Phenology: An Integrative Environmental Science

PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE

Tasks for vegetation science 39 SERIES EDITORS A. Kratochwil, University of Osnabrück, Germany H. Lieth, University of Osnabrück, Germany

The titles published in this series are listed at the end of this volume.

PHENOLOGY: An Integrative Environmental Science

Edited by

MARK D. SCHWARTZ Department of Geography, University of Wisconsin – Milwaukee, Milwaukee, WI, U.S.A.

KLUWER ACADEMIC PUBLISHERS DORDRECHT / BOSTON / LONDON

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN 1-4020-1580-1

Published by Kluwer Academic Publishers, P.O. Box 17, 3300 AA Dordrecht, The Netherlands. Sold and distributed in North, Central and South America by Kluwer Academic Publishers, 101 Philip Drive, Norwell, MA 02061, U.S.A. In all other countries, sold and distributed by Kluwer Academic Publishers, P.O. Box 322, 3300 AH Dordrecht, The Netherlands.

Printed on acid-free paper

All Rights Reserved © 2003 Kluwer Academic Publishers No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed in the Netherlands

Dedication

This book is dedicated to my parents, Marjorie H. and the late Donald J. Schwartz, who nurtured my early interest in science

v

Contents

Dedication

v

Contributing Authors

xi

Preface

xvii

Color Plates

xxi

Foreword

xxvii

Part 1: INTRODUCTION

1

1.1 Introduction MARK D. SCHWARTZ

3

Part 2: PHENOLOGICAL DATA, NETWORKS, AND RESEARCH

9

2.1 East Asia XIAOQIU CHEN

11

2.2 Australia MARIE R. KEATLEY AND TIM D. FLETCHER

27

vii

viii 2.3 Europe ANNETTE MENZEL

45

2.4 North America MARK D. SCHWARTZ AND ELISABETH G. BEAUBIEN

57

2.5 South America L. PATRÍCIA C. MORELLATO

75

2.6 The Global Phenological Monitoring Concept 93 EKKO BRUNS, FRANK-M. CHMIELEWSKI, AND ARNOLD J. H. VANVLIET 2.7 Toward a Multifunctional European Phenology Network ARNOLD J. H. VANVLIET AND RUDOLF S. DEGROOT

105

Part 3: PHENOLOGY OF SELECTED BIOCLIMATIC ZONES

119

3.1 Tropical Dry Climates 121 ARTURO SANCHEZ-AZOFEIFA, MARGARET E. KALACSKA, MAURICIO QUESADA, KATHRYN E. STONER, JORGE A. LOBO, AND PABLO ARROYO-MORA 3.2 Mediterranean Climates 139 DONATELLA SPANO, RICHARD L. SNYDER, AND CARLA CESARACCIO 3.3 Grasslands of the North American Great Plains GEOFFREY M. HENEBRY

157

3.4 High Latitude Climates FRANS E. WIELGOLASKI AND DAVID W. INOUYE

175

3.5 High Altitude Climates DAVID W. INOUYE AND FRANS E. WIELGOLASKI

195

Part 4: PHENOLOGICAL MODELS AND TECHNIQUES

215

4.1 Plant Development Models ISABELLE CHUINE, KOEN KRAMER, AND HEIKKI HÄNNINEN

217

ix 4.2 Animal Life Cycle Models JACQUES RÉGNIÈRE AND JESSE A. LOGAN

237

4.3 Phenological Variation of Forest Trees 255 ROBERT BRÜGGER, MATTHIAS DOBBERTIN, AND NORBERT KRÄUCHI 4.4 Phenological Growth Stages UWE MEIER

269

4.5 Assessing Phenology at the Biome Level XIAOQIU CHEN

285

4.6 Developing Comparative Phenological Calendars REIN AHAS AND ANTO AASA

301

4.7 Plant Phenological "Fingerprints" ANNETTE MENZEL

319

4.8 Phenoclimatic Measures MARK D. SCHWARTZ

331

4.9 Weather Station Siting 345 RICHARD L. SNYDER, DONATELLA SPANO, AND PIERPAOLO DUCE Part 5: REMOTE SENSING PHENOLOGY

363

5.1 Remote Sensing Phenology BRADLEY C. REED, MICHAEL WHITE, AND JESSLYN F. BROWN

365

Part 6: PHENOLOGY OF SELECTED LIFEFORMS

383

6.1 Aquatic Plants and Animals WULF GREVE

385

6.2 Insects KAREN DELAHAUT

405

6.3 Birds 421 TIM H. SPARKS, HUMPHREY Q. P. CRICK, PETER O. DUNN, AND LEONID V. SOKOLOV

x 6.4 Timing of Reproduction in Large Mammals ERIC POST

437

Part 7: APPLICATIONS OF PHENOLOGY

451

7.1 Vegetation Phenology inn Global Change Studies 453 MICHAEL A. WHITE, NATHANIEL BRUNSELL, AND MARK D. SCHWARTZ 7.2 Phenology of Vegetation Photosynthesis 467 LIANHONG GU, WILFRED M. POST, DENNIS BALDOCCHI, T. ANDY BLACK, SHASHI B. VERMA, TIMO VESALA, AND STEVE C. WOFSY 7.3 Radiation Measurements JIE SONG

487

7.4 Phenology and Agriculture FRANK-M. CHMIELEWSKI

505

7.5 Winegrape Phenology GREGORY V. JONES

523

7.6 Long-Term Urban-Rural Comparisons CLAUDIO DEFILA AND BERNARD CLOT

541

Acknowledgments

555

Index

557

Contributing Authors

Aasa, Anto, Institute of Geography, University of Tartu, Tartu, Estonia Ahas, Rein, Institute of Geography, University of Tartu, Tartu, Estonia Arroyo-Mora, Pablo, Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT, USA Baldocchi, Dennis, Department of Environmental Science, Policy & Management, University of California, Berkeley, CA, USA Beaubien, Elisabeth G., Devonian Botanic Garden, University of Alberta, Edmonton, Alberta, Canada Black, T. Andy, Faculty of Agricultural Sciences, University of British Columbia, Vancouver, Canada Brown, Jesslyn F., SAIC, USGS EROS Data Center, Sioux Falls, SD, USA Brügger, Robert, PHENOTOP, Institute off Geography of the University of Berne, Berne, Switzerland Brunns, Ekko, Department of Networks and Data, German Meteorological Service, Offenbach, Germany Brunsell, Nathaniel, Department of Civil Engineering, Duke University, Research Triangle, NC, USA

xi

xii Cesaraccio, Carla, Agroecosystem Monitoring Laboratory, Institute of Biometeorology, National Research Council, Sassari, Italy Chen, Xiaoqiu, Department of Geography, College of Environmental Sciences, Peking University, Beijing, China Chmielewski, Frank-M., Subdivision of Agricultural Meteorology, Institute of Crop Sciences, Faculty of Agriculture and Horticulture, HumboldtUniversity, Berlin, Germany Chuine, Isabelle, CEFE-CNRS, Montpellier, France Clot, Bernard, Biometeorology, MeteoSwiss, Zürich and Payerne, Switzerland Crick, Humphrey Q. P., British Trust for Ornithology, Thetford, UK Defila, Claudio, Biometeorology, MeteoSwiss, Zürich and Payerne, Switzerland deGroot, Rudolf S., Environmental Systems Analysis Group, Department of Environmental Sciences, Wageningen University, Wageningen, The Netherlands Delahaut, Karen, Department of Horticulture, University of WisconsinMadison, Madison, WI, USA Dobbertin, Matthias, WSL Swiss Federal Institute for Forest, Snow and Landscape Research, Forest Ecosystems and Ecological Risks Division, Birmensdorf, Switzerland Duce, Pierpaolo, Agroecosystem Monitoring Laboratory, Institute of Biometeorology, National Research Council, Sassari, Italy Dunn, Peter O., Department of Biological Sciences, University of Wisconsin-Milwaukee, Milwaukee, WI, USA Fletcher, Tim D., Department of Civil Engineering, Monash University, Clayton, Victoria, Australia Greve, Wulf, German Center for Marine Biodiversity Research (Senckenberg Research Institute), Hamburg, Germany

xiii Gu, Lianhong, Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA Hänninen, Heikki, Department of Ecology and Systematics, University of Helsinki, Helsinki, Finland Henebry, Geoffrey M., Center for Advanced Land Management Information Technologies (CALMIT), School of Natural Resources, University of Nebraska, Lincoln, NE, USA Inouye, David W., Department of Biology, University of Maryland, College Park, MD, USA Jones, Gregory V., Department of Geography, Southern Oregon University, Ashland, OR, USA Kalacska, Margaret E., Earth and Atmospheric Sciences Department, University of Alberta, Edmonton, Alberta, Canada Keatley, Marie R., School of Resource Management, University of Melbourne, Creswick, Victoria, Australia Kramer, Koen, Alterra, Department of Ecology and Environment, Wageningen University, Wageningen, The Netherlands Kräuchi, Norbert, WSL Swiss Federal Institute for Forest, Snow and Landscape Research, Forest Ecosystems and Ecological Risks Division, Birmensdorf, Switzerland Lobo, Jorge A., Biology Department, Universidad de Costa Rica, San Jose, Costa Rica Logan, Jesse A., USDA Forest Service, Logan, Utah, USA Meier, Uwe, Federal Biological Research Center for Agriculture and Forestry, Braunschweig, Germany Menzel, Annette, Department of Ecology, TU Munich, Freising, Germany Morellato, L. Patrícia C., Departmento de Botânica, Plant Phenology and Seed Dispersal Research Group, Universidade Estadual Paulista, São Paulo, Brasil

xiv Post, Eric, Department of Biology, The Pennsylvania State University, University Park, PA, USA Post, Wilfred M., Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA Quesada, Mauricio, Centro de Investigaciones en Ecosistemas, Universidad Nacional Autónoma de México, Morelia, México Reed, Bradley C., SAIC, USGS EROS Data Center, Sioux Falls, SD, USA Régnière, Jacques, Natural Resources Canada, Canadian Forest Service, Quebec, Canada Sanchez-Azofeifa, Arturo, Earth and Atmospheric Sciences Department, University of Alberta, Edmonton, Alberta, Canada Schwartz, Mark D., Department of Geography, University of WisconsinMilwaukee, Milwaukee, WI, USA Snyder, Richard L., Department of Land, Air, and Water Resources, University of California, Davis, CA, USA Sokolov, Leonid V., Russian Academy of Sciences, St. Petersburg, Russia Song, Jie, Department of Geography, Northern Illinois University, Dekalb, IL, USA Spano, Donatella, Department of Economics and Woody Plant Ecosystems, University of Sassari, Sassari, Italy Sparks, Tim H., Centre for Ecology and Hydrology, Monks Wood, UK Stoner, Kathryn E., Centro de Investigaciones en Ecosistemas, Universidad Nacional Autónoma de México, Morelia, México vanVliet, Arnold J. H., Environmental Systems Analysis Group, Department of Environmental Sciences, Wageningen University, Wageningen, The Netherlands Verma, Shashi B., School of Natural Resource Sciences, University of Nebraska, Lincoln NE, USA

xv Vesala, Timo, Department of Physical Sciences, University of Helsinki, Helsinki, Finland White, Michael, Department of Aquatic, Watershed, and Earth Resources, Utah State University, Logan, UT, USA Wielgolaski, Frans E., Department of Biology, University of Oslo, Oslo, Norway Wofsy, Steve C., Department of Earth and Planetary Sciences, Harvard University, Cambridge, MA, USA

Preface

I recall as a doctoral student at the University of Kansas discussing dissertation topics with my advisor, Prof. Glen A. Marotz, one day in 1983. He had just suggested to me that phenology was an interesting topic, and one that held promise for important research contributions in the future. “What’s that?” I asked, thus beginning my career as a phenologist, and the long path that led to my editorship of this volume. Skipping ahead a decade, I was encouraged by my colleague Elisabeth Beaubien to attend the 13th International Congress of Biometeorology, which was being held in Calgary, Alberta, Canada that year. I did attend, and also met Prof. Dr. Helmut Lieth there for the first time. I had corresponded with him while writing my dissertation, having gained much insight from his seminal book, Phenology and Seasonality Modeling g at that time. At the Calgary meetings Prof. Lieth helped Elisabeth and I reactivate a Phenology Study Group within the International Society of Biometeorology (ISB). The first workshop of the new group was organized by Dipl.-Met. Hartmut Scharrer of the German Weather Service (DWD) Phenology unit, and scheduled for May 1995. As a UW-Milwaukee assistant professor in the Geography Department at the time, I had never traveled outside of North America, and further did not have a source of travel funding for the trip to Offenbach (just outside Frankfurt), Germany. So I consulted the associate dean responsible for our department, G. Richard Meadows (now Dean of the College of Letters and Science) and he was able to provide me with funds to cover the airfare (after I assured him that this trip would be an important one for establishing my connection to international phenological research). At the Offenbach workshop the thirteen participants proposed an organizational structure and laid out a set of objectives for the Phenology Study Group.

xvii

xviii Some subsequent early activities included the launching of a new journal Phenology and Seasonality (unfortunately discontinued after one issue), and participation in the 14th International Congress of Biometeorology (Ljubljana, Slovenia) in 1996. The study group’s first international scientific meeting was a “Phenology Symposium” that I organized in 1998 as a group of four paper, one poster and one discussion sessions (21 participants) held within the Association of American Geographers Annual meeting in Boston, MA, USA. The number of individuals involved with, amount of research being conducted in, and level of interest by scientists from other disciplines for phenology had all been slowly rising since the early 1990s. However, a series of papers published in Nature (over the 1997-2000 period) dramatically accelerated these trends, especially the interest of global change researchers in remote sensing and biology for phenological data and techniques. In recent years, this surge in interest from the global change research community, and corresponding funding by the European Union of several projects (POSITIVE and EPN, European Phenology Network) have led to a greater number of scientific conferences with increasing numbers of participants. Specifically, a first “stand alone” international phenology conference, organized by Dr. Annette Menzel and colleagues (2000, Freising, Germany, 70 participants), and two subsequent international conferences (organized by Arnold vanVliet and associates) held in Wageningen, The Netherlands in 2001 and 2003 connected with the EPN project (each had just over 100 participants). The two European projects have also supported a large number of workshops on specialized phenology topics for smaller groups of participants (including individuals from other parts of the world). Within the ISB, the study group participated in the 15th International Congress of Biometeorology (held in Sydney, Australia in 1999) and due to reorganization within the society was renamed the Vegetation Dynamics, Climate, and Biodiversity Commission after that meeting. Members of the new ISB commission also participated in the most recent ISB Congress (16th International Congress of Biometeorology, held in Kansas City, MO, USA in 2002). During that meeting the group requested, and was subsequently granted by ISB, the simplified current name “Phenology Commission.” So the sequence of events I have described created the conditions and provided resources to make development of this book possible, namely sufficient interest in the topic by the general scientific community, and an interconnected community of phenological researchers with the necessary diversity of research expertise to cover the range of required topics. Jacco Flipsen, a Kluwer editor, who wrote me a letter in early 2001 stating the need for and asking if I was interested in editing a book on plant phenology,

xix initiated the actual development of this volume. After some negotiation, specifically to allow the book to cover a broader range of phenological topics, the project began in earnest during the first months of 2002. The book was seamlessly transferred into Prof. Lieth’s “Tasks for Vegetation Science” series at Kluwer (supervised by Helen Buitenkamp) in early 2003, and completed later that year.

