken up in Chapter 23. Despite ecological stratification of samples throughout different environments, in the MDEI study nearly 33% of all the samples analyzed were placed within the Larrea zone where Larrea and/or Ambrosia were> 1% cover (Franklin et al. 2001). Due to the broad ecological amplitude of vegetation with Larrea and/or Ambrosia, the environmental correlations across this zone resemble a somewhat narrowed cross-section of the entire Mojave. An ordination of the Larrea zone MDEI samples shows the same general trends as ordination of all plots within the desert. Correlation values are very close to the correlation values within the full data set (see Table 22.1). Of the 23 environmental variables tested 5 showed significance (r values in parentheses), all with the first axis: Max July temp (-0.589), Min January temp (-0.531), Summer ppt (0.494), Elevation (0.607), and Winter ppt. (0.626). Indicator species analysis and cluster analysis run within this same dataset of 709 plots identified several major groupings associated with this data set strongly driven by precipitation and temperature. Following analysis of the retrospective data the samples containing Larrea and/or Ambrosia were combined into a classification using the rules of the national vegetation classification hierearchy (Grossman et al. 1998). Depending on the cover and constancy of shrubs (see Thomas et al. 2004, for descision rules) these included four main alliances and a total of 66 plant associations (Table 22.5).

MOJAVE DESERT SCRUB VEGETATION

627

TABLE 22.5

Alliances and Associations Within the Larrea-Ambrosia Zone Categorized by General Environmental Characteristics

Larrea tridentata Alliance (n = 148) Disturbed types associated with washes, sand, or grazing 6 associations with Ambrosia salsolal Brickelia incanal Pleuraphis rigida l and Lycium andersonii indicative. Low elevation types 4 associations with Eriogonum inflatuml Atriplex polycarpal Atriplex hymenelytral and Acamptopappus shockleyi indicative. High elevation types 4 associations with Yucca schidigeral Ephedra califomical Poa secundal Ephedra nevadensisl and Atriplex confertifolia indicative.

Larrea tridentata-Ambrosia dumosa Alliance (n

=

694)

Upper elevation types 10 associations with Yucca schidigeral Salazaria mexicanal Ephedra nevadensisl Grayia spinosal Opuntia acanthocarpal Eriogonum fasciculatum l Ericameria cooperil Ephedra viridisl Gutierrezia sarothrael and Atriplex confertifolia indicative. Rocky upland types 4 associations with Viguieria parishiil Echinocactus polycephalusl Ephedra funereal and Amphipappus fremontii indicative. Wash types 6 associations with Senna armatal Ambrosia salsolal Petalonyx thurberil Ephedra californical Bebbia juncea, and Encelia virginensis indicative. Bajada types 13 associations including those with Krameria grayil Krameria erectal Psorothamnus fremontii l Psorothamnus arborescensl Opuntia ramosissima l Opuntia basilarisl Lycium andersoniil Lepidium fremontii l Eriogonum inflatuml Encelia farinosa l Cryptogam crust, and Atriplex hymenelytra indicative. Sandy or basin types 4 associations including those with Atriplex polycarpal Dalea mollosimal Pleuraphis rigida l and Atriplex canescens indicative.

Larrea tridentata-Encelia farinosa Shrubland Alliance (n

=

87)

5 associations; hotter, usually rockier, and lower elevation than Larrea-Ambrosia alliance with Bebbia junceal Pleurocoronis pleurosetal and Peucephyllum schottii as associates.

Ambrosia dumosa Dwarf Shrubland Alliance (n = 13) 3 associations; generally sandy or mechanically altered, with Atriplex confertifolial Atriplex hymenelytral and Pleuraphis rigida as associates.

Atriplex hymenelytra Shrubland Alliance (n

=

30)

5 associations; very hot and or harsh sites, with Tidestromia oblongifolial Encelia farinosa l and Ambrosia dumosa as very low density associates. NOTE:

Data from Thomas et al. (2004) combines historical data and MDEI releves collected in 1997-2000.

LARREA-ENCEL/A FAR/NOSA SCRUB (CREOSOTEBUSH-BRITTLEBUSH) SCRU B

This variant of the typical Larrea-Ambrosia scrub is found commonly on southerly facing slopes in the southern portions of the Mojave up into the lower elevations of the Death Valley and Panamint valleys. In this case the droughtdeciduous subshrub Encelia farinosa replaces Ambrosia as the major short shrub. Larrea tends to become relatively reduced cover and density in the overstory, but is still the visual overstory dominant. This type of vegetation continues widely

628

MOJAVE DESERT SCRUB VEGETATION

south into the Colorado Desert and will be discussed more completely in the Colorado Desert section (Chapter 23). ATRIPLEX HYMENELYTRA (DESERT HOLLY) SCRUB

Desert holly is in many ways the most xerophytic shrub in the Mojave Desert. It thrives in the hottest and most arid portions of the Mojave and is the most drought-tolerant Atriplex in North America (Turner, Bowers, and Burgess 1995). It has many physiological and physical adaptations enabling its persistence in such environments. New leaves

grow in December to April and tend to be pale green, but as water content drops in the leaves through the spring, the bladder-like hairs on the surface of the leaves tend to concentrate salts and increase reflectance. By the summer the leaves are shiny white. Reflectance of leaves leaps from 35(N, in the early spring to over 60% by summer (Mooney, Bjorkman, and Troughton 1974). They thus absorb light for photosysthesis best in the cool wet season and switch to high reflectivity during the hot dry season, when little photosysthesis takes place (Turner, Bowers, and Burgess). The high reflectance and the steep leaf angle relative to sun position enables this species to remain evergreen. Because the shrub is evergreen and relatively long-lived it can photosynthesize over a long period of time, thus compensating for its low rates of carbon dioxide uptake (Pearcy et al. 1974). Desert holly vegetation is usually very simple with A. hymenelytra being the only conspicuous shrub (Fig. 22.11). Although most characteristic of hot rocky slopes it may also occur at the edges of alkaline sinks and on fine sediments of Pleistocene lakebeds. It also may occur in washes and road cuts suggesting its colonizing abilities within the hot dry zone it usually enhabits. Although most characteristic of the lowest and hottest parts of the desert, it also forms stands on dark, coarse basaltic slopes in the Saline Range in the northern Mojave at elevations up to over 1,000 m. Among the most common associates of A. hymenelytra scrub are the annual or biennial Tidestromia oblongi(oJia and the desert trumpet buckwheat Eriogonum inflatllm. Both of these species are tolerant of extremely hot dry summers and appear to colonize disturbed areas in these settings. Tidestromia tends to occur in sandy areas and may even form monospecific stands around the skirts of dune such as the Dumont Dunes with high off highway vehicle use. Eriogonum inflatum conversely tends to occur on rocky bajadas and slopes, but frequently colonizes rocky washes and road cuts. MOJAVE WASH VEGETATION

As with all vegetation in the Mojave, the vegetation of its washes also varies primarily as a result of its relationship to temperature and moisture. Lower elevation washes of the southern Mojave have much in common with washes of the adjacent Colorado Desert. For example, although lacking the taller microphyll trees such as Olneya tesota (ironwood) and Cercidium floridum (blue paloverde) they contain Psorothamnus spinosus (smoketree) and Hyptis emoryi (desert lavender). The upper elevation washes contain an array of species that merge with the adjacent sagebrush, pinyon, and juniper vegetation of the Great Basin Desert such as Quercus chrysolepis, Quercus turbinella, Prunus (asciculata, Ambrosia eriocentra, and Baccharis sergilloides. As a result of the wide ranging sampling of vegetation in the MDEI study (Thomas, Keeler-Wolf, and Thorne 2002)

FIGURE 22.11 Late-summer aspect of sparse Atriplex l1yl1lenelytm scrub on low gravelly hills near Tecopa, Inyo County. Note white color of leaves. Photo taken September 1998.

and the specific work of Evens (2000, 2003) we now know much more about the vegetation of intermittent and ephemeral drainages in the California Mojave. Evens' work summarized vegetation from and different drainage systems in the California Eastern Mojave (Evens). Although some of her plots include truly riparian and Great Basin vegetation, the majority are applicable to this chapter. She compared vegetation along a topographic and elevation gradient from calcareous and noncalcareous mountain massifs. Watercourses were followed from their origins in rocky uplands to where they debauched from lower bajadas to wide playas or basins. Over three hundred 1,000-m 2 reieves were laid out systematically within the watercourse channels along lower (bajada), middle (arroyo), and upper (canyon) reaches of these watercourses. Evens found significant differences in the wide variety of vegetation between canyons, arroyos, and bajadas. She also found significant differences in the vegetation among these three different physical settings depending on whether they were on calcarous or granitic substrates (Table 22.6). Ordination of Evens's plot data within the main Mojave Desert ecosystem (Table 22.7, arroyo-wash plots) showed strong positive correlations with elevation, geographic

MOJAVE DESERT SCRUB VEGETATION

629

TABLE 22.6

Watercourse Vegetation Types and Habitats

National Vegetation Classification Group and Alliance Narne

Canyon

Arroyo

Wash

Forest

Quercus chrysolepis (canyon live oak)

B

Woodland

Pinlls monophylla (single-leaf pinyon pine)

B

Temporarily Flooded Woodland

Populus fremontii (Fremont cottonwood)

G

Temporarily Flooded Shrubland

Salix exigua (narrowleaf willow)

G

Dwarf-Shrubland

Salvia dorrii (desert sage)

L

L

Shrubland

Qllercus turbinella (desert scrub oak)

G

Prunus fasciculata (desert almond)

B

B

Acacia greggii (catclaw)

B

B

B

Encelia virginensis (Virgin River brittlebush)

L

Arnbrosia salsola (cheesebush)

G

G

B

B

G

G

Intermittently Flooded Shrubland

Baccharis sergiloides (desert mulefat)

G

Chilopsis linearis (desert-willow)

B

Ephedra cali(ornica (California desert tea) Chrysothamnus paniculatus (sticky rabbitbrush)

B

Psorothamnus spinosus (smoketree)

G

NOTE:

From Evens (2003). Substrates: G = granitic, L = calcareous, B = both.

location (UTM northings and eastings), slope steepness, and the number of species per plot as plots rose from the lower bases to the upper bajadas and arroyos. Very strong correlations also existed with increasing cover of cobble and stone coarse fragements along this gradient, whereas a strong negative correlation existed with the relative percentage of fine soil fragments (sands and smaller particles), thus emphasizing the natural sorting of alluvium sizes from coarse at the upper ends to fine at the lower ends of desert washes. Several of the alliances determined in this study were further broken into associations. Evens defined 30 associations within 15 alliances from the study area. The broader and less intensive work by Thomas et al. (2004) and Keeler-Wolf, San, and Hickson (2005) substantiated many of these alliances and associations elsewhere in the Mojave. In addition, several other alliances were found to occur in washes in western Mojave including Lepidospartum squarnatuln, Baccharis elnoryi,

630

MOJAVE DESERT SCRUB VEGETATION

Viguiera reticulata, and Hyptis emoryi. The L. squarnatlun alliance is restricted to western Mojave wash systems with substantial winter and spring flows and is an extension of the same alliance characteristic of such watercourses in southern cismontane California (Barbour and Wirka 1997). Baccharis emoryi alliance is widespread and scattered and prefers more reliable water sources and is thus often found downstream frorn permanent springs and seeps, but in settings with less permanent water. Another Widely scattered vegetation type of washes and seeps throughout the mountains of the Mojave is the desert olive (Forestiera pubescens) scrub. Desert olive forms dense thickets in arroyos in the upper elevations of the Mojave Scrub area often associated with Baccharis sergilloides. Probably the most widely distributed vegetation type of Mojave washes is the Ambrosia salsola (cheesebush) alliance. This single species is the primary coIonizer and