Mark D. Schwartz Milwaukee, March 2003

xxi

xxii

xxiii

xxvi

Foreword

I was pleased when Mark Schwartz invited me to write a foreword to his volume. And even more so after I had read the content and many of the papers contributed to the volume. My own book on the subject matter (Lieth 1974) appeared as vol. 8 in the famous ecological studies series by Springer, and a contribution to the then fully operating Analysis of Ecosystem program of the U.S. International Biological Program (US-IBP). Phenology was a rather quiet scientific objective at that time. Some operational networks existed in Europe and America mainly in agriculture. Only a few researchers in biology, ecology and meteorology were using the accumulated datasets at that point. Satellite image analyses and the development of new remote sensing techniques were of interest then, but the ground truth observation of biological fluctuating phenomena were regarded as outmoded. The common thrust of the papers presented at the 1972 phenology symposium of the American Institute of Biological Sciences conference in Minneapolis gave phenology work in the U.S. and Europe a big push, and ground truth observations in ecosystems studies were initiated in many parts of the world. The initial successes in modeling phenological events, the comparisons with meteorological parameters, and the correlation attempts with global remote sensing data sets caught the attention of the scientists, working at that time on global change initiated directly and indirectly by humans. This interest continues, and a book presenting the achievements of the last 30 years (two or three generations of graduate students) is very much needed. I followed the results with interest, because I had earlier made predictions that had to be tested, verified or modified through field

xxvii

xxviii observations. Phenological observations and experiments undertaken during the last 30 years have greatly improved insights into ecosystems operation. One of the major values for phenological data is their validation value for seasonality models. These models have gained prominence in global climatic change models to predict biosphere responses to climatic parameter changes. This, however, is by no means the only value of phenological work. The book presented here by Dr. Schwartz includes many other fields of biology for which phenological investigations are needed. The reliance of species association in ecosystems upon a quasi-correct seasonal behavior in a seasonal climate is so prominent, that most investigations and experiments include phenological aspects, be they climatic, physiologic or biochemical. Throughout the historical development of phenology, its practical applications in agriculture and forestry have dominated the field. The chapters in this volume dealing with the history of phenology by Menzel (Chapter 2.3) and Chen (Chapter 2.1) uncovered many local networks that I had not found in the early 1970s. While this is a valuable addition to the field, I found that several important networks and papers had still been neglected. The 1974 volume has, therefore, not completely lost its relevance for future generations of phenologists. The history of European phenology emphasizes agricultural and forest phenology and neglects the body of work started by Heinrich Walter, whose students and coworkers (e.g., Kreeb and Ellenberg) and these together with their coworkers made substantial contributions to phenology in Europe (see Walter 1960, which shows that he had much more influence on phenology than providing the widely used climate diagrams which are so easily available in the climate diagram world atlas by Walter and Lieth 1960ff., and now available on CD by Lieth et al. 1999). The Russian work on phenology is only partly recognized. For me a major omission appears to be the book by Alexander Podolski, which appeared about 2 decades ago in English (1984). His approach to identifying the start of a phenologically valid period from physiological data, rather than an arbitrary chosen, convenient calendar date, still warrants further analyses in relevant cases. Podolski’s volume also includes a wealth of literature otherwise not mentioned in Russian books that mostly refer to papers from west Russian institutes (covered by Dr. Menzel’s historical treatment in this volume). It appears to me that students interested in phenology should be encouraged to read some of the older papers by Hopkins (1938), Thornthwaite (several papers), Hopp, Caprio, a Schnelle and Volkert, and all the others as cited in Lieth (1974) and in this volume, as well as Walter (1960) and Podolski (1984). The literature on remote sensing and global change applications is so new, that for the purpose of this book’s users, the

xxix authors in this area of research will be available in current relevant journals. Many authors of these papers will not include their contribution as part of phenology, but their work deals very often with topics that would be included in seasonality, climate and species fluctuations, global change and methods for the investigation of these topics. All this is phenology in the wider sense. The historical assessment in another 30 years will evaluate the importance of these authors and developments for phenological work. Work on satellite remote sensing had just started around the time I compiled my phenology book. The same was true for computer mapping, which was in its infancy as well. But the combination of data available from different phenological networks in the U.S. through computer modeling and computer mapping was so attractive to graduate students, that many of them choose phenological topics for their degree papers. When I developed my volume in the early 1970s I was greatly supported by Forrest Stearns who was a professor at the University of Wisconsin-Milwaukee. Wisconsin was an intellectual center for phenology at that time where the Lettaus (Heinz and Katharina) provided guidance in meteorological and phenological observations. No wonder then that phenology received new impulses from Wisconsin. In summary I can say that this book edited by Dr. Schwartz shows that phenology is as alive and important as ever. Like any other field of research it undergoes peaks and valleys in recognition. As long as planet earth tumbles around the sun, there will be ecologists and meteorologists, foresters and agronomists, insurance people and a wealth of other specialists observing, measuring and evaluating phenological data. Many of them will use this book. I thank also the responsible persons in Kluwer academic publishers for their interest in presenting this volume with the usual Kluwer quality. I am sure that the book will obtain the worldwide reception accorded many of other previous volumes of the T:VS series.

Helmut Lieth Osnabrück, February 2003

REFERENCES CITED Hopkins, A. D., Bioclimatics–A science of life and climate relations, U.S. Dept. Agr. Misc. Publ. 280, 1938. Lieth, H., editor, Phenology and Seasonality Modeling, Springer-Verlag, New York, 444 pp., 1974.

xxx Lieth, H., J. Berlekamp, S. Fuest, and S. Riediger, Climate Diagram World Atlas on CD (unpaginated electronic publication), Backhuys Publishers, Leiden, Netherlands, 1999. Podolski, A. S., New Phenology: Elements of mathematical forecasting in ecology, John Wiley and Sons, New York, 504 pp., 1984. Walter, H., Grundlagen der Pflanzenverbreitung, part 1 Standortslehre, Eugen Ulmer Verlag, Stuttgart, Germany, 566pp., 1960. Walter, H. and H. Lieth, Klimadiagramm-Weltatlas (unpaginated), VEB Gustav Fischer, Jena, 1960ff.

PART 1

INTRODUCTION

Chapter 1.1 INTRODUCTION Mark D. Schwartz Department of Geography, University of Wisconsin-Milwaukee, Milwaukee, WI, USA

Key words:

1.

Definitions, Environment, Organization, Modeling, Global Change

BASIC CONCEPTS AND BACKGROUND

Phenology, which is derived from the Greek word phaino meaning to show or to appear, is the study of periodic biological events in the animal and plant world as influenced by the environment, especially temperature changes driven by weather and climate. Sprouting and flowering of plants in the spring, color changes of leaves in the fall, bird migration and nesting, insect hatches, and animal hibernation are all examples of phenological events (Dubé et al. 1984). Seasonality is a related term, referring to similar non-biological events, such as timing of the fall formation and spring breakup of ice on fresh water lakes. Human knowledge and activities connected to what is now called phenology are probably as old as civilization itself. Surely, soon after farmers began to continuously dwell in one place—planting seeds, observing crop growth, and carrying out the harvest year after year—they quickly became aware of the connection of changes in their environment to plant development. Ancient records and literature, such as observations taken up to 3000 years ago in China (see Chapter 2.1), and references in the Christian Bible, testify to a common level of understanding about phenology among early peoples: “Learn a lesson from the fig tree. Once the sap of its branches runs high and it begins to sprout leaves, you know that summer is near.” Gospel of Mark 13:28 Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 3-7 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

4

Phenology: An Integrative Environmental Science

Unfortunately, these ancient “roots” did not translate into systematic data collection across large areas over the centuries, nor did they provide impetus for the early development of phenology as a scientific endeavor and discipline. For a long time the field remained tied almost exclusively to agricultural applications, and even those were only deemed practical on the local scale (i.e., every place was different, and generalizations difficult or impossible). With the establishment of continuous and continental-scale observation networks by the mid-1900s (though still largely confined to Europe, see Chapter 2.3), and contributions of early researchers such as Schnelle (1955), phenology began to emerge as an environmental science. Lieth’s (1974) book was the first modern synthesis to chart the interdisciplinary extent of the field, and demonstrate its potential for addressing a variety of ecological system and management issues. These foundations have prepared the way for this volume.

2.

ORGANIZATION AND USE

Phenological research has traditionally been identified with studies of mid-latitude plants (mostly trees and shrubs) in seasonal climates, but other areas of the field are also progressing. Thus, a principal goal in organizing this book was to overcome this mid-latitude plant bias with a structure that would facilitate a thorough examination of wider aspects of phenology. After this introduction, the second section, Phenological Data, Networks, and Research, adopts a regional approach to assess the state and scope of phenological research with chapters on East Asia (2.1), Australia (2.2), Europe (2.3), North America (2.4, excluding Mexico), and South America (2.5). Several major regions, most notably Africa and central Asia were not included due to my inability to identify researchers working in these geographical areas. While some efforts were made in these chapters to survey the history of regional data collection and research, more emphasis was given to an assessment of recent developments. My assumption was that since Lieth’s (1974) book had make an extensive survey of the history of phenological research up to the early 1970s, there was no great need to reproduce all that historical information in this volume. Two other chapters in this section explore a plan for a global monitoring network (2.6), and the multifunctional capabilities and uses of continental-scale phenological network data (2.7). Section 3, Phenology of Selected Bioclimatic Zones, examines phenological research in areas outside of mid-latitudes, with chapters on Tropical Dry Climates (3.1) and High Latitude Climates (3.4). Other chapters document phenology in drier mid-latitude biomes, including

Chapter 1.1: Introduction

5

Mediterranean Climates (3.2) and Grasslands of the North American Great Plains (3.3). Lastly, the special phenological responses of High Altitude Climates are explored in Chapter 3.5. Phenological Models and Techniques (Section 4) presents a survey of phenological research methodologies and strategies. Model building and development is outlined in chapters addressing plants (4.1), animal life cycles (4.2, concentrating on insects), and Phenoclimatic Measures (4.8). The challenges of phenological variability within species are explored in Chapter 3.3, and other chapters address the issues of temperature measurement (4.9), standardization of phenological event definitions (4.4), and development of phenological calendars (4.6). The remaining chapters in this section detail methods to detect climate change (4.7) and assess biome level phenology (4.5). The next section (5) is devoted entirely to the emerging area of remote sensing phenology. Section 6, Phenology of Selected Lifeforms looks at research and developments in animal phenology, including chapters on Aquatic Plants and Animals (6.1), Insects (6.2), Birds (6.3), and Timing of Reproduction in Large Mammals (6.4). The final section of the book (7) details Applications of Phenology to a variety of topics. Chapter 7.1 looks specifically at Vegetation Phenology in Global Change Studies, Chapter 7.2 explores frontiers related to the Phenology of Vegetation Photosynthesis, and Chapter 7.3 Phenological Effect on Radiation Measurements. Several remaining chapters in this section explore applications in traditional field agriculture (7.4) and winegrape growth and care (7.5). Lastly, phenological applications to Long-Term Urban-Rural Comparisons are examined in the final chapter of this section (7.6). Therefore, this volume’s structure is primarily designed to serve the basic reference needs of phenological researchers and students interested in learning more about specific aspects off the field, or evaluating the feasibility of new ideas and projects. However, it is also an ideal primer for ecologists, climatologists, remote sensing specialists, global change scientists, and motivated members of the public who wish to gain a deeper understanding of phenology and its potentials.

3.

FUTURE DIRECTIONS AND CHALLENGES

When I chose the name for this book, I deliberately selected the word “integrative” because of its implication of a process. Phenology is an interdisciplinary environmental science, and as such brings together individuals from many different scientific backgrounds, but the full benefits of their combined disciplinary perspectives to enrich phenological research

6

Phenology: An Integrative Environmental Science

have yet to be realized. Thus, the term “integrative” as in moving together, rather than “integrated,” implying already being together. The last five years have seen rapid progress in the transmission of “phenological perspectives” into the mainstream of science, especially related to the needs of global change research. While other parts of phenological research are still important and need to progress, it is global change science that will stimulate, challenge, and transform the discipline of phenology most in the coming decades. In order to maximize the benefits of phenology for global change research as rapidly as possible, commitments to integrative thinking and large-scale data collection must continue. First of all, the limitations of the primary forms of data collection (remote sensing derived, native species, cloned indicator species, and model output) must be accepted. None of these data sources can meet the needs of all research questions, and an “integrative approach” that combines data types provides synergistic benefits (Schwartz 1994, 1999). The most needed data are traditional native and cloned plant species observations. Networks that select a small number of common plants for coordinated observation among national and global scale networks will prove the most useful. These networks should be embraced and integrated into the missions of national weather services around the world, as is now the case in many European countries (see Chapter 2.3). A little more than one hundred years ago, the countries of the world began to cooperate in a global-scale network of weather and climate monitoring stations. The results of this long-term investment are the considerable progress that has been made in understanding the workings of the earth’s climate systems. I believe that we have a similar opportunity with phenological data, and that small investments in national and global-scale observation networks are crucial to global change science, and will yield an impressive return in the years ahead.