TABLE 22.7

Correlations of Environmental Variables with Canyon and Arroyo/Wash Axes of the Bray-Curtis Ordination Analysis (Pearson and Kendall Correlations are Shown as r-values for All Variables Except Aspect, Shown As a Tau-Value)

Arroyo/Wash Plots

Canyon Plots Axis 1

Axis 2

Axis 1

Axis 2

Variable UTME

0.472

w-(t5;69

0.448

0.113

UTMN

0.744

-0.148

O.,~77

-0.26

Elevation

0.7.1.8

-0.266

0.691

-0.061

Wash Width

0.005

0.436

-0.385

0.285

Slope

0.177

-0.323

0.,~:;71

0.042

---O/~7

0.393

-0.13

0.183

0.001

-0.15

0.804

-0.342

Fines

m'~O.697

0.004

Bedrock

-0.336

v~-O.476

0.073

-0.049

Gravel

0.074

0.644

-0.205

-0.111

Cobble

0.321

0.,.5.92

0.701

-0.227

0.22

-0.027

0.64,5

-0.135

0.138

0.6.52

0.396

0.058

0.409

-~O.47,l

0.022

0.319

Aspect No. Spp.

Stone Boulder Litter

NOli:

~~·O.(j69

0.274

Data froIn Evens (2003). Strong correlations are boIded.

often dominant of wash systems throughout most of the desert from low elevations below 600 m to over 1,400 m elevation. It has occupied parts of the Mojave for over 11,000 years (Cole 1986) and is frost tolerant and well adapted to drought through facultative leaf loss (Comstock, Cooper, and Ehleringer 1988). It can resprout after fire and has fruits that are adapted to flooding dispersal and wind dispersal. Webb, Steiger, and Turner (1987) describe A. sa/so/a as a "stress-tolerent ruderal" with a short lifespan, high seed production, and the ability to rapidly colonize disturbed sites. It may be equally at home in washes or in recently burned upper bajada scrub, or heavily grazed or Inechanically degraded landscapes. Other widespread wash alliances in the central and southern Mojave include those dominated by the tall shrubs or low trees Acacia greggii (catclaw) and Chilopsis lincaris (desert willow, Bignoniaceae) and Chrysothalnnus palliCll/atlls (black band rabbitbrush), characteristic of active sandy to gravelly washes throughout much of the desert. 'rhese three alliances often occupy the same reaches of the same strearTI and appear to segregate based on flooding cycles and intensities of events with the Chilopsis alliance

being most faithful to reliable subterranean water within active flooding zones. ANNUAL VEGETATION OF DESERT PAVEMENTS AND HERBACEOUS VEGETATION OF THE WESTERN MOJAVE

This category is a somewhat artificial group for the annual vegetation of the Mojave. Although cluster analysis (Group 116 in Table 22.2) suggests that the annual ephemeral vegetation of the hot desert pavements of the lower fans and bajadas naturally falls into one general group, the quantitative ecological position of the showy and important annual/herbaceous vegetation of the western Mojave has not yet been analyzed, and if so would probably fall in a different part of desert plot cluster groups. It is treated here as a matter of convenience. Annual vegetation of sand dunes and sand sheets has already been discussed. In addition, annuals are important constituents of several other Mojavean vegetation zones. The environments described herein range from the most xeric upland conditions in the Mojave, so hot and or dry to preclude the development of any shrub cover, to the Larrea-Ambrosia core of the Mojave where many annual species form an ephemeral understory

MOJAVE DESERT SCRUB VEGETATION

631

beneath shrubs, to relatively moist and cool conditions of the predominately winter rainfall regime of the western Mojave. The latter situation resembles the conditions of much of Cismontane central and southern California. Most of these settings lllay be devoid of any low growing ground vegetation for several consecutive years and only have significant cover of annual species following adequate and well-timed precipitation. These"good wildt10wer years" are spectacular events and draw many tourists to the deserts. Rainfall patterns that encompass these conditions range from bimodal winter and summer to exclusively winter precipitation. Annuals, of course, also occur within shrub-dominated communities throughout the desert and strongly change the character of these types for short periods. Patterns of distribution were described by Went (1942) in which certain species tended to occur under shrubs and other species tended to occur in intershrub areas. Regional differences in species association patterns were also explicitly indicated by Went. Some of the possible explanations for the occurrence of annuals under some shrubs but not under others have been explored by Muller (1953), Muller and Muller (1956), and Adams, Strain, and Adams (1970). Rickard and Beatley (1968) noted that sites dOlllinated by Larrea had more annual species and greater annual plant cover than any of eight other shrub-dominated comrnunities in southern Nevada. The MDEI data (Thomas, Keeler-Wolf, and Thorne 2002) suggest that a single alliance represents the ephemeral vegetation of the lower bajadas and hills of the MOjave although complete sampling of other annual, nonshrubby vegetation of badlands, mud hills, and other largely unvegetated surfaces was not possible. This is the Geraea canescensChorizanthe rigida alliance, typical of old desert pavement surfaces throughout the Mojave and adjacent Sonoran deserts. Due to the extreme age and development of armored clay-rich subsoils these desert pavements do not typically support perennial shrubs except at very low density and low stature (McAuliffe 1994, 1999). When winter rainfall approaches a minimum threshold (Venable and Pake 1999), large numbers of annuals of these and other species germinate and flower in the early spring. The tall, showy composite Geraea canescens (known as "desert gold") brightens thousands of acres of otherwise harsh, inhospitable lower bajadas and fans in springs throughout the desert in good wildflower years. Chorizanthe rigida is a short extremely spiny annual (the lllost wildely used common name is simply "spiny herb") in the Polygonaceae, which is most notable if you happen to sit on it unsuspectingly. These two species, along with several annuals in the Borginaceae, Polygonaceae, Asteraceae, Plantaganaceae, and Resediaceae, show up regularly on these desert pavements. The annual herbaceous vegetation of the western MOjave consists of many speCies, most of which have greater similarity to communities of annuals in cismontane central and southern California. Considerable variation in winter annual vegetation is apparent regionally, locally, and seasonally. The composition of the annual vegetation differs

632

MOJAVE DESERT SCRUB VEGETATION

troln year to year, depending on the time and amount of rainfall, as suggested by Shreve (1942), and inferred from the seed germination patterns described by Went (1948, 1949) and Tevis (1958a, 1958b). Beatley (1969b) catalogued differences in density of annual plants over a 5-year period (1969a) and variations in annual plant biomass over 3 years. She also correlated general phenology of annuals with climatic events (Beatley 1967, 1974a). The following paragraph is drawn from her work in southern Nevada. Winter annuals germinate after autumn or winter rains in excess of 15 mm, and the life span extends to late May, when drought and high temperatures cause plant death. The winter-hardy plants (surviving at least down to -18°C) grow slowly until March or April. Then rapidly increasing soil and air temperatures lead to completion of vegetative growth and onset of flowering and fruiting by May. Depending mainly on rainfall, winter annual density may approach 1,000 per m 2 (but more typically less than 100 per m 2), a cover of 30(}() and a biomass of over 600 kg ha~l. Herbivory accounts for some plant loss, but is insignificant compared to mortality from drought stress in the top 25 cm of soil. Most winter annuals do not survive to maturity. Two different annual plant communities have been noted on three substrate types in the Whipple Mts. of eastern San Bernardino Co. (Vasek and Barbour 1977). A large, rocky slope, heterogeneous in surface conformation, supported 18 species with an average density of 55 plants per m 2. A more uniform, more open slope with stony soil had only 9 species, but small plants, such as Plantago ovata, were abundant and total plant density reached 190 per m 2. A similar local difference in annual vegetation between sandy bajada and a rocky slope occurs in the western Mojave Desert (Johnson 1976). Furthermore, the spatial heterogeneity of annual communities seemed greater than that of perennial communities. In addition to the winter annuals discussed earlier, a few species of summer annuals germinate and grow in response to summer rain. Summer annuals grow in the intershrub areas, and all have a C4 carbon pathway in contrast to the winter annuals, which have a C3 pathway. Partial suppression of winter annuals was evident after growth of one species of summer annual (Pectis papposa), but the mechanism has yet to be explained (Johnson 1976). On a sandy bajada, the average density of winter annuals was 39.8 plants per m 2 on a plot irrigated the previous summer and 52.1 plants per m 2 on the adjacent nonirrigated area. In the Nevada Test Site, the summer annual flora consists of only seven species, though in the middle elevations they may occur in large numbers (84 plants per 1ll2, 8(~'h cover: Beatley 1974b). These germinate in August or September after heavy rains, remain small and inconspicuous, and flower and fruit until autumn frosts kill thelll. Unlike the winter annuals, their life span is measured in

terms of weeks. Mulroy and Rundel (1977) compiled a useful summary of 130 common species of winter and summer annuals of the

Mojave and Sonoran Deserts. The summary lists type of leaf anatomy (Kranz or normal), flowering period, and leaf arrangement (rosette, cauline, prostrate, and combinations). Of 63 summer annuals, 42 (67°/3 m) Y. brevifolia is an indicator for the upper bajadas zone, whereas lower stature «3 m) Y. brevifolia is a significant indicator for the upper slope Mojave scrub, to be discussed in the next section. As Webber (1953); Hogan (1977); Vasek and Barbour (1977); Phillips, Page, and Knapp (1980); and others have pointed out, tree-stature Joshua tree "woodland" usually occurs on loose soils and gentle substrates, whereas blackbush scrub and shadscale scrub are often found on heavy or rocky soils. Presumably shorter stature Y. brevifolia occurs as a stronger indicator in the upper slope Mojave Scrub group because it tends to occur on shallower, rocky soils and at cooler winter conditions less condusive to tall growth than the larger individuals indicative of the upper bajada scrub. The quantitative rules developed from the MDEI data analysis indicate that any stands with over 1% evenly dispersed Y. brevifolia with a variety of understory species meets the criteria for falling within several related Y. brevifolia vegetation types (see Table 22.3). Although on rare occasion, tree cover of Y. brevifolia exceeds 10%, it averages only 1.5% in the 75 samples from the MDEl data set. Y. brevifolia is a conspicuous element in over 274,280 ha and is the dominant vegetation in 1,277 polygons of the MDEI vegetation map. It is within this upper bajada zone that Joshua trees reach their greatest densities and heights. One such area, Cima Dome in the eastern Mojave, is summarized by Vasek and Barbour (1977). In a single plot, the estimated density of Joshua trees in a sample area near was between 104 and 125 trees per hectare. Average Joshua tree height was 2.24 m, and average stem diameter 30 cm above the ground was 18.75 cm. On one plot (near Cima) in the MDEI data set, tree-size Y. brevifolia attained a cover value of 19%, by far the highest recorded. Within this upper bajada zone and including the shortstature Joshua trees that occupy the upper slope Mojave scrub, Y. brevifolia occurs associated with several different f10ristic groupings that may be called alliances, suballiances, or plant associations depending on the rules applied. Thomas et al. (2004) have identified several "suballiances" based on the rules of the National Vegetation Classification. These include: (a) Yucca brevifolia/Coleogyne ramosissima at higher elevations on shallow rocky soils; (b) Y. brevifolia sparsely wooded shrubland, which includes a mixture of understory shrubs; and (c) Y. brevifolia/herbaceous type where the grasses Pleu-

634

MOJAVE DESERT SCRUB VEGETATION

FIGURE 22.12 YI/cca brevi(o/ia savanna with the grass P/el/raphis rigida as the principal understory, Wild Horse Valley, Joshua Tree

National Park.