REFERENCES CITED Dubé, P. A., L. P. Perry, and M. T. Vittum, Instructions for phenological observations: Lilac and honeysuckle, Vermont Agricultural Experiment Station Bulletin 692, University of Vermont, Burlington, 7 pp., 1984. Lieth, H., editor, Phenology and Seasonality Modeling, Springer-Verlag, New York, 444 pp., 1974. Schnelle, F., Pflanzen-Phänologie, Akademische Verlagsgesellschaft, Geest and Portig, Leipzig, 299 pp., 1955. Schwartz, M. D., Monitoring global change with phenology: the case of the spring green wave, Int. J. Biometeorol., 38, 18-22, 1994.

Chapter 1.1: Introduction

7

Schwartz, M. D., Advancing to full bloom: planning phenological research for the 21st century, Int. J. Biometeorol., 42, 113-118, 1999.

PART 2

PHENOLOGICAL DATA, NETWORKS, AND RESEARCH

Chapter 2.1 EAST ASIA Xiaoqiu Chen Department of Geography, College of Environmental Sciences, Peking University, Beijing, China

Key words:

China, Japan, Networks, Models, Data

1.

PHENOLOGICAL OBSERVATION AND RESEARCH IN CHINA

1.1

Historical Background

Modern phenological observation and research in China started in the 1920s with Dr. Kezhen Zhu (1890-1974), who may be regarded as the founder of modern Chinese phenology. As early as 1921 he observed spring phenophases of several trees and birds in Nanjing. In 1931, he summarized phenological knowledge from the last 3000 years in China. He also introduced phenological principles (e.g. species selection, criteria of phenological observations and phenological laws) developed in Europe and the United States from the middle of the eighteenth to the early twentieth century (Zhu 1931). In 1934, he organized and established the first phenological network in China. Observations of some 21 species of wild plants, 9 species of fauna, some crops, and several hydro-meteorological events ceased in 1937 because of the War of Resistance Against Japan (1937-1945). Twenty-five years later the Chinese Academy of Sciences (CAS) established a countrywide phenological network under the guidance of Dr. Zhu. The observations began in 1963 and continued until 1996. Observations resumed in 2003, but with a reduced number of stations, Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 11-25 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

12

Phenology: An Integrative Environmental Science

species, and phenophases. In addition, the Chinese Meteorological Administration (CMA) established a countrywide phenological network in the 1980s.

1.2

Networks and Data

The observation program of the CAS network included a total of 173 observed species. Of these, 127 species of woody and herbaceous plants had a localized distribution. Table 1 lists the 33 species of woody plants, two species of herbaceous plants, and 11 species of fauna that were observed across the network (Institute of Geography at the Chinese Academy of Sciences 1965, Table 1). Since 1973, several stations added phenological observation of major crops. These observations were carried out mainly by botanical gardens, research institutes, universities and middle schools according to uniform observation criteria (Institute of Geography at the Chinese Academy of Sciences 1965; Wan and Liu 1979). The phenophases of woody plants included bud-burst, first leaf unfolding, 50% leaf unfolding, flower bud or inflorescence appearance, first bloom, 50% bloom, the end of blooming, fruit or seed maturing, fruit orr seed shedding, first leaf coloration, full leaf coloration, first defoliation, and the end of defoliation. The Institute of Geography at the Chinese Academy of Sciences took responsibility for collecting the phenological data and publishing them. Changes to the stations and in observers over the years resulted in data that were spatially and temporally inhomogeneous. The number of active stations has varied over time. The largest number of stations operating was 69 in 1964 and the lowest number occurred between 1969 and 1972 with only four to six stations active. The phenological data from 1963 to 1988 were published in form of Yearbooks of Chinese Animal and Plant Phenological Observation. Table 2.1-1. Common observation species of the CAS phenological network in China. Woody plants Ginkgo biloba L. Metasequoia glyptostroboides Hu et Cheng Thuja orientalis L. Juniperus chinensis L. Populus simoniii Carr. Populus canadensis Moench. Salix babylonica L. Juglans regia L. Castanea mollissima Blume. Quercus variabilis Blume.

Chapter 2.1: East Asia

13

Woody plants Ulmus pumila L. Morus alba L. Broussonetia papyrifera (L.) Vent. Paeonia suffruticosa Andr. Magnolia denudata Desr. Firmiana simplex W. F. Wight. Malus pumila Mill. Prunus armeniaca L. Prunus persica Stokes. Prunus davidiana (Carr.) Franch. Albizzia julibrissin Durazz. Cercis chinensis Bge. Sophora japonica L. Robinia pseudoacacia L. Wisteria sinensis Sweet. Melia azedarach L. Koelreuteria paniculata Laxm. Zizyphus jujuba Thunb. Hibiscus syriacus L. Lagerstroemia indica L. Osmanthus fragrans Lour. Syringa oblata Lindl. Fraxinus chinensis Roxb. Herbaceous plants Paeonia lactiflora Pall. Chrysanthemum indicum L. Fauna Apis mellifera L. Apus apus pekinensis (Swinhoe) Hirundo rustica gutturalis Scopoli. Hirundo daurica japonica Temminck et Schlegel. Cuculus canorus Subspp. Cuculus micropterus micropterus Gould. Cryptotympana atrata Fabr. Gryllulus chinensis Weber (Gryllus berthallus Sauss.) Anser fabalis Subspp. Oriolus chinensis diffusus Sharpe. Rana esculenta L.

The CMA phenological network is affiliated with the national-level agrometeorological monitoring network and came into operation in 1980.

14

Phenology: An Integrative Environmental Science

The phenological observation criteria for woody and herbaceous plants, and fauna were adopted from the CAS network. r There are 28 common species of woody plants, one common species of herbaceous plant and 11 common species of fauna. The main phenophases are the same as those of the CAS network. In addition to the natural phenological observations, the network also carries out professional phenological observation of crops on the basis of a specific observation criterion (National Meteorological Administration 1993). The main crop varieties include rice, wheat, corn, grain sorghum, millet, sweet potato, potato, cotton, soybean, rape, peanut, sesame, sunflower, sugarcane, sugar beet, and tobacco. The CMA network is the largest phenological observation system in China at present. There were 587 agrometeorological measurement stations in 1990, of these about 400 stations were undertaking phenological observations. As the phenological and meteorological observations are parallel in this network, the data are especially valuable for understanding phenology-climate relationships. These data can also be used to provide agrometeorological service and prediction on crop yield, soil moisture and irrigation amounts, plant diseases and insect pests, and forest fire danger (Cheng et al. 1993). Guodong Yang and Xiaoqiu Chen established another phenological observation network in 1979, which operated until 1990. The network consisted of approximately 30 stations in the Beijing area under a research project financially supported by the Beijing Higher Education Bureau. Using these data, they worked out and published a series of phenological calendars of the Beijing area (Yang and Chen 1995).

1.3

Research and Applications

Modern phenology research in China focuses mainly on the following: – The development and application of phenological calendars, – Defining phenological seasons and phenological growing seasons, – Phenological mapping, – Phenological modeling and prediction, – Phenology and historical climate change, – Remote sensing of phenophases, and etc. Some important aspects are summarized below. 1.3.1

Phenological calendars

After obtaining the raw phenological data, the primary aim is to compile local phenological calendars that can be used as biological indicators to detect seasonality and do farm work in the right season. Kezhen Zhu and

Chapter 2.1: East Asia

15

Minwei Wan compiled the first phenological calendar based on observational data from 1950 to 1972 in Beijing. This phenological calendar was published in the book Phenology (Zhu and Wan 1973) and consisted of the average, earliest, and latest dates of 129 phenological events. In the 1980s, the Institute of Geography at CAS devised uniform criteria to compile phenological calendar at stations of the CAS network. All together 45 phenological calendars in China were published (Wan 1986, 1987). In each phenological calendar, the main phenological events of plants and fauna, and hydro-climatic events were chosen to represent an ordinal succession of phenophases in the annual cycle at each station. In order to detect the spatial difference of phenological occurrence dates in a relatively small area, a specific observation network was established in the Beijing area (16807.8 km2), which operated between 1979 and 1990. Based on the observed data of this network, 16 phenological calendars were compiled, about one phenological calendar per 1000 km2 (Yang and Chen 1995). In contrast to the phenological calendars of the CAS observation network, each phenological calendar in the Beijing area included almost all observed phenological occurrence dates in order to represent a more detailed and continuous succession of phenophases at the location. In addition, except for the average, the earliest and latest dates of phenological events as well as the standard deviation were calculated to describe the general temporal fluctuation of each phenological event. The results showed that the spatial difference of the average occurrence dates of a spring phenophase was 3-7 days between urban and rural areas on the plain, but it reached 10 days to one month between plain and mountain areas. Generally speaking, phenophases during spring and summer appeared first in the urban area and then in rural and mountainous areas; in contrast, phenophases during autumn and early winter appeared first in mountainous and rural areas and then in the urban area. 1.3.2

Phenological seasons

Phenological calendars describe the occurrence dates of various phenophases and their sequence in the annual cycle, whereas the phenological season represents characteristic stages of the phenological landscape. Several methods have been developed to determine phenological seasons at a station. An earlier method selected representative phenophases as indicators of particular seasons (Schnelle 1955). Since there were few common species of plants at some stations, using the same phenophase to identify a phenological season in a large region like China, is difficult. In order to be able to compare phenological seasons among different stations, both temperature and phenology indicators were applied to determine

16

Phenology: An Integrative Environmental Science

seasons. According to Wan (1986), daily mean temperatures of 3°C and 19°C were thresholds indicating the beginning dates of spring and summer, whereas 19°C and 10°C were thresholds indicating the beginning dates of autumn and winter in China. Beginning dates of sub-seasons were also defined using other specific temperature thresholds. Based on beginning dates of the temperature seasons, corresponding phenological indicators were fixed by referring to the local phenological calendar. The phenological seasons in Beijing are shown in Table 2. However, observations showed that the same plant phenophase occurred under different air temperatures in different areas (Japanese Agrometeorological Society 1963; Reader et al. 1974). This indicates that the occurrence date of a phenophase results from the influence of a combination of environmental factors, including air temperature, precipitation, atmospheric humidity, radiation, soil conditions, etc. Therefore, in order to determine phenological seasons accurately, we should use pure phenological data. The relevant methods will be introduced in Chapter 4.5. Table 2.1-2. Phenological seasons in Beijing (1931-1982). Season

Temperature

Phenological indicator

Period (m/d)

Days

Early spring >3°C Prunus davidiana bb 3/8 – 3/14 7 Mid-spring >5°C Ulmus pumila fb 3/15 – 4/3 20 Late spring >10°C Prunus armeniaca fb 4/4 – 5/7 34 Early summer >19°C Robinia pseudoacacia 50%b 5/8 – 6/11 35 Mid-summer >24°C Albizzia julibrissin 50%b 6/12 – 7/18 37 Late summer <27°C Sophora japonica fb 7/19 – 9/18 62 Early autumn <19°C Fraxinus chinensis sm 9/19 – 9/27 9 Mid-autumn <16°C Morus alba lc-first 9/28– 10/12 15 Late autumn <13°C Populus canadensis fd 10/13– 10/24 12 Early winter <10°C Acer mono lc-full 10/25– 11/11 18 11/12 – 3/7 116 Mid-winter <5°C Prunus davidiana ed bb: bud-burst; fb: first bloom; 50%b: 50% bloom; sm: seed maturing; lc-first: first leaf coloration; fd: first defoliation; lc-full: full leaf coloration; ed: the end of defoliation

1.3.3

Phenology and historical climate change

In China, phenological observation can be traced back to the eleventh century B.C. The earliest phenological calendar, Xia Xiao Zheng, stems from this period and recorded (on a monthly basis) phenological events, weather, astronomical phenomena, and farming activities in the region between the Huai River drainage area and the lower reaches of Yangtze River. In addition, extensive phenological data were recorded in other

Chapter 2.1: East Asia

17

ancient literature over the last 3000 years. These data could to some extent, reflect past temperature and therefore may be used to model past climate. Kezhen Zhu is the pioneer in revealing historical climate change using phenological evidences. Using ancient phenological data and other data, he reconstructed a temperature series of the past 5000 years in China. The main results showed that: (1) during the first 2000 years, the annual mean temperature in most eras may have been 2°C higher than the present with the winter temperature being 3-5°C higher; (2) afterwards there were several fluctuations with the lower temperatures in 1000 B.C., 400 A.D., 1200 A.D. and 1700 A.D. with an amplitude of 1-2°C; and (3) in each period of 400800 years, several smaller cycles of 50-100 years with an amplitude of 0.51°C could be identified. A strong agreement was found between Zhu’s temperature series and the temperature series obtained by variations of the isotope O18 content of the Greenland ice sheet during the last 1700 years. This agreement shows that phenological data acquired from the ancient literature are one of the effective tools, which may be used to study historical climate change (Zhu 1973). According to a rough estimate, about 200 kinds of archaic personal diaries remain in the Chinese literature, in which 10-20% of the diaries recorded phenological data. Using the historical phenological data (flowering dates of Prunus davidiana, Prunus persica, Prunus armeniaca, Syringa oblata, etc.) and other evidence, as well as the modern phenological data since 1950, Gong, et al. (1984) reconstructed the spring phenological series from 1849 to 1981 in Beijing. The statistical analyses indicated that there were 7.4-year, 4-year, and 2-year cycles in the time series. The maximum amplitude in the occurrence dates of spring phenology was 26 days. In general, the onset of spring phenology has advanced by 2.8 days since 1959. 1.3.4

Phenological models and prediction

Several statistical models were established to simulate temporal and spatial phenological performance. In order to predict the flowering of ornamentals, Yang and Chen (1985) established linear regression equations between flowering dates of different trees (1950-1973) in Beijing based on correlation analysis. Some examples are:

yapricot = 19.8 + 0.85 x peach (r = 0.903, p < 0.01) ylilac = 27.5 + 0.82 xapricot (r = 0.834, ρ < 0.01) ylocust = 91.7 + 0.35 xlilac (r = 0.650, ρ < 0.01)

(1)

18

Phenology: An Integrative Environmental Science

where y is the occurrence date of a later phenophase, and x is the occurrence date of an earlier phenophase. This kind of regression equation was also established between spatialtemporal series of flowering dates acquired from three sample stations (Beijing, Tai’an, and Xi’an, 1963-1978) in North China. Spatial extrapolation tests showed that the “regional” statistical models were capable of estimating flowering dates at the other 14 sites within the research region during 1963 and 1978 (Chen and Yang 1988; Chen 1990). Table 2.1-3. Spatial progress rates of plant phenophases in China (Gong and Jian 1983, abridged). Species

Phenophase

b (day/degree)

c (day/degree)

d day/100m

Salix bud burst +3.88 +0.78 +0.97 abylonica Prunus first bloom +3.28 +0.55 +0.81 davidiana Prunus first bloom +3.74 +0.78 +1.54 armeniaca first bloom +3.98 +0.71 +1.36 Prunus persica Wisteria first bloom +2.40 +1.10 +0.73 sinensis Albizzia first bloom +2.53 +0.07 +0.60 julibrissin Sophora first bloom +0.72 +0.19 +0.32 japonica Lagerstroemia first bloom +0.49 +0.25 +0.53 indica Chrysanthemum first bloom -3.81 -0.08 -0.69 indicum Ulmus -0.36 -0.77 End of -3.62 pumila defoliation +: The occurrence date delays from south to north, from west to east, from low altitude to high altitude. -: The occurrence date advances from south to north, from west to east, from low altitude to high altitude.