raphis rigida or P. jamesii are the conspicuous understory species (Figure 22.12). Sawyer, Keeler-Wolf, and Evens (ms) describe a Joshua tree/shrubland alliance and a Joshua tree/grass alliance and summarize the associations found within each of these defined from data analyzed in the MDEI project and in Joshua Tree National Park, analyzed by Keeler-Wolf, San, and Hickson (2005). Gradations occur toward creosotebush scrub at low elevations, where Yucca brevifolia may be in codominance with Larrea tridentata. Gradations at higher elevations occur with pinion or juniper wood-land, where Y. brevifolia may occur as a co-dominant with /uniperus californica or /. osteosperma and Pinus monophylla. The described plant associations (Thomas et aI. 2004; Keeler-Wolf, San, and Hickson 2005) are arranged below along an idealized elevational gradient from coolest and wettest to hottest and driest: Yucca brevifolia/Artemisia tridentata-Atriplex confertifolia (n = 4)

Yucca brevifolia-/uniperus californica/Coleogyne ramosissima (n = 3) Yucca brevifolia/Yucca baccata/Pleuraphis jamesii Yucca brevifo1ia/Coleogyne ramosissima

(n

=

7)

(n = 10)

Yucca brevifo1ia/Pleuraphis rigida-Muhlenbergia porteri (n = 14)

Yucca brevifolia/Guiterrezia microcephala/Pleuraphis rigida (n = 10)

Yucca brevifolia/Opuntia acanthocarpa/Pleuraphis jamesii (n = 3)

Yucca brevifolia/Lycium andersonii

(n = 10)

Yucca brevifolia/Salazaria mexicana

(n = 3)

Yucca brevifolia/Opuntia acanthocarpa

(n = 6)

Yucca brevifo1ia/Larrea tridentata-Ephedra nevadensis (n = 31)

probably been historically more frequent. Minnich (2000 personal communication) has observed the response of a Joshua tree (Y. brevifolia var. brevifolia, sensu Munz 1974) stand follOWing the 1999 fire in Covington Flat, in Joshua Tree National Park. Although some resprouting did occur in the year follOWing the fire, there was near complete die-off of those sprouts, many as the result of browsing activity by native herbivores. Increase in the numbers and cover of Achnatherum speciosum, Sphaeralcea ambigua, Guierrezia microcephala, and especially the nonnative grasses Bromus madritensis and B. tectorum, were substantial. Although difficult to age using traditional methods, Comanor and Clark (2000) have shown that Yucca brevifolia is a relatively short-lived and fast-growing tree. Repeat photography of permanent plots in Nevada (Webb et al. 2003) demonstrates the rapid waxing and waning of individuals and stands of this species.

Mojave Yucca (Yucca schidigeraJ Scrub

FIGURE 22.13 Resprouting Yucca brevi(oJia (formerly called Y. b. f. lIerbertii) approximately 5 years following a fire in the Antelope Valley

at the southern base of the Tehachapi Mountains, Kern Co. Plants in the background include the postfire increasers ClIrysotlIamnus nauseosus and AcllIIatl1erum speciosum.

Yucca brevifolia/Yucca schidigera-Larrea tridentata (n = 21)

Yucca brevifolia/Larrea tridentata-Eriogonum fasiculatum (n = 7)

Within Yucca brevifolia, variety jaegeriana (McKelvy 1938) has been described in the eastern Mojave Desert and form herbertii, characterized by rhizomatous clumps, occurs along the western margin of the Mojave Desert and on the slopes of the Tehachapi Mts., the southern Sierra Nevada, and the Antelope Valley (Webber 1953). Although, the taxonomic fluidity of the species is now better understood (see Hickman 1993; Baldwin et al. 2002) and these subspecific taxa are now not widely accepted, there is some ecological validity in recognizing them. The low, rhizomatous forms of Y. brevifolia corresponding to forma herbertii are Widely known to resprout well follOWing fire (Fig. 22.13), whereas other forms tend to either resprout weakly or are killed by fire. Perhaps this is the reason why the rhizomatous form is the most Widespread in the western Mojave where extensive fire has

Yucca schidigera is one of the most characteristic shrubs of the midelevation eastern and central Mojave Desert. Mojave yucca, though not forming such spectacular stands as Y. brevifolia is also a characteristic species of the upper bajada zone. Like its congener, it rarely dominates in cover greater than 4% or 5%, but is a strong indicator (Table 22.2). Ecologically both Y. brevifolia and Y. schidigera occupy similar niches (see Figs. 22.3 d and e). But it appears that Y. schidigera requires more reliable summer moisture and may tolerate slightly higher summer temperatures. Thus, unlike Y. brevifolia, it does not occur in the western or northern Mojave where summer rain is less significant. Ordination and classification of 94 releves in MDEl project suggest that Yucca schidigera shrubland is ecologically similar to the Yucca brevifolia shrubland, but it tends to occur at slightly lower elevations and on shallower soils. Yucca schidigera Shrubland grades into Larrea tridentata-Ambrosia dumosa shrubland at lower elevations and is similar to several other midelevation alliances including: Ephedra nevadensis (rockier slopes), Coleogyne ramosissima (often caliche layer), Grayia spinosa (deeper alluvial soils), and Eriogonum fasciculatum and Salazaria mexicana (higher disturbance) shrublands. Yucca schidigera is also tolerant of calcareous soils and regularly occurs in skeletal soils or even on exposed outcrops of limestone in the eastern Mojave. Mojave yucca vegetation was represented in several associations defined in the MDEI study and in the JOTR study. These include: Yucca schidigera-Coleogyne ramosissima: upper bajadas old geomorphic surfaces, pediments Yucca schidigera/Pleuraphis rigida: well-drained sandy soils or burned upper bajadas Yucca schidigera-Ephedra nevadensis: rocky, usually granitic slopes

MOJAVE DESERT SCRUB VEGETATION

635

TABLE 22.8

Frequency and Range and Average Cover of Species in 54 Samples of Vegetation Within the Upper Slope Mojave Scrub Zone