Another kind of statistical model was constructed between the average occurrence date of a phenophase and geographical coordinates at different stations. The general model for China was described as follows (Gong and Jian 1983):

Chapter 2.1: East Asia

y = a + b(ϕ − 30$ ) + c(λ − 110$ ) + dh

19 (2)

where y is the average occurrence date (day of year) of a phenophase at various stations; ϕ , λ , and h are latitude, longitude, and altitude (unit: 100m) of the stations; and a, b, c, d are coefficients. Table 3 shows that from spring to summer the representative phenophases tended to be delayed by 0.49 to 3.98 days per degree of latitude to the north, 0.07 to 1.1 days per degree of longitude to the east, and 0.32 to 1.54 days per 100m toward higher altitude. In autumn, however, the two phenophases tended to advance by 3.62 to 3.81 days per degree of latitude to the north, 0.08 to 0.36 days per degree of longitude to the east, and 0.69 to 0.77 days per 100m toward higher altitude. In addition, based on a linear regression equation established between average occurrence date of a phenophase and annual mean temperature at all stations, possible effects of a temperature rise of 0.5°C, 1.0°C, 1.5°C, 2.0°C, and the CO2 doubling scenario on plant phenology were extrapolated. For example, the corresponding phenophases may advance four to six days in spring and summer, and delay four to six days in autumn under the scenario of double CO2 content (Zhang 1995). 1.3.5

Teaching phenology at universities

The best way to spread phenological knowledge and methodology, and promote phenological research is to teach phenology at universities. Guodong Yang and Xiaoqiu Chen made an early attempt. In September 1985, they began to give a course entitled “Phenology” for senior students of the Department of Geography at Beijing Teacher’s College. During 1985 and 2002 about 500 students selected this course. Some of them participated in phenological observation and research on phenological calendars in the Beijing area. Since spring 1997, Xiaoqiu Chen has been teaching “Ecological Phenology” for graduate students in fields of Physical Geography, Ecology and Environmental Science at Peking University. In this two-credit course (36 hours per semester, from March to July), the following knowledge and techniques on phenology have been introduced: – what is phenology about? – appearance and development of phenology in China and on the world, – method of phenological observation, – natural laws of phenological event occurrence, – phenological calendars, – phenological seasons and phenological growing seasons,

20

Phenology: An Integrative Environmental Science

– – – –

causes of spatial and temporal differences of phenological events, phenological modeling and prediction, remote sensing of phenophases and phenology and global climate change. About 50 graduate students have undertaken this elective course during the last six years. The teaching process has raised their curiosity in relation to natural seasonality and enhanced their ability to observe nature. Several students have also chosen phenological themes as the basis for their master’s theses.

2.

PHENOLOGICAL OBSERVATION AND RESEARCH IN JAPAN

2.1

Networks and Data

In 1953, the Japanese Meteorological Agency (JMA) established a national phenological observation network consisting of approximately 100 stations. The aims were to monitor local climate via phenological phenomena of some specific plants and fauna. The Observation Division at the Headquarters of the Japanese Meteorological Agency (JMA) in Tokyo is responsible for phenological observations in Japan. The observation program consists of 12 species of plants and 11 species of fauna, and related phenophases (Table 4), whereas the observation criteria were defined in “The Manual of Phenological Observation” (JMA 1953). Phenological data are published monthly in the “Geophysical Review” under categories of “Agrometeorological Summary” or “Applied Meteorology”. The phenological network in Japan is now encountering difficulties in continuing reliable observations because of the effects of urbanization, and the function of phenological observation in monitoring local climate has been weakened with the modernization of surface weather observation network (from personal correspondence with Dr. Mitsuhiko Hatori, Director of Observations Division, Observations Department, JMA).

2.2 Research and Applications The research undertaken on phenology in Japan focuses mainly on phenological models and their applications in predicting phenophases, estimating effects of urban warming on plant phenology, reconstructing historical climate using phenological data, and detecting the influence of climate change on plant phenology. The study of spatial distribution models

Chapter 2.1: East Asia

21

of phenophases in Japan can be traced to the 1940s. Nakahara (1948) established the relationship between mean flowering date (y ( ) of Prunus yedoensis and geographical coordinates of phenological stations. Park-Ono et al. (1993) modified this model using phenological data of 37 years (19531989) as follows:

y = 92.56 + 4.77(ϕ − 35° ) − 0.59(λ − 135° ) + 1.28h

(3)

On the basis of this kind of model, isolines of Prunus yedoensis mean flowering dates were drawn and published in an atlas (Momose 1974). Since Prunus yedoensis is the national flower of Japan, the Japanese Meteorological Agency officially issues the predicted flowering date each year for sightseers. Table 2.1-2. Phenological observation program in Japan. Plants B* First-F* Full-F* Color changing Prunus yedoensis Matsum. x x Prunus mume Sieb. et Zucc. x Camellia japonica L. x Taraxacum (T. platycarpum x Dahlst., T. albidum Dahlst., and T. japonicum Koidz.) Rhododendron kaempferi x Planch. Wisteria floribunda DC. x Lespedeza bicolor Turcz. x Var. japonica Nakai x Hydrangea macrophylla Seringe var. otaksa Makino Lagerstroemia indica L. x Miscanthus sinensis Anderss. x x x Ginkgo biloba L. Acer palmatum Thunb. x *

B: budding, First-F: first flowering, Full-F: full flowering. Fauna First voicing

Alauda arvensis Linnaeus Cettia diphone (Kittlitz) Lanius bucephalus Temminck et Schlegel

x x x

First seeing

Leaf falling

x x

Last seeing

22

Phenology: An Integrative Environmental Science

Fauna Graptopsaltria nigrofuscata Motschulsky Tanna japonensis Distant Hirundo rustica Linnaeus Pieris rapae crucivora Boisduval Papilio machaon hippocrates C. et R. Felder Orthetrum albistylum speciosum Uhler Lampyridae (Luciola ( cruciata Motschulsky and L. lateralis Motschulsky) Rana nigromaculata Hallowell

First voicing x

First seeing

Last seeing

x x x x

x

x x

x

Another widely applied model was established based on the concept of “the number of days transformed to standard temperature” (DTS), by which the relationship between flowering date of Prunus yedoensis and environmental factors was described as follows (Aono and Omoto, 1990):

D* = 110.05 + 1.763ϕ − 0.062r + 0.403CI − 0.234WI + 0.011WI 2 (4) where ϕ is latitude, r is distance from coast (km), CII is coldness index (monthly mean air temperature below 5°C; °C month), and WII is warmth index (monthly mean air temperature above 5°C; °C month). Using the DTS method, Omoto and Aono (1990/1991) estimated the shift of flowering dates of Prunus yedoensis attributable to urban warming in Japan, and found that the amount of shift in the flowering date was almost proportional to the amount of urban warming at each site. In addition, it was also shown that the proportional constants were closely related to mean flowering dates at each site. Trends in the advancement of flowering dates due to an increase in temperature because of urban warming were observed in large cities of Japan. The estimated effect on flowering dates ranged from about 2.5 to 5.5 days/°C with smaller values in southern and larger values in northern Japan. From the temporal perspective, a relatively rapid increase to earlier flowering dates was detected from the 1950s to mid-1970s, and then the increasing tendency stopped. A reversed procedure of DTS method was used to deduce the March mean temperature based on full flowering dates of cherry trees (Prunus ( jamasakura) recorded in chronicles and ancient diaries in Kyoto since the

Chapter 2.1: East Asia

23

14th century. The error of calculated March decadal mean temperatures was less than 0.5°C during the 20th century. The results showed that there were relatively cold periods around the mid-fourteenth and early sixteenth centuries, as well as from the late seventeenth to early nineteenth century. Warm springs appeared during the early seventeenth century. Since the early nineteenth century, a steady warming trend has been detected, which may have been induced by both urban warming and larger scale warming (Aono and Omoto 1993). In order to estimate possible effects of climate change on plant phenology, Park-Ono, et al. (1993) established linear regression equations between flowering dates of Prunus yedoensis and monthly mean air temperature for different stations in Japan and South Korea. They showed that flowering dates of Prunus yedoensis were significantly influenced by the monthly mean temperatures of January (winter), March and April (spring). For example, if the March mean temperature increased by 1°C, the flowering date of Prunus yedoensis would advance by 3-4 days. Kai et al. (1993) used simple and multiple regression analyses between spring and autumn phenophases, and meteorological factors (monthly mean temperatures, warming indices and cold indices) to examine the potential effects of global warming on plant phenology at 102 stations in Japan from 1953 to 1990. Results showed that if the monthly mean temperature increased by 1°C, the flowering dates of Prunus yedoensis and Prunus mume would advance by 2.7-4.8 days and by 4-13 days, respectively. In autumn, the leaf color changing dates of Ginkgo biloba and Acer palmatum would delay by two to seven days if the monthly mean temperature increased by 1°C. There were significant spatial differences in the effects of temperature increase on plant phenology. In addition, they also presented the predicted shift in the distribution patterns of flowering dates of Prunus yedoensis and Prunus mume under a monthly mean temperature rise of 3°C in Japan. In addition to the above studies, Shigehara et al. (1991) compared the linear trend of the first flowering date of Prunus yedoensis with that of average temperature in February and March for the 12 climatic regions of Japan, and in March and April for the northern regions (“eastern and western Hokkaido”) directly. The study period was from 1953 to 1990 for most regions. The first flowering date showed a significant positive (delayed) linear trend with a slope of 0.6-1.8 days per 10 years in nine regions from 1953 to 1987, whereas the mean temperature indicated a significant negative linear trend (decreased) with a slope of 0.16-0.4°C per 10 years in the nine regions and during the same period. On average, if the mean temperature decreased by 1°C, the first flowering date would be delayed by four days. Generally speaking, northern regions showed a larger impact (5 days/°C) than southern regions (3 days/°C). However, from 1988 to 1990, the

24

Phenology: An Integrative Environmental Science

flowering date tended to be earlier; this coincided with an increase in the average temperature. In 2001, the Ministry of Environment of Japan published a report addressing the effects of global warming on Japan, in which influences of climate change on plant and animal phenology were summarized (Ministry of Environment 2001).

REFERENCES CITED Aono, Y., and Y. Omoto, Estimation of blooming date forr Prunus yedoensis using DTS combined with chill-unit accumulations, J. Agricult. Meteorol., 45, 243-249, 1990. Aono, Y., and Y. Omoto, Variation in the March mean temperature deduced from cherry blossom in Kyoto since the 14thh century, J. Agricult. Meteorol., 48, 635-638, 1993. Chen, X., A study of phenological seasonal rhythm in spring and summer of the North China region (in Chinese), Scientia Geographica Sinica, 10(1), 69-76, 1990. Chen, X., and G. Yang, A study on regional forecast model of tree’s phenology, a case of the North China region (in Chinese), Agricultural Meteorology, 9(3), 42-44, 1988. Cheng, C., X. Feng, L. Gao, and G. Shen (Editors), Climate and Agriculture in China, China Meteorological Press, Beijing, 519 pp., 1993. Gong, G., and W. Jian, On the geographical distribution of phenodate in China, Acta Geographica Sinica, 38 (1), 33-40, 1983. Gong, G., P. Zhang, and J. Zhang, Changes in natural phenological periods of Beijing area, in Environmental Changes Studies, Vol. 1 (in Chinese), edited by R. Hou, J. Xing, T. Su, pp. 64-75, Ocean Press, Beijing, 1984. Institute of Geography at the Chinese Academy of Sciences, Yearbook of Chinese Animal and Plant Phenological Observation No. 1 (in Chinese), Science Press, Beijing, 122 pp., 1965. Japanese Meteorological Agency, The Manual of Phenological Observation (in Japanese), Tokyo, 88 pp., 1953. Japanese Agrometeorological Society, Fundamental Agrometeorology, Chinese translation by H. Hou, Science Press, Beijing, 426 pp., 1963. Kai, K., M. Kainuma, N. Murakoshi, and K. Omasa, Potential effects on the phenological observation of plants by global warming in Japan, J. Agricult. Meteorol., 48 (5), 771-774, 1993. Ministry of Environment, Investigation Committee of Global Warming Problems, Impact of Global Warming on Japan 2001 (in Japanese), pp. 403-434, March, 2001. Momose, N., Atlas of Animal and Plant Phenology of Japan (in Japanese), Marunouchi Shuppan Book Co., Tokyo, 180 pp., 1974. Nakahara, M., Phenology (in Japanese), Kawadesyobo Press, Tokyo, 209 pp., 1948. National Meteorological Administration, Agrometeorological Observation Criterion Vol. 1 (in Chinese), Meteorological Press, Beijing, 212 pp., 1993. Omoto, Y., and T. Aono, Effect of urban warming on blooming of Prunus yedoensis, Energy and Buildings, 15-16, 205-212, 1990/1991. Park-Ono, H. S., T. Kawamura, and M. Yoshino, Relationships between flowering date of cherry blossom (Prunus yedoensis) and air temperature in East Asia, Proceedings of the 13th International Congress of Biometeorology 1993 September 12-18, pp. 207-220, Calgary, 1993.

Chapter 2.1: East Asia

25

Reader, R., J. S. Radford, and H. Lieth, Modeling important phytophenological events in Eastern North America, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 329-342, Springer-Verlag, New York, 1974. Schnelle, F., Pflanzen-Phänologie, Akademische Verlagsgesellschaft, Geest and Portig, Leipzig, 299 pp., 1955. Shigehara, K., T. Okamura, N. Nakayama, and F. Watanabe, Phenological observation data in Japan to be utilized as an indicator of climatic variation, Proceedings of the International Conference on Climatic Impacts on the Environment and Society 1991 January 27February 1, pp. C1-C6, Ibaraki, Japan, 1991. Wan, M. (Ed.), Natural Calendar of China I (in Chinese), Science Press, Beijing, 421 pp., 1986. Wan, M. (Ed.), Natural Calendar of China II (in Chinese), Science Press, Beijing, 437 pp., 1987. Wan, M., and X. Liu, Method of Chinese Phenological Observation (in Chinese), Science Press, Beijing, 136 pp., 1979. Yang, G., and X. Chen, Correlation and regression analysis between tree’s phenophases (in Chinese), Agricultural Meteorology, 6(1), 49-52, 1985. Yang, G., and X. Chen, Phenological Calendars and their Applications in the Beijing area (in Chinese), Capital Normal University Press, Beijing, 309 pp., 1995. Zhang, F., Effects of global warming on plant phenological events in China (in Chinese), Acta Geographica Sinica, 50(5), 402-410, 1995. Zhu, K. New monthly calendar (in Chinese), Bulletin of Chinese Meteorological Society, 6, 114, 1931. Zhu, K. A preliminary study on the climate fluctuation during the last 5000 years in China (in Chinese), Scientia Sinica, 16, 226-256, 1973 Zhu, K., and M. Wan, Phenology (in Chinese), Science Press, Beijing, 131 pp., 1973.

Chapter 2.2 AUSTRALIA Marie R. Keatley1 and Tim D. Fletcherr2 1

School of Resource Management, University of Melbourne, Creswick, Victoria, Australia; Department of Civil Engineering, Monash University, Clayton, Victoria, Australia

2

Key words:

Australia, History, Seasonality, Community networks, Observation

1.

HISTORICAL CONTEXT

1.1

Aboriginal Understanding of Seasonality

Aboriginals have occupied the Australian continent for at least 50,000 years (Head 1993). Their culture had therefore, prior to European settlement in 1788, developed a deep understanding of the interrelationships between the environment and its influence on the seasonality of fauna and flora (Davis 1989). This is partly because their survival depended heavily on their understanding of phenology or seasonality of food resources (Reid 1995). Aboriginal calendars recognize between 5 and 10 seasons, varying in length from 2 weeks to 4 months (Davis 1989; Baker 1993; Jones et al. 1997). Each season is defined by the changes in flora and fauna as well as the strength of wind, amount of rain and temperature (Davis 1989; Baker 1993; Jones et al. 1997). An example of this is from the Yolngu people in northeast Arnhem land, whose calendar contains 6 major seasons. The commencement of Midawarr (the fruiting season) is signaled by a wind change from the northwest to the east. It lasts 8 weeks (March to April) and includes the harvest period (Ngathangamakulingamirri, Davis 1989). The calendars that remain “intact” are mainly associated with communities from Northern Australia (e.g., Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 27-44 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

28

Phenology: An Integrative Environmental Science

Aurukun, Alngith, and Bunitj); those in southern states have been lost (Reid and Beckett 1995).

1.2

Early European Phenological Observation

It is not known when the systematic recording of phenological data by Europeans commenced. Table 1 covers 98 years (1856-1954) of Australian phenological history and focuses on plants, reflecting the extant records. Table 2.2-1. Early phenological studies undertaken in Australia. Date Comments 1856 - ?

1857 - 1895 (?) 1858 - Dec 1885

1886-1887 1891 1895

1905, 1908

1907, 1925, 1954

1909 - 1921

1909, 1911, 1924

Phenological recordings undertaken by von Mueller, Vic. Government botanist Irregular collection of flowering observations Leafing, flowering, and fruiting recorded in the Royal Society of Tasmania's Garden Monthly listing of plants in flower around Sydney Call for the establishment of phenological network in Victoria Flowering phenology of orchids, methodology outlined for recording phenological observations New South Wales Undersecretary for Lands requests that foresters record flowering within their district Meteorological observers requested to undertake phenological observations The broad flowering times of honey flora published in Apiarists' journals Book published on general flora and fauna observations taken throughout the year

Reference (Prince 1891)

(Maplestone 1895a) (Royal Society of Tasmania 1858-c1883, 1860, 1896) (Haviland 1886a, b, 1887a, b, c, d, e, 1888) (Prince 1891) (French 1895; Maplestone 1895a, b) (Maiden 1910)

(Commonwealth Meteorology 1907; Bureau of Meteorology 1925a, b; 1954) (Penglase and Armour 1909; Beuhne 1914; McLachlan 1921) (Mack 1909, 1911, 1924)

Chapter 2.2: Australia Date 1910, 1922

1929-1949 1925 - 1981

1949

1949 - 1954 (?)

Comments "A plea for the study of phenological phenomena in Australia" by New South Wales government botanist Flowering dates of 7 orchid species at 3 locations in Western Australia Monitoring undertaken by Forest Commission of the various states

Another call for phenological studies to be undertaken "Meteorological Service" established a program for phenological observations Australia-wide ornithological program

29 Reference (Maiden 1910, 1922)

(Erickson 1950) (Steane 1931; Tout 1935; Loneragan 1979; Dale and Hawkins 1983; Keatley et al. 1999) (Gentilli 1949) (Wang 1967)

(Anon. 1949; Jarman 1950)

The earliest cited phenological recording is the work of von Mueller, which commenced in 1856 (Prince 1891)—21 years after the official European settlement of Victoria (Carron 1985). Baron von Mueller was, at this time, the government botanist and from 1857 superintendent of the Royal Botanic Gardens (Hall 1978). Prince (1891) indicates that von Mueller ceased recordings sometime prior to 1891 because of work commitments. In his article Prince (1891) not only outlines the usefulness of phenology to ornithology, botany, agriculture and human health, but also calls for a phenological network to be established. He also notes that 20 years or more are required for phenological observations to be of any scientific merit, the same standard that is recommended today (Menzel et al. 2001). Other early observations were primarily undertaken by individuals (e.g., Haviland 1886 - 1888, Maplestone 1857-1895) who were associated with scientific organizations. It was not until the late part of the eighteenth century that the general public in Australia was encouraged to be involved in science (Newell and Sutherland 1997). Haviland’s recordings (Haviland 1886a, b, 1887a, b, c, d, e, 1888) are a list of what species were in flower in each month around Sydney. While such a list without specific dates of flowering is now considered limited, it is highly likely that the range and months of flowering of many of the species were unknown at the time. The longer set of observations (1857-1895) undertaken by Maplestone are only referred to in his 1895 article, and the original data are probably lost. The Royal Society of Tasmania (the first such organization in Australia, Baker and Rae 2000) appears to have instigated the earliest set of

30

Phenology: An Integrative Environmental Science

phenological (1858 - December 1885?) records in Australia by an organization. The Society was granted “a large portion of the Government garden” in 1843 which was maintained by the society until December 31st, 1885 when ownership was returned to the State of Tasmania (Royal Society of Tasmania 1896). From 1858 (Royal Society of Tasmania 1858-c1883) records of which plants were in flower for each month were kept. In 1860, this practice was formalized when regulations for the garden were published which required, in part: “a nominal return of plants which have flowered and of fruits in season at the Gardens during each month, shall in like manner be furnished periodically by the Superintendent” (Royal Society of Tasmania 1860). From January 1864 until December 1885, observations from a list of “standard” plants (mainly exotics) were recorded in the Proceedings of the Royal Society (Royal Society of Tasmania 1865). It seems that phenological recordings may have ceased by 1900 (Morton 1901). Between 1905 and 1922, Maiden, the New South Wales government botanist and director of the Sydney Botanical Gardens (1896-1924) (Hall 1978) made requests of the Bureau of Meteorology and then Forests Department of New South Wales (through the Undersecretary for Lands), to undertake phenological observations (Maiden 1910; 1922). The Bureau of Meteorology had recognized the value of undertaking phenological observations and requests were made of rainfall and meteorological observers to record the first flowering of native plants, the arrival of migratory birds and butterflies in their weather notes (Bureau of Meteorology 1925a, b, 1954; Commonwealth Meteorology 1907). The Bureau did not, however, insist on these being taken: “from occasional remarks furnished with returns, the recording of fundamental climatological data becomes at times irksome, and a tax upon the time of our worthy settlers; I have (H.A. Hunt, Commonwealth Meteorologist), therefore, hesitated to press for phenological observations” (Maiden 1910) There is a specific mention of a phenological network being established in 1949 by the “Meteorological Service” (Wang 1967). Unfortunately, extensive research has not found evidence of the network that Wang (1967) was referring to. If it was the Bureau of Meteorology that was responsible for these observations it is possible that the data were destroyed in the 1960s (A. Brewster, Data Services, National Climate Centre, Bureau of Meteorology, Victoria personal communication, 2002), a not uncommon practice (Keatley et al. 1999; Sparks 1999).

Chapter 2.2: Australia

31

In 1948, the Annual congress of the Royal Australian Ornithologists’ Union (now Birds Australia) adopted a proposal for the establishment of an Australian wide community network of observers to provide data on distribution, bird movements and the influence of climate on such. It was to be organized through the very popular Wild Life journal (Sedgwick 1949). The program had commenced by 1949 with 200 observers. It seems that Tasmania and South Australia were not included as they already had similar networks operating (Jarman 1950). This scheme probably ceased operation in 1954, when Wild Life was no longer published (Mulligan and Hill 2001).

2.

AGENCY DATA AND RESEARCH

Figure 1 summarizes forest agency phenological records known in five of the Australian states (New South Wales, Queensland, Tasmania, Victoria and Western Australia). These records cover various durations of flowering and/or budding of forest commercial and/or “honey” trees. The primary aim seems to have been the enumeration of seed crop for silvicultural

Figure 2.2-1. Overview of Australian Forests Commissions' phenological studies (1925 1981).

32

Phenology: An Integrative Environmental Science

management. The surviving length of observations varies from, as yet unknown in New South Wales and Tasmania to 51 years in Victoria.

2.1

New South Wales

The date of commencement and cessation of long-term phenological recordings in New South Wales undertaken by the Forestry Commission is unknown. Correspondence between the Secretary of the Forestry Commission in New South Wales and his Victorian counterpart indicate that “for some years” prior to 1935 the Commission in New South Wales had been “obtaining sylvicultural (sic) notes …on the flowering and fruiting of various species” on a monthly basis (Tout 1935). This practice seems to have continued until at the least the 1950s (Neil Humphreys, former Forester, Forestry Commission of New South Wales, personal communication, 2002). Two research notes; Number 10 (Floyd 1962) and Number 19 (Van Loon 1966), indicate that specific phenological research was undertaken on the seedfall and flowering of Eucalyptus pilularis and E. microcorys between October in the late 1950s and early 1960s. More recently (Law et al. 2000), results of a ten-year study (1982-1992) on the flowering phenology of 20 myrtaceous species phenological studies have been published.

2.2

Queensland

Queensland’s then Forestry Department established flowering and fruiting “experiments” in the early 1930s, which appeared to have continued until 1958 (Dale and Hawkins 1983). Surviving summary sheets (in possession of the authors) indicate that these experiments were, in the main, observations that focused on quantifying the duration, intensity and time of year of budding, flowering, capsule development and seedfall. These sheets indicate observations were carried out in 19 locations throughout the State on 12 Myrtaceous species, as well as Araucaria cumminghamii and Pinus spp. (R. Lott, Scientist, Queensland Forestry Research Institute, Gympie, Queensland, personal communication, 2002). There is also the indication that these observations were written up as internal Queensland Forest Service unpublished reports (Kluver 1942 as noted in Hawkins 1959). Some quantitative studies were undertaken during this period and again in 1966 to 1973 (Dale and Hawkins 1983) examining the reproductive ecology of Corymbia maculata (Spotted gum). In recent times (post 1986) similar species (Araucaria ( cumminghamii, Corymbia maculata, Eucalyptus cloeziana) have been examined but for shorter periods (less than 5 years). Between 1995 and 1997, pollination,

Chapter 2.2: Australia

33

summer flowering abundance and annual fruit crop size were recorded for Flindersia brayleyana on the Atherton Tableland. The flowering abundance and fruit crop size measures in small remnants and larger tracts of native forest are also being compared to determine if there is a difference attributable to forest fragmentation (R. Lott, Scientist, Queensland Forestry Research Institute, Gympie, Queensland, personal communication, 2002).

2.3

Victoria

In 1930, Officers-in-Charge of forest districts were instructed to establish flowering and seeding plots for each of the commercial Eucalypt species in their district, to monitor the time and pattern of flowering, the duration between flowering and the availability of mature seed (Strahan 1930). Within five years the focus had changed from seeding and flowering alone, to budding, flowering and potential honey production. These were codified into “Quarterly Budding and Flowering Reports” (Keatley et al. 1999). The observations were continued until 1981. Table 2 outlines the surviving phenological data of greater than 20 years in length. In total, 92 locations and 51 species are known and indicate what resource might have been available in other states. Table 2.2-2. Surviving Victorian Forests Commission records of greater than 20 years duration. District Length of Record Barmah 30 Beaufort Broadford Macallister/Heyfield Maryborough Rushworth Yarram

July 1933 – December 1974 December 1934 – September 1975 January 1948 – September 1975 October 1949 – May 1975 January 1940 – December 1978 April 1945 – December 1964 July 1943 – December 1973

Quantitative phenological studies (i.e., traps are laid out in the forest to collect floral components) have also been undertaken by the Forest Commission (e.g., Dexter 1968) and its successors (e.g., Bassett 1995), although usually limited to one or two locations. Table 3 outlines the species studied to date and shows that the longest duration has been six years.