Species

Frequency

% Cover Range

Mean % Cover

Ephedra nevadensis

0.850

0.5-5

2.05

Sphaeralcea alnbiglla

0.760

0.5-2

0.42

Lycium andersonii

0.700

0.5-7

1.57

Grayia spinosa

0.650

0.5-19

2.55

Atriplex confertifolia

0.560

0.5-3

1.06

Krascheninnikovia lanata

0.560

0.5-6

0.7

Achnatherum specioszlln

0.540

0.5-4

0.63

Eriogonum inf1atllm

0.500

0.5-1

0.3

Yucca brevifolia

0.430

0.5-4

0.94

Menodora spinescens

0.370

0.5-8

1.21

Artemisia spinescens

0.330

0.5-1

0.28

Eriogonum fasciculatum

0.330

0.5-4

0.46

Larrea tridentata

0.330

0.5-6

0.75

Ambrosia salsola

0.310

0.5-5

0.49

Opuntia basilaris

0.310

0.50

Xylorhiza tortifolia

0.310

0.5-2

Alrzbrosia dumosa

0.280

0.5-4

0.44

Ericameria cooped

0.260

0.5-4

0.3

Pleuraphis jamesii

0.260

0.5-5

0.39

Tetradynlia axillaris

0.240

0.5-4

0.38

Lepidium fremontii

0.200

0.5-1

Mirabilis bigelovii

0.200

0.50

Salazaria mexicana

0.200

0.5-4

0.44

Stanleya elata

0.200

0.5-1

Achnatherum hylnenoides

0.200

0.5-1

Coleogyne ramosissima

0.190

0.5-4

0.29

Gutierrezia lnicrocephala

0.190

0.5-6

0.25

Opuntia echinocarpa

0.170

0.50

Stephanolneria paucitlora

0.170

0.5-1

Poa secunda

0.150

0.5-4

Thamnoslna lnontana

0.150

0.5-1

Acalnptopappus shockleyi

0.150

0.5-1

Arabis pulchra

0.130

0.50

Castilleja angustifolia

0.130

0.50

ElylrzUS elymoides

0.130

0.5-1

Ericameria teretifolia

0.130

0.5-2

Kralneria erecta

0.130

0.5-1

NOTE: Data from MDEI (Thomas et al. 2003). Only those species with frequency >0.12 are listed. This includes samples assigned to Lycium (mdersonii, Gmyia spinosa, Yucca brevi(olia, Menodom spinescens, and Ephedm nevadensis alliances within the Mojavean upper slope group. t = trace tyo cover.

Yucca sclJid(s;era-Opulltia acanthocarpa: upper bajadas, often with grazing history Yucca sclzidigera-Larrea tridentata-(Sinunondsia clzinensis): local in low-energy washes in Newberry Mountains and Joshua Tree National Park, transition with Sonoran Desert Yucco sclzidigera-Larrea tridentata-Ambrosia dunlosa: 111iddle bajadas transition with Larrea-Anlbrosia Yucca schidigera-Arnbrosia durnosa: Iniddle bajadas disturbed areas with Larrea reduction The range in associated species within these plant associations again suggests a range of temperature and precipitation less broad than the Yucca brevifolia vegetation and more ske\,yed toward the warmer lower elevations. Yeaton et a1. (1985) cOlupared the three species of cooccluing yuccas (Y. brevifolia/ Y. schidigera/ and Y. baccota) along a 700-m elevational gradient in the eastern Mojave. They found that vegetation increases in density and complexity froln lower elevations to higher. Species composition also changes constantly with increasing elevation from creosote bush scrub to pinyon-juniper woodland. This indicates that teluperature decreases and rainfall increases with increasing elevation over this transect. 17. schidigera and Y. hrevitc)/ia are found together in the lower portion of the elevational gradient. Y. baccota replaces Y. schidigera at higher elevations. The maximulll density and biomass for all three Yucca species totaled occurs at 1,375 m. Y. schidigero and Y. brevitc)/ia/ which are associated extensively over the gradient, reach peaks of abundance toward the upper edges of their range/ but Y. baccata shows no change in its density or bio111ass fro111 the upper portions of its range down to 1,450 m where it drops out suddenly, to be replaced down slope by Y. sclzidigera. Yeaton et a1. (1985) measured several physiological parameters for the three species and found Y schidigera and Y. baccata to be lllOSt similar in all of them including: transpiration rates, water potential, sto111atal density, and temperature of upper leaf surface. They also report that competition (n1easured by individual sizes and spacing differences) is greatest between Y. schidigera and 17. baccata. They proposed that the similarity of Y schidigera and Y baccata causes interspecific competition that limits the lower edge of the distribution of 17. baccata. They further proposed that the lTIode of cOlnpetition between the two species is water utilization, in particular, which subsurface zones might be used by the various species. In the MDEI data low stature «0.5 m) Y. baccata is an indicator of the upper elevation "Great Basin" woodlands and shrublands, while taller "shrubby" stature Y. baccata is a slightly weaker indicator of the upper bajadas than Y schidigera. The core of the range of Mojave Yucca vegetation is from the ()rd and Newberry mountains southeast of Barstow to the Kingston Range in southeast Inyo County and on east and south to the upper bajadas and slopes of the Clark, New York, Providence and Granite Mountains and at upper ele-

vations ranging south through Joshua Tree NP and along the eastern base of the Peninsular Range as least as far as the southern end of Anza Borrego Desert State Park. Yucca schidigera is a long-lived species indicative of long persisting stands of vegetation. Rowlands et a1. (1982), Vasek (1995), and others describe Y schidigera clonal rings and ascribe great age to them. Evidence suggests that the slow-growing Yucca schidigera is particularly susceptible to deep soil disturbances and recovers very slowly (Tratz 1978). Although Yucca schidigera may persist for long periods, other components of the stand may be less persistent. Unlike a number of associated desert species, fire usually does not kill Yucca schidigera even when aboveground vegetation is totally consumed. In chaparral-desert ecotones of southern California less than 10(Y() of all Mojave yuccas were actually killed by fire (Tratz). In desert grassland, only a few plants were killed by a summer fire, which removed old shoots to or near the ground level (Vasek, Johnson, andH. Eslinger 1975). Mechanical injury other than fire can also result in re-sprouting, although the 1110re severe the injury, the less vigorous the sprouting (Vasek 1995). It can sprout frolll roots protected by overlying soil, or from surviving active tissues at the stem base. Certain dry, rocky sites occupied by Yucca schidigera may lack sufficient fuels to carry a fire in ordinary circumstances. It is likely that stands with a high understory cover of Pleuraphis spp. or disturbance related shrubs might have had higher fire frequencies than those with long-lived nonsprouting desert shrubs. Very few seedlings have been observed on many of the harsher Yucca schidigera shrubland sites. Reproduction by seed may have been much more important during 1110re favorable climatic regimes. Most regeneration now probably occurs through root sprouting, after fire or mechanical disturbance.