2.4

Tasmania and Western Australia

There is little known about phenological observations undertaken by the Tasmanian Forestry Department and the Forests Department of Western

34

Phenology: An Integrative Environmental Science

Australia. In Tasmania, observation of seeding of five Eucalypt species commenced in 1929 in the North East and North West districts of the state (Steane 1931). The cessation dates of observations are unknown. With the exception of a seven-year study (1989 - 1996) on the seeding of Eucalyptus obliqua (Neyland et al., in press), there does not appear to be any systematic studies undertaken (M. Neyland, Native Forest Research Officer, Forestry Tasmania, personal communication, 2002). Table 2.2-3. Quantitative phenological studies undertaken by the Victorian Forest Commission and its successors. Species Duration References E. camaldulensis E. denticulata

Sept. 1961 - Apr. 1975 Mar. 1975 - Apr. 1980, 1994 - 2000, Nov. 1999 - May 2001

E. fastigata

Apr. 1975 - Apr. 1980

E. regnans

1986 - 1989, 1994 - 2000

E. globoidea E. sieberi E. tricarpa E. cypellocarpa E. obliqua

Feb. 1989 - Nov. 1992 Feb. 1989 - Nov. 1992 Aug. 1993 - July 1997 Sept. 1994 - Jan 1998 Sept. 1994 - Jan 1998 1997 - ongoing

E. nitens

1994 - 2000

E. viminalis

Dec. 2000 - ongoing

E. delegatensis

2001 - ongoing

(Dexter 1968) M. Murray, Officer in Charge, Eastern Research Centre, Dept. of Natural Resources and Environment, personal communication, 2002 M. Murray, personal communication, 2002 (Squire 1990), B. Roberts, Seed Management Officer, Dept. of Natural Resources and Environment, personal communication, 2002 (Bassett 1995) (Bassett 1995) (Keatley and Murray 2001) (Murray and Lutze 2000) (Murray and Lutze 2000) B. Roberts, personal communication, 2002 B. Roberts, personal communication, 2002 M. Murray, personal communication, 2002 B. Roberts, personal communication, 2002

One study (Loneragan 1979) indicates that the Forests Department of Western Australia undertook early phenological studies. The data in this study come from the West Pemberton forest district and cover the

Chapter 2.2: Australia

35

development, timing and pattern of the buds, flowers and seed fall of Eucalyptus diversicolorr from 1925 until 1971.

3.

COMMUNITY-BASED PHENOLOGICAL NETWORKS

The two most notable community-based networks in Australia are “Timelines Australia Project” (Timelines) and Faunawatch. Each network, voluntarily coordinated, aims to engage the general public in monitoring of flora and fauna. While neither of these organizations regard phenology as their primary focus, it does form a substantial part of their work.

3.1

Timelines

Timelines’ national program was launched in 1997, although individual programs have operated at a local level from 1994 (Jameson 2001; McDonald 2002). It is sponsored by The Gould League of Victoria, an environmental education organization, and coordinated by Alan Reid1 (Reid 2001). The main philosophy behind Timelines is that European seasons are inappropriate for Australia (Reid 1994; Reid and Beckett 1995). It aims to develop, through research, appropriate Australian seasonal calendars (as shown in Figure 2) similar to the aboriginal calendars of Northern Australia (Jameson 2001). The project also aims to: – recover information held by naturalists in their diaries, notebooks or information from library files, – analyze the information to identify environmental change, either seasonal, successional, or changes in populations, – promote the establishment of monitoring programs across Australia which data may then assist in environmental management and planning (Reid and Beckett 1995). To meet these aims The Gould League has published a recording diary called “Banksias and Bilbies” (Reid and Beckett 1995), and a CD called “Timelines” (Gould League of Victoria 1998). Participants in the project are encouraged to record anything of interest to them as long as the reason for recording the data is also listed either within the diary or on the database. Hence, people may concentrate on birds, insects, flowers or any one particular species of these. The database also asks for the month, species, activity (e.g. preening), number and location. Members of Timelimes, either individuals or groups, have also received a bimonthly newsletter since 1998 (Reid 2001), which covers a wide range of topics (e.g., a scoring scheme for

36

Phenology: An Integrative Environmental Science

fauna quality, climate projections and seasons, as well as news from other Timelines groups).

Figure 2.2-2. Calendar originally developed for Middle Yarra. Dot indicates locality (reproduced with the kind permission of The Gould League and Alan Reid).

There are approximately 50 active groups and individuals currently involved in Timelines. Three groups—Middle Yarra Timelines project, Timelines Hunter and Nature Watch—appear to be the most active. Both Timelines Hunter and Nature Watch are supported by local government. Middle Yarra Timelines has been supported by the Gould League, The Field Naturalist Club of Victoria and Parks Victoria (Jameson 1996). This group has delineated a six-season calendar—Early spring, True Spring, High summer, Late summer, Early winter and Late winter (Jameson 2001). Timelines Hunter has its annual diary “Nature Watch Diary” printed and distributed free-of-charge by the local government. Within this group there is a core of 24 dedicated people (although in 2002 130 diaries were distributed (K. McDonald, coordinator, Timeslines Hunter, personal communication, 2002)). Nature Watch Newcastle uses its diary as part of a multi-media program to engage the general public to “look and talk” about nature; phenological monitoring is not the primary aim. In addition, a website, monthly electronic bulletin, as well as a monthly broadcast on the national broadcaster's regional network (Radio 1233 ABC), form part of the program. An

Chapter 2.2: Australia

37

environmental education officer is also employed as by Nature Watch (Fran Beilby, Newcastle City Council, personal communication, 2002). The Timelines Australia Project itself has had a long development phase, with similar versions of the recording diary published between 1979 and 1989 (Gould League of Victoria 1979; Mason 1989; Reid and Beckett 1984). “Gum Leaves and Geckos” (Reid and Beckett 1984) inspired the “Nature Watch Observation Diary” which was used by geographically isolated children in years 4 to 7 attending the School of the Air in the Northern Territory. Observations formed the basis for a monthly discussion on environmental issues during radio class lessons (Sim 1987).

3.2

Faunawatch

Faunawatch is a volunteer fauna monitoring project, covering the Sunshine Coast in Queensland (Jameson 2001). Observations are collected on birds, butterflies/moths, mammals, fish, reptiles, insects and spiders. The majority of observations, however, are concentrated on birds (87.7%), covering 320 species, and butterflies and moths (8.4%), with 106 species recorded (Hickman 2002). Data collected from this program are fed into the Queensland government’s “Wildnet” and Birds Australia’s Atlas. Initially inspired by the Timelines project (Jameson 2001), this community-based program commenced in 1998 with funding from the Federal government, and was known as “The Rhythms of Life” until September 2002 (K. Hickman, Faunawatch, volunteer coordinator, personal communication, 2002). It has nine aims that include raising public awareness and knowledge of wildlife in the area, and the creation of a comprehensive database of fauna occurring on the Sunshine Coast. Timelines and Faunawatch have demonstrated that there are many “closet phenologists” (a term first used by Tim Sparks, Centre for Ecology and Hydrology, Monks Wood, United Kingdom) within Australia. These programs have also helped to demonstrate and promote the value of phenological data to the community.

3.3

Other Organizations

3.3.1

Globe

Australia has been involved with GLOBE (Global Learning and Observations to Benefit the Environment) since April 1995, coordinated through the Federal Department, Environment Australia. Teachers were initially trained in phenological methods, t but difficulties were encountered entering southern hemisphere data into a database designed for the northern hemisphere. There was also a problem in identification of native species.

38

Phenology: An Integrative Environmental Science

Hence, now only a few schools undertake measurements for their own research use (J. Stuart, Environmental Education Unit, Environment Australia, personal communication, 2001). 3.3.2

Birds Australia

Ornithology is one science where phenological data may play a significant role (Ahas 1999; Sparks 1999). Within Australia, Birds Australia (formerly the Royal Australiasian Ornithologists’ Union) is the national body (non-government) whose history (1901 to present) is closely aligned with the history of the discipline over the last hundred years (Robin 2001). The electronic databases of Birds Australia contain phenological information such as movements and breeding, with a temporal time scale of as little as one year (e.g., Birds in Tree Hollows) to 38 years and continuing (Nest Record Scheme, Dr. M. Weston, Manager, Research and Conservation Department, Birds Australia, personal communication, 2002). The Birds Australia archives are held at the State Library of Victoria, and contain many bird lists. They have not, however, been examined in much detail and are not scheduled to be sorted and catalogued until 2032! In addition, some of the earlier longer-term observation schemes were in essence just informal networks (Dr Libby Robin, Research Fellow, Australian National University, Canberra, personal communication, 2002).

4.

RECENT PHENOLOGICAL RESEARCH

Table 4 summarizes recent (i.e., 1975-2002) phenological research reported in scholarly journals, which relate to Australia. A search of relevant databases (e.g., Biological abstracts) found 341 publications. This excludes, however, the period of the late fifties to middle sixties, when ecological studies on particular Eucalypt species were prevalent as thesis topics (e.g. Ashton 1956; Cunningham 1958; Dexter 1960; Florence 1961; Gill 1966; Loneragan 1961). Table 4 includes those papers (125 of 341) whose abstract described the length, discipline and geographical location of the research. While not exhaustive, this summary is representative of the research areas and duration of published studies undertaken in the last 27 years within Australia. The table shows that phenological research within Australia is centered on agriculture (49.2% of studies, including modeling). The majority of the studies (86.0%, excluding modeling), across all fields of research, span a period of five years or less. This is comparable to that found in an

Chapter 2.2: Australia

39

examination of field experiments published between 1977-1987 in Ecology g (Tilman 1989) where 93% were of durations of five years or less. Table 2.2-4. Recent phenological research in Australia. Study duration (years) Research Area <1 1-5 5-10 10-20 20-30 >30 Agriculture Ecology/phytophenology Forestry Entomology Horticulture Fisheries Human Health

26 4 2 4 3

Sub-total

39

Modeling Agriculture Ecology Entomology Horticulture Sub-total

10 1 6 1 18

23 21 1 3 3

2

3

1

1

1 1

2 2

1 2

1

53

4

0

0

4

0

3

0

Total 49 32 6 10 6 1 3

4

107

0

10 1 6 1 18

Studies extending beyond 20 years, the minimum time recommended for phenological studies (Menzel 2001), represent only 6.5%. This is partly attributable to the convention of what is considered long-term within a particular branch of science. It also reflects that funding for long-term research is restricted by funding organizations and community perception, as well as being determined by changing government policy and management (Keatley et al. 1999). The figure in Table 4 for ecological/ phytophenological studies greater than five years represents 2.2 percent. Again, this figure is similar to the Tilman (1989) paper that found less than 1.7 percent of the studies were greater than five years. The above only reinforces the need for public participation in phenological monitoring but coordinated by dedicated individual(s) (Strayer et al. 1986).

5.

FUTURE DEVELOPMENTS

Where to go from here? Phenological research has been and is currently being undertaken within Australia. Unfortunately, much of this has been short term and in the case of longer-term data, much of it has been lost. The value of phenology as a science that can assist sciences other than agriculture and forestry is yet to be realized. Climate change and the need to

40

Phenology: An Integrative Environmental Science

understand its impacts may change this. A country with the size and population of Australia cannot undertake phenological monitoring unless the general public is involved, as is occurring with programs such as Timelines and Faunawatch. The collection of data is a vital first step, but unless scientists are involved in the interpretation and analysis of these data, their full value will not be realized. There also needs to be a focus for those interested in phenology. A website, similar to the European Phenological Network website, currently under development by Macquarie University, University of Melbourne and the Royal Botanic Gardens Sydney, will hopefully provide this.

ACKNOWLEDGMENTS The information presented in this chapter has only been possible because of the assistance of the following people: Mr. Alan Reid, Timelines National Coordinator, Camberwell, Victoria for his enthusiasm for this work and also for permission, along with The Gould League to reproduce the seasonal calendar, Ms. Jane Catford, for undertaking the database search, Ms. Anne Brewster, Data Services, National Climate Centre, Bureau of Meteorology, Victoria, a Ms. Robyn Eastley, Archives Office of Tasmania, Mr. Keith Hickman, Faunawatch, Queensland, Mr. Andrew Hollis, National Meteorological Library, Victoria, Mr. Neil 'Curly' Humphreys, Forpac, Victoria, Mr. Glen Jameson, Middle Yarra Timelines Coordinator, Dr. Bradley Law, State Forests of New South Wales, Dr. Rosemary Lott, Queensland Forestry Research Institute, Mr. Kevin MacDonald, Timelines Hunter, Dr. Maureen Murray, Forest Science Centre, Dept. of Natural Resources and Environment, Victoria, Mr. Mark Neyland, Forestry Tasmania, Mr. Barry Roberts, Dept. of Natural Resources and Environment, Victoria, Dr. Libby Robin, Centre for Resource and Environmental Studies, Australian National University, Canberra, Ms. Julia Stuart, Environment Australia, Canberra, Ms. Emilia Ward, University of Tasmania Library and the Royal Society of Tasmania, Dr. Michael Weston, Birds Australia, Victoria.

NOTES 1

Alan Reid is one of Australia's foremost environmental educators, ornithologist and winner of many environmental awards, among them the Australian Natural History Medal (Houghton 1993).

Chapter 2.2: Australia

41

REFERENCES CITED Ahas, R., Long-term phyto-, ornitho- and ichthyophenological time-series analysis in Estonia, Int. J. Biometeorol., 42, 119-123, 1999. Anon., Are you a bird observer?, Wild Life, 11, 207, 208, 238, 1949. Ashton, D. H., Studies on the Autecology of Eucalyptus regnans, Ph.D. thesis, The University of Melbourne, Parkville, 1956. Baker, A. T., and I. D. Rae, More than bugs and stones: Chemistry in the Royal Society of New South Wales, Historical Records of Australian Science, 13, 117-130, 2000. Baker, R., Traditional Aboriginal land use in the Borroloola region, in Traditional Ecological Knowledge: Wisdom for Sustainable Development, edited by N. M. Williams and G. Baines, pp. 126-143, Centre for Resources and Environmental Studies, Canberra, 1993. Bassett, O. D., Development of seed crop in Eucalyptus sieberi L. Johnson and E. globoidea Blakely in a lowland sclerophyll forest of East Gippsland, d Dept. of Conservation and Natural Resources, Victoria, 38 pp., 1995. I 610-618, 1914. Beuhne, F. R., The Honey Flora of Victoria, J. Dept. Agric. Vic., XII, Bureau of Meteorology, Australian Meteorological Observer's Handbook, H. J. Green, Government Printer, Melbourne, 171 pp., 1925a. Bureau of Meteorology, Australian Rainfall Observer's Handbook, H. J. Green Government Printer, Melbourne, 48 pp., 1925b. Bureau of Meteorology, Australian Meteorological Observers' Handbook, Commonwealth of Australia, Melbourne, 148 pp., 1954. Carron, L. T., A History of Forestry in Australia, Australian National University Press, Canberra, 355 pp., 1985. Commonwealth Meteorology, Instructions to Country Observers, William Applegate Gullick, Government Printer, Sydney, 29 pp., 1907. Cunningham, T. M., The natural regeneration of Eucalyptus regnans in association with logging, Ph.D. thesis, University of Melbourne, Parkville, 1958. Dale, J. A., and P. J. Hawkins, Phenological studies of spotted gum in southern inland Queensland, Queensland Department of Forestry, Technical Paper No. 35, Brisbane, 10 pp., 1983. Davis, S., Man of All Seasons, Angus and Robertson, North Ryde, New South Wales, 82 pp., 1989. Dexter, B. D., Seed Supply and Field Germination in the Natural Regeneration of Eucalyptus sideroxylon A. Cunn., Fourth year B.For.Sci. thesis, University of Melbourne, 1960. Dexter, B. D., Flooding and regeneration of River Red Gum, Eucalyptus camaldulensis, Dehn, Forests Commission of Victoria, Melbourne, 35 pp., 1968. Erickson, R., Flowering dates of orchids, West. Aust. Nat., 2, 72, 1950. Florence, R. G., Studies in the ecology of blackbuttt (Eucalyptus pilularis Sm.), Ph.D. thesis, University of Sydney, Sydney, 1961. Floyd, A. G., Investigations into the natural regeneration of blackbutt - E. pilularis. Research Note No. 10, Forestry Commission of N.S.W, Sydney, 20 pp., 1962. French, F. J., Observations on the flowering times and habitats of some Victorian Orchids, Victorian Nat., 12, 31-34, 1895. Gentilli, J., Phenology-A new field for Australian Naturalists, West. Aust. Nat., 2, 15-20, 1949. Gill, A. M., The Ecology of Mixed species Forests of Eucalyptus in Central Victoria, Australia, Ph.D. thesis, University of Melbourne, 1966.

42

Phenology: An Integrative Environmental Science

Gould League of Victoria, Environmental Log Book, Gould League of Victoria, Prahran, 72 pp., 1979. Gould League of Victoria, Timelines, Viridans Biological Databases, Brighton East, Australia, (CD-ROM), 1998. Hall, N., Botanists of the eucalypts, CSIRO, Melbourne, 160 pp., 1978. Haviland, E., Flowering seasons of Australian plants No 1, Proc. Linn. Soc. N.S.W., 1, 1049, 1886a. Haviland, E., Flowering seasons of Australian plants No 2, Proc. Linn. Soc. N.S.W., 1, 11021103, 1886b. Haviland, E., Flowering seasons of Australian plants No 3, Proc. Linn. Soc. N.S.W., 2, 105106, 1887a. Haviland, E., Flowering seasons of Australian plants No 4, Proc. Linn. Soc. N.S.W., 2, 135136, 1887b. Haviland, E., Flowering seasons of Australian plants No 5, Proc. Linn. Soc. N.S.W., 2, 185186, 1887c. Haviland, E., Flowering seasons of Australian plants No 6, Proc. Linn. Soc. N.S.W., 2, 348, 1887d. Haviland, E., Flowering seasons of Australian plants No 7, Proc. Linn. Soc. N.S.W., 3, 565, 1887e. Haviland, E., Flowering seasons of Australian plants No 8, Proc. Linn. Soc. N.S.W., 3, 267268, 1888. Hawkins, P. J., Establishment of regeneration of narrow-leaf ironbark (Eucalyptus crebra) by top burning following logging, Research Note No. 7, Queensland Forest Service, Brisbane, 18 pp., 1959. Head, L., The value of long-term perspective: environmental history and traditional ecological knowledge, in Traditional Ecological Knowledge: Wisdom for Sustainable Development, edited by N.M. Williams and G. Baines, pp. 66-70, Centre for Resources and Environmental Studies, Canberra, 1993. Hickman, K., Rhythms of Life Update, in Timelines Australia Project News, vol. 6, pp. 5, 2002. Houghton, S., Australian Natural History Medal: Alan Reid, Victorian Nat., 110, 228-229, 1993. Jameson, G., Middle Yarra Timelines: High Summer, Victorian Nat., 113, 26-28, 1996. Jameson, G., Timelines Calendars: Entering the landscape, in Getting to the heart of it: Connecting people with heritage. The Ninth Annual Conference Interpretation Australia Association, edited by I. A. Association, pp. 110-114, Interpretation Australia Association, Alice Springs, 2001. Jarman, H. E. A., Proceedings of the Annual congress of the Royal Australian Ornithological Union, Emu, 49, 238, 1950. Jones, D., S. Mackay, and A.-M. Pisani, Patterns in the valley of the christmas bush: A seasonal calendar for the Upper Yarra Valley, Victorian Nat., 114, 246-249, 1997. Keatley, M. R., I. L. Hudson, and T. D. Fletcher, The use of long-term records for describing flowering behaviour: A case-study in Victorian Box-Ironbark Forests., in Australia's Everchanging Forests IV, V edited by J. Dargavel and B. Wasser, pp. 311-328, Australian University Press, Canberra, 1999. Keatley, M. R., and M. Murray, An examination of the reproductive phenology of Eucalyptus tricarpa (DRAFT), Forest Science Centre, Orbost, Victoria, 31 pp., 2001.

Chapter 2.2: Australia

43

Law, B., L. Mackowski, and T. Tweedie, Flowering phenology of myrtaceous trees and their relation to climate, environmental and disturbance variables in northern New South Wales, Austral Ecol., 25, 160-178, 2000. Loneragan, O. W., Jarrah (Eucalyptus marginata Sm.) and karri (Eucalyptus diversicolor F.v.M.) regeneration in south-west Western Australia, MSc. thesis, University of Western Australia, Nedlands, 1961. Loneragan, O. W., Karri (Eucalyptus diversicolor F. Muell.) phenological studies in relation to reforestation, Bulletin 90, Forest Department of Western Australia, Perth, 37 pp., 1979. Mack, A. E., A Bush Calendar, Angus and Robertson, Sydney, 110 pp., 1909. Mack, A. E., A Bush Calendar, Angus and Robertson, Sydney, 110 pp., 1911. Mack, A. E., A Bush Calendar, Cornstalk Publishing Company, Arnold Place, Sydney, 110 pp., 1924. Maiden, J. H., A plea for the study of phenological phenomena in Australia, Proc. R. Soc. N.S.W, W 157-170, 1910. Maiden, J. H., Phenology: A form of nature study with very practical applications, in The Forest Flora of New South Wales, vol. 7, pp. 166-191, John Spence, Acting Government Printer, Sydney, 1922. Maplestone, C., Calendars and the indexing of natural history observations, Victorian Nat., 12, 120-122, 1895a. Maplestone, C., Flowering times of Orchards, Victorian Nat., 12, 82-83, 1895b. Mason, P., Environmental Log Book: For Keeping Notes on Nature in Australia, Gould League of Victoria, Prahran, 82 pp., 1989. McDonald, K., Nature watching in the Seaham area, in Essays on Seaham, edited by C. Hunter, pp. 157-162, Seaham Public School P & C Association, Seaham, 2002. McLachlan, R. G., Victoria Valley experiences, The Victorian Bee Journal, 2, 64-65, 1921. Menzel, A., N. Estrella, and P. Fabian, Spatial and temporal variability of the phenological seasons in Germany from 1951 to 1996, Global Change Biol., 7, 657-666, 2001. Morton, A., Some account of the work and workers of the Tasmanian Society and the Royal Society of Tasmania, from the year 1840 to the close of 1900, Pap. & Proc. R. Soc. Tas., 1900-1901, 109-126, 1901. Mulligan, M. J., and S.B. Hill, Ecological pioneers: a social history of Australian ecological thought and action, Cambridge University Press, Cambridge, 338 pp., 2001. Murray, M., and M. Lutze, Case study of seedcrop development in Eucalyptus obliqua (Messmate) and Eucalyptus cypellocarpa (Mountain Grey Gum) in High elevation Mixed Species forests of East Gippslandd (DRAFT), Centre for Forest Tree Technology, Eastern Research Centre, Orbost, 30 pp., 2000. Newell, J., and D. Sutherland, Scientists and Colonists, Australasian Science, 18, 56, 1997. Neyland, M. G., L. G. Edwards, and N. J. Kelly, Seedfall of Eucalyptus obliqua at two sites within the Forestier silvicultural systems trial, Tasmania., Tasforests, 14, in press. Penglase, and J. Armour, Victorian Honeys and Where They Come From, The Federal Independent Beekeeper, 2-4, 1909. Prince, J. E., Phenology and rural biology, Victorian Nat., 8, 119-127, 1891. Reid, A. J., Exploring local seasonality, Victorian Nat., 111, 35-37, 1994. Reid, A. J., A plan for all seasons, Habitat Australia, April, 14-15, 1995. Reid, A. J., The timelines Australia project, Timelines Australia Project News, 4, supplement 4 pp., 2001. Reid, A. J., and A. Beckett, Gum leaves and Geckos, Gould League of Victoria, Prahran, 96 pp., 1984.

44

Phenology: An Integrative Environmental Science

Reid, A. J., and A. Beckett, Banksias and Bilbies: Seasons of Australia, Gould League of Victoria, Moorabbin, Australia, 81 pp., 1995. Robin, L., The Flight of the Emu: A hundred years of Australian Ornithology 1901 - 2001, Melbourne University Press, Carlton South, Victoria, 492 pp., 2001. Royal Society of Tasmania, RSA/E29 Work and notebook, Monthly list of plants in flower in the Botanic Gardens, in possession of Special and Rare Materials, University of Tasmania Library, Hobart, 30 pp., 1858-c1883. Royal Society of Tasmania, Bye-laws for the regulation and management of the gardens of the Royal Society of Tasmania, Royal Society of Tasmania, Printed at the Mercury Steam Press Office, Hobart, 6 pp., 1860. Royal Society of Tasmania, Time of leafing, flowering, and fruiting for a few standard plants, in the Royal Society's Garden for the month, Pap. & Proc. R. Soc. Tas., 1864, 5, 1865. Royal Society of Tasmania, The Royal Society of Tasmania, Papers and Proceedings of the Royal Society of Tasmania, 1894-95, 1-5, 1896. Sedgwick, E. H., Proceedings of the Annual congress of the Royal Australian Ornithological Union, Perth 1948, Emu, 48, 177-211, 1949. Sim, J., Nature Watch Observation Diary, Northern Territory, 10 pp., 1987. Sparks, T. H., Phenology and the changing pattern of bird migration in Britain, Int. J. Biometeorol., 42, 134-138, 1999. Squire, R. O., Report on the Progress of the Silvicultural Systems Project July 1986 - June 1989, Dept. of Conservation and Environment, East Melbourne, Victoria, 90 pp., 1990. Steane, S. W., Report of the Forestry Department for the year ended 30th June, 1930, Forestry Department, Hobart, 6 pp., 1931. Strahan, A., Circular 144, Forests Commission Victoria, Melbourne, Unpublished Correspondence in VPRS 11563/P/0001, Unit 000131, File FCV 35/3123 HONEY, Location L/AZ/068/01/08, Public Records Office, 1 pp., 1930. Strayer, D., J. F. Glitzenstein, C. G. Jones, G. E. Likens, M. J. McDonnell, G. G. Parker, and T. A. P. Steward, Long-term Ecological Studies: An Illustrated Account of their Design, Operation, and Importance to Ecology, Institute of Ecosystem Studies, The New York Botanical Garden, Mary Flagler Cary Arboretum, Millbrook, New York, 32 pp., 1986. Tilman, D., Ecological Experimentation: Strengths and Conceptual Problems, in Long-term Studies in Ecology: Approaches and Alternatives, edited by G. E. Likens, pp. 136-157, Springer-Verlag, New York, 1989. Tout, S. M., Enquires on method of collection of data in regard to flowering and fruiting of native trees from Forestry Commission of New South Wales, Forest Commission of Victoria, Melbourne, Unpublished Correspondence in VPRS 11563/P/0001, Unit 000131, File FCV 35/3123 HONEY, Location L/AZ/068/01/08, Public Records Office, 1 pp., 1935. Van Loon, A. P., Investigations in regenerating the Tallowwood - Blue Gum forest type, Research Note 19, Forestry Commission of N.S.W, Sydney, 24 pp., 1966. Wang, J. Y., Agricultural Meteorology, Agriculture Weather Information Service, San Jose, California, 693 pp., 1967.

Chapter 2.3 EUROPE Annette Menzel Department of Ecology, TU Munich, Freising, Germany

Key words:

1.