Other Upper Bajada Vegetation Types

Other vegetation types of this zone include: Prunus fasciculata intermittently flooded shrubland alliance, Ericarneria teretifolia shrubland alliance, Eriogonurn fasciclllatllrn shrubland alliance/ Salvia dorrii Dwarf-shrubland Alliance, Pleuraphis rigida Herbaceous Alliance (aka Hilaria rigida) , Achnatherufn specioslun herbaceous alliance (aka. Stipa speciosa), Salazaria mexicana shrubland alliance, and Viguiera parishii shrubland alliance. Many of these alliances are disturbance-related, often found in channels or in sites disturbed by grazing, fire, or other means, and are relatively poorly defined by strong indicators, or by nondisturbance related environmental variables. See Sawyer, Keeler-Wolf, and Evens (ms 2007) or Tho111as et a1. (2004) for additional information. Cody (1986) studied shrub spacing patterns on the upper bajadas of the Granite Mountains and in the Mid Hills of the Californian eastern Mojave. Species tending to be uniformly distributed, especially at higher densities, include Yucca brevifolia/ 0pulltia echillocarpa/ Opulltia rafnosissirna/ and Yucca schidigera. Species that tend toward a clu111ped

MOJAVE DESERT SCRUB VEGETATION

637

distribution, especially at higher densities, include Alnbrosia (Hymenoclea) salsola, Acamptopappus sphaerocephalus, Ericarneria cooped, Salazaria mexicana, and Salvia dorrii. The latter clumped species are also the relatively short-lived disturbance followers. Perhaps the clumped distributions observed by Cody are mostly a result of their opportunism (including high dispersibility and fecundity) and thus their ability to dominate on localized disturbances. Some of the highest cover of cactus in the Mojave is also found in this zone. In many cases the dominants are buckhorn cholla (Opuntia acanthocarpa) and silver cholla (0. echinocarpa). However, several other species such as Echinocactus polycephalus, Echinocereus engelmannii, E. triglochidiatus, Ferocactus cylindraceus, Opllntia basilaris, O. chlorotica, and O. rarnosissima also may occur in these stands. Although high cover of O. acanthocarpa and O. echinocarpa is visually striking, they are such widespread species throughout the desert (including lower bajadas and upper slopes) that they are not good indicators of vegetation, and are most often indicative of past or ongoing grazing or other disturbance. Cody (1993) has shown that several species of Opuntia in the eastern Mojave have nonrandom clumped distributions and are associated with the grass Plellraphis rigida. Cody suggested that P. rigida may play the role of a nurse plant for these cactus species.

UPPER SLOPE MOJAVE SCRUB ZONE

This zone is probably best individuated from the previous upper bajada zone by its slightly cooler yearly temperatures and higher precipitation both in summer and the winter. It is also more likely to occur on shallower soils of hills and mountain slopes than on deep alluvium of upper alluvial fans and bajadas. However, there are numerous examples of stands of this vegetation occurring on alluvial as well as colluvial soils. This zone is similar to the upper bajada zone in that it is composed of a mixture of shrubs that may form different combinations of dominants, which are poorly defined by simple climatic or substrate environmental correlates. Thus, although fine-scale classification at the National vegetation classification level of the alliance and association has been attempted, further work needs to be done to synthesize the ecological relationships between the floristic alliances mentioned in the preceeding paragraph. As with the upper bajada zone, it is likely that many shifts in species composition have more to do with disturbance history than with other environmental correlations. This zone includes a varying palette of scrubs including hopsage and Anderson's wolfberry scrub (Grayia spinosa, Lycium andersonii, respectively) and the short-stature version (generally under 3 m tall) of Joshua tree scrub. Intensive sampling and analysis in this zone identified several new types of vegetation such as spiny menodora (Menodora spinescens) and Nevada joint-fir (Ephedra nevadensis) scrub. The eastern portion of this zone also includes little galletta

638

MOJAVE DESERT SCRUB VEGETATION

(Pleuraphis jalnesii) scrub steppe, which is more widespread to the east in the Great Basin and Colorado Plateau.

Grayia spinosa (Hopsage) and Lycium andersonii (Anderson's Wolfberry) Scrub

Grayia spinosa scrub (Sawyer and Keeler-Wolf 1995; Rowlands 1995) or Grayia-Lycizl1n scrub (Beatley 1976) has been discussed by several authors. Beatley suggests that this vegetation best characterizes the transitional zone of her study of the Nevada Nuclear Test site. It is considered intermediate in climate between the vegetation of the lower Mojave Desert, with a high Larrea and Ambrosia dllmosa component, and the higher elevation and higher precipitation Coleogyne zone. The main shrubs in this group are generally not as well represented in the modal climatic zone of either the Great Basin or the lower Mojave. Rowlands (1995) cites extensive stands of hopsage scrub in the Panamint Mountains of Inyo Co., where he describes the component species, in addition to Grayia, to be Lyciurn andersonii, L. pallidum, Ericameria cooperi, as well as Larrea tridentata and Ambrosia dumosa at lower elevations on locally xeric sites. Rowlands characterizes the soils as moderately sandy to loamy and the climatic conditions to be similar to the blackbush zone (see following account). Beatley (1976) suggests that such sites fall in a narrow range of mean annual rainfall (166-179 mm) with low minimum temperatures of - 20. 5°C to -17.2°C. A number of the species she lists as most characteristic of this zone (in addition to Lyciurn andersonii and Grayia), including Artemisia spinescens, Tetradymia axillaris, Krascheninnikovia lanata, Ephedra nevadensis, Achnatherllm hymenoides, Yucca brevifolia, Ericameria cooperi, and Acamptopappus shockleyi, are also among the principal indicators defined in the analysis of the MDEI data (Table 22.2). The general aspect of the Grayia and L. andersonii zone is of a scrub averaging 32(Yo to 3791) cover (Beatley 1976) with an average shrub height of about 0.4 to 0.5 m. Vegetation cover from the 27 plots sampled in the MDEI study ranged between 7qt(J and 40(~6. Although several species are centered in this zone, the samples from MDEl suggest that there is no single species that is an overall indicator of this zone. Even some of the best indicators of this zone only occur in slightly greater than 250/0 of the samples (Table 22.8). The best indicator, Ephedra nevadensis, only occurs in 8Sq,() of the samples. One of the most distinctive vegetation types of this zone is the Menodora spinescens scrub. This type was not widely recognized before the analysis of the MDEI data, but is a distinctive low thorny scrub scattered throughout the mountains of the northeastern Mojave (Fig. 22.14). These stands, characterized by a high cover of Menodora spinescens relative to other shrubs and herbs, are usually found on shallow, welldrained, gravelly soils of upper bajadas and slopes. Menodora, despite its occupation of the transitional upper Mojave Desert climatic belt, is apparently frost sensitive. Webb et al. (2003) report severe reduction of Menodora cover probably as a result of a hard freeze in 1990. Menodora spinescens, like

years of 1996 and 1999 to 2002 as the principle cause of this shift. They suggest that the increases in Larrea involved both its ability to persist after the demise of the more moisturedemanding shrubs (e.g., Grayia) and its ability to increase in size and density during the good rainfall years. BLACKBUSH-DESERT CHAPARRAL-CALlFORNIA JUNIPER ZONE

This zone not only includes the well-known blackbrush (Coleogyne ramosissima) scrub but also includes stands of desert transition chaparral and California Juniper (Juniperus californica), and several disturbance-related vegetation types. It occurs in the wetter portions of the desert including the higher mountains of the eastern Mojave, as well as those bordering the desert on the west. Beatley (1976) characterizes this zone as haVing higher precipitation (annual totals ranging from 225 to 240 mm in the area of the best Coleogyne development in south-central Nevada) than the Larrea communities. However, the temperature ranges are more similar to the Larrea communities than the cooler Great Basin communities. The inclusion of transitional desert chaparral and California Juniper as a part of this zone in the California Mojave makes ecological sense, as these types are also known to be representatives of the warmer and relatively moister cismontane parts of the state. Blackbush Scrub FIGURE 22.14 Menodora spineseens dwarf scrub on the NE-facing slopes of the Inyo Mountains, Inyo Co. Note the scattered low-stature YI/eea brevi(olia. Plel/raphis ;amesii, Ephedra nevadensis, Castille;a angl/sti(olia, and Kraseheninnikovia lanata are also present in the stand.

Coleogyne ramosissima, extends beyond the confines of the

Mojave to the adjacent southern Great Basin Desert. Often found in similar areas are stands of Pleuraphis jamesii, the little galletta grass. Although it may form large open stands in northern Arizona, southeastern Utah, and south and central Nevada (West 1988; Eric Peterson NV heritage ecologist, personal communication 2002), in California, these stands are usually shrub-steppes with scattered shrubs of Ephedra nevadensis, Grayia spinosa, Yucca brevifolia, Lycium sp.,and so forth. They tend to occur on lower to midslopes in the mountains of the northeastern Mojave and are particularly notable in portions of the Inyo Mountains. In some cases, both Pleuraphis rigida (big galletta grass) and P. jamesii may be found growing under similar conditions. However, P. rigida is usually found in sandier soils or if on rocky slopes is at lower elevations. Following a 37-year monitoring interval, from 1964 to 2002, Webb et al. (2003) have demonstrated a marked decrease in density and cover of Grayia spinosa, and in some cases Lycium andersonii and Menodora spinescens at the Nevada Test Site. Plots that were dominated by Grayia and Lycium in 1964 were in 2001 to 2002 in many cases dominated by Larrea tridentata. In other cases sites that were dominated by Menodora now have more cover of other species. Webb et al. invoke the drought of 1989 to 1991 and to a lesser extent the drought

Blackbush, or "blackbrush " (Coleogyne ramosissima), is a monotypic paleoendemic (Raven and Axelrod 1978) shrub of the American Southwest, which ranges from the Colorado Plateau to the Mojave Desert and the southern Great Basin and the western borders of the Sonoran Desert (Keeler-Wolf, Roye, and Lewis 1998). It dominates on rocky or shallow soils on upper bajadas, pediments, and rocky (sometimes calcareous) slopes in the transition zone between the warm deserts and the Great Basin. Some considered the vegetation dominated by Coleogyne as part of the Great Basin (Rowlands 1995), Mojave (Charlton 2003), or a separate transitional setting (Beatley 1976). This debate is still open, judging from the conflicting results from the MDEI (Table 22.2) and data from Joshua Tree National Park, but the more complete data set from the MDEI suggests that it is more closely related to the Mojave than to the Great Basin. The blackbush community is Widespread from elevations of 1,200 to > 1,800 m. The climate is colder than that in areas occupied by creosote bush, and snow is common for short periods during the winter. Several authors (Sawyer and Keeler-Wolf 1995; West 1988; Vasek and Barbour 1977) have commented on the stands of Coleogyne occurring on cemented duripan (calliche) or older alluvial fans. In general, where Coleogyne stands occur on fans and bajadas, they tend to occupy the older geomorphic surfaces and do not occur on more recently eroded sediments (Spolsky 1979; Webb, Steiger, and Turner 1987). However, blackbush communities also occur on steeper colluvial surfaces with little soil development. Blackbush stands also tend to have relatively high cover and uniform height. Beatley

MOJAVE DESERT SCRUB VEGETATION

639

(1976) reports cover of up to 45