Europe, Networks, ICP Forests, International Phenological Gardens, National Weather Services

INTRODUCTION

While the longest written phenological record originates in Japan at the royal court of Kyoto (the beginning of cherry flowering since 705 AD), the most vital and broadest tradition of phenological monitoring is found in Europe (Menzel 2002). In many countries long-term data sets exist, thus Europe is a particularly suitable region for investigating phenological changes or providing “ground truth” to satellite data. However, phenological information exists in numerous countries at a local, regional, or national level with quite different histories and traditions of observation. Thus, we are far away from a homogenous plant phenological data set at a continental level necessary for the applications indicated above. Similarly, it is impossible to provide here a complete overview of European phenology. Current important national networks are compared to new schemes, such as Nature’s Calendar, and to international initiatives (International Phenological Gardens, ICP Forests). One of the oldest European phenological records is the famous Marsham family record in Norfolk 1736–1947 (Sparks and Carey 1995). Following Schnelle’s (1955) good historical overview, the first phenological network was then established by Linné (1750-1752) in Sweden. The first international (European) phenological network was run by the Societas Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 45-56 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

46

Phenology: An Integrative Environmental Science

Meteorologica Palatina at Mannheim (1781-1792), and a second famous one by Hoffmann and Ihne (1883-1941). Schnelle (1955) summarized the development of phenological observations in Germany as well as in other countries, such as (considering only European ones) Austria, Poland, Czech Republic and Slovakia, Russia, Finland, Sweden, Norway, United Kingdom, The Netherlands, Belgium, France, Switzerland, Spain, Italy, Greece and the former Yugoslav Republic through the middle of the last century. However, their subsequent history was very patchy.

2.

INTERNATIONAL NETWORKS

2.1

International Phenological Gardens

The International Phenological Gardens (IPG) is a unique phenological network in Europe, which was founded in 1957 by F. Schnelle and E. Volkert. Since then, it has been maintained on a voluntary basis, coordinated by a chairman (Chmielewski 1996). Manuals and annual observations have been published in the journal of the IPGs, the Arboreta Phaenologica. At present the network is coordinated by the Humboldt University Berlin (www.agrar.hu-berlin.de/pflanzenbau/agrarmet/ipg.html). The core idea of this network was to obtain comparable phenological data across Europe by observing genetically identical plants (clones), which permanently remained at one site. Thus, the records are not influenced by different genetic codes of the plants and the variability and potential inaccuracy of the observations is reduced compared to data from the national phenological networks (Baumgartner and Schnelle 1976). In 1959, the first IPG started its observations in Offenbach (near Frankfurt am Main), Germany. Subsequently, additional IPGs were established with vegetative propagated species of trees and shrubs at different sites across Europe. Today, about 50 IPGs record up to seven phases of 23 plants species (~ 50 clones). The network covers a large area from 42 to 69°N (Macedonia to Scandinavia) and from 10°W to 27°E (Ireland / north Portugal to Finland), comprising different climate regions in Europe. Recent studies comprehensively analyzing the data of the IPGs revealed a lengthening of the growing season across Europe, provided the necessary ground truth to satellite data and CO2 records, and linked the changed onset of spring to spring temperature and the North Atlantic Oscillation Index (NAO) (Menzel 1997; Menzel and Fabian 1999; Menzel 2000; Chmielewski and Rötzer 2001, 2002).

Chapter 2.3: Europe

2.2

47

ICP Forests

Phenological observations are also made at Level II plots of ICP Forests. ICP Forests is the International Co-operative Programme on Assessment and Monitoring of Air Pollution Effects on Forests, which was launched in 1985 under the convention on long-range transboundary air pollution of the United Nations Economic Commission for Europe. ICP Forests monitors the forest condition in Europe using two monitoring intensity levels. The second level (so called Level II) has been operating since 1994 in selected forest ecosystems. On these plots, soil and soil solution chemistry, foliar nutrient status, increment, meteorological condition, ground vegetation and deposition of air pollutants are measured in addition to the annual crown condition assessments. On an optional basis, phenological observations are made to provide supplementary information on the status and development of forest tree condition during the year. Since 1999, additional phenological phases are recorded to determine the course of the annual development states of forest trees, to explain possible changes in relation to environmental factors, and to utilize this knowledge in interpreting observed changes in tree condition. Information about this network including the complete manual for phenological observations is available at http://www.icp-forests.org/.

3.

NATIONAL NETWORKS

In Europe two major types of national (countrywide) phenological networks can be distinguished. In several countries, such as Albania, Austria, Czech Republic, Estonia, Germany, Poland, Russia, Slovak Republic, Slovenia, Spain, and Switzerland, the National Weather Services have been running (plant) phenological schemes during the second half of the 20th century, and some networks already existed at the beginning of the 20th century (see Schnelle 1955). In contrast to these “traditional” networks, “younger” networks have been (re)established recently. Basic information concerning the phenological networks of selected countries, such as web site and network contact, number of observers, and species and phases, are listed in Tables 1 and 2. The recorded phenological information is used by the monitoring networks themselves or externally, primarily for research in agri-, horti-, and viniculture, forestry, ecology, human health, as well as for climatic evaluations and evaluation of potential global change impacts. In most cases, National Weather Services’ phenological networks were intended to gather additional (integrated) climate information. Thus, observations of wild plants were used to monitor phenological seasons; agricultural observations of different crops and fruit

48

Phenology: An Integrative Environmental Science

trees mainly served to predict growing success, delivered data for modeling, and facilitated agro-meteorological consulting. Other purposes include the prognosis of onset dates, pollen forecasts, frost risk management, and monitoring of biotic damage. Table 2.3-1. Basic identification information about the selected European phenological networks portrayed. Network Contact Internet Current Manual number of stations Central Institute for www.zamg.ac.at 80 ZAMG Meteorology and (2000) Geodynamics, Austria Czech Hydrometeorological www.chmi.cz 158 (46 forests, Guidebooks Institute 84 crops, 28 available fruit) Slovakk Hydrometeorological www.shmu.sk 221 (61 forests, SHMÚ Institute Bratislava 53 crops, 15 fruit, 92 com. (1988, a, b, phenology) 1996 a, b) German Weather Service www.dwd.de/de/FundE/ ~ 1700 DWD (1991) Klima/KLIS/daten/nkdz/ fachdatenbank/ k datenkollekitve/ phaenologie/index.htm

Estonian Hydrometeorological Institute Environmental Agency of Slovenia MeteoSchweiz

www.emhi.ee

21

EMHI (1987)

www.rzs-hm.si/

61

Observation guidelines MeteoSchweiz (2003)

projekti/fe f no/ www.meteoschweiz.ch/de/ Beru r f/ Landwirtschaft/

160

IndexLandwirtschaft.shtml

The recent history (since 1950) and special characteristic of selected networks are quite similar: In Germany, the Deutscher Wetterdienst in 1949 and the Hydrometeorologische Dienst der DDR in 1951 took over the phenological network, which has been founded by the Reichswetterdienst in 1936 by combining different regional phenological networks. Since 1991, these networks are unified again, and the phenological network is managed by the Deutscher Wetterdienst. At the same time, the observational program was adjusted. Similarly, in Austria and Switzerland phenological phases

Chapter 2.3: Europe

49

have been recorded continuously since 1951 in networks, run by their National Weather Services. In Austria after the World War II, a new phenological network was built up by the Zentralanstalt für Meteorologie und Geodynamik in Vienna, based on an older network started in 1928. Table 2.3-2. Basic observation program information about the selected European phenological networks portrayed (S species, P phases). Network Contact First Fruit trees + Wild plant + Agricultural trees grape vines (current) crops + network farmer’s activities Central Institute for (1928) 15 + 9 S 11 S 5+1S Meteorology and 19517 P (68 Ptot ) 12 P (78 Ptot ) 9 P (26 Ptot ) Geodynamics, Austria today (1923) Czech Hydrometeorological 45 S 19 S 14 + 1 S 1986Institute 26 P 33 P 16 / 2 P today Slovakk Hydro(1923) meteorological Institute 1986today 30 S German Weather Service (1936) 10 S 10 S 59 Ptot 195166 Ptot 32 Ptot today Russian Geographical >102 S (wild Society plants, trees, crops, animals, agrometeorlogical events), >32 P

Estonian Hydrometeorlogical Institute Environmental Agency of Slovenia

1948today

68* S 7P

9S 16 P

1951today

38 S

23 S

MeteoSchweiz

1951today

11 + 11S 6P 60 Ptot

1P

* The selection of native species is voluntary for the observers.

8S 11 P

3 + 1S 3P 8 Ptot

50

Phenology: An Integrative Environmental Science

Unfortunately, the number of stations decreased from around 500 in the 1970s to 80 currently. The Swiss phenological observation network was founded in 1951 and initially consisted of 70 observation posts; the phenological observation program was slightly modified in 1996. The first phenological record in Slovenia is Scopoli's work Calendarium Florae Carniolicae from 1761. Modern phenology data collection started in 1950/1951 with the establishment of a phenological network within the Agrometeorological service, thus data are mainly used for research and applicable agriculture purposes. The recent network consists of 61 phenological stations, well distributed by a regional climatic key over the whole territory of Slovenia. The observations are carried out on species of non-cultivated plants (herbaceous plants, forest trees and bushes, clover and grasses) and of cultivated plants, such as field crops and fruit trees. In some portions of the Slovak and the Czech Republic, phenological observation was also conducted for a short time in the last half of the 19th century, but regular and managed phenological observation did not start until the 20th century. From 1923-1955, the observational program comprised more than 80 plant species (crops, fruit trees, native plants), but also some migratory birds, insects, as well as agro-technical data. From 1956 to 1985 observations were made following the first instruction guide edited by the Hydrometeorological Institute, with an enlarged program including, for example, agro-meteorological observations and crop diseases. In 1985/1986 a new system of phenological observation (including new guides) was instituted with three special sub-networks for field crops, fruit trees, and forest plants, and respective stations in regions with intensive agricultural production, in orchards and vineyard regions, and in forest regions. Three developmental stages (10%, 50%, and 100%) are now observed. In Slovakia, some historical stations were maintained in the (so called) “common” phenological network. In that network, general phenological observations (crops, fruit trees and grapevine, forest plants, migrating birds, some agro-meteorological and agro-technical data) are made by volunteers, in contrast to the “special” networks, where experts (e.g., with agronomic education) do the recording. In 1996 the guides for the common and special observation of forest plants were modified. The species and scales of phenophases are now very close to those in use before 1985. R area has many phenological observation programs The former USSR supervised mainly by Russian organizations. The Russian Geographical Society started phenological studies in 1850s with more than 600 observers, mostly in European Russia. Today, the archive of this voluntary network in St. Petersburg is one of the most important phenological centers in Russia, with more than two thousand observation sites all over the former USSR area, and regional subprograms which have different observation manuals

Chapter 2.3: Europe

51

and species lists (Hydrometeorological Printing House 1965; Schultz 1981). The second important network is the Hydrometeorological Service’s agrimeteorological observation program, organized by Schigolev in 1930, with a unique and strict methodology and very detailed observations of agricultural crops and some natural trees species, including climate parameters, such as soil temperature and moisture, snow and precipitation, at same site (Hydrometeorological Printing House 1973; Davitaja 1958). Today, the database at the central archive in Obninsk is not actively used, because the data is not digitized. Several other phenological observation programs exist that are run by the plant protection service, agricultural selection service, forestry department (Schultz 1982) or in Nature Conservation areas that use their materials for study and educational purposes (Kokorin et al. 2001). In Estonia, the first scientific phenological observation program (with more than 30 observed species) was set up in the botanical garden of University of Tartu in 1869, by the Estonian Naturalists Society (Oettingen 1882). The Estonian Naturalists Society started organized phenological studies in the 1920s / 1930s and a broad observation program in 1951 (Eilart, 1959). Today, the society is the most active voluntary observer of plant, bird, fish, phenology and seasonal phenomena in the country (Eilart 1968, ENS http://www.loodus.ee/lus/). The agri-phenological network of the Estonian Meteorological and Hydrometeorological Institute (started in 1948) used standard observation methods similar to those used in the former USSR (Hydrometeorological Printing House 1973; EMHI 1987). Their observation list consisted of agricultural plants, selected tree species, and main characteristics of the physical environment. Until the 1990s, 21 stations were still in operation (Ahas 2001), but that number diminished to 10 in 2001, and six in 2002. In Poland, the Hydrometeorological Institute ran a phenological network from 1951-1990 composed of around 70 stations. In the manual by Sokolowska (1980), phenological observations are described, and the main results are reported by Tomsazewska, and Rutkowski (1999). In Spain, the phenological network organized by the Spanish Meteorological Institute is characterized by an enormous number of stations, species, and phases, but less continuity of observations at single sites. The two examples of the British “Nature’s Calendar” and the Dutch “De Natuurkalender” stand as examples of phenological networks which have been set up recently, mostly run on the Internet, and organized by nongovernmental organizations (NGOs), media, and research institutions. They include a lot of observations on animals (e.g., birds and butterflies). A national phenological network in the United Kingdom was established by the Royal Meteorological Society in 1857. However, the subsequent development of British phenology was quite different from the continental

52

Phenology: An Integrative Environmental Science

central European countries, as annual reports were published only up until 1948. In 1998 a pilot scheme to revive a phenology network in the UK was started by Tim Sparks, research biologist at the Centre for Ecology & Hydrology in Cambridge, comprising both plant and animal phases. In autumn 2000 the Woodland Trust forces joined with the Centre for Ecology & Hydrology to promote phenology to a far wider and larger audience. In 2001 the number of registered recorders across the UK rose to over 11,700, and by August 2002 it was 16,809 and still growing, with around half of these being online observers. The Nature’s calendar’s website (http://www.phenology.org.uk) provides information about the species observed, an online list of observations, and graphic presentations of trends. In February 2001, Wageningen University and the national radio program VARA Vroege Vogels (Early Birds) started a phenological monitoring network in the Netherlands, called De Natuurkalender. This network aimed to increase understanding of changes in the onset of phenological phases, also due to climate change, for human health, agriculture, and forestry. Other aims were to strengthen the engagement of the public in their natural surroundings and to develop interactive educational programs for school children and adults. The observation program includes over 100 species of plants, birds, and butterflies with at least one phenophase per species. The phenophases are clearly defined in an observation manual. Over 2000 volunteers subscribed to the program, and send their observations via the Internet, a paper form, or a special telephone line (Fenolijn) to the coordinators of the network. The observers and other potentially interested people are informed about the results of the observations by a weekly report during the radio program, which is followed by 500,000 people every Sunday morning. Furthermore, the network uses an interactive website to provide direct feedback to the observers.

4.

OTHER NETWORKS

This rough overview of phenological networks leaves many regions in Europe blank, due to the lack of current phenological networks in those places (e.g., Portugal, Greece). Other countries only have local networks due to regional organized research structure (e.g., Italy), current networks that are run by other institutions (e.g. Finland METLA, Norway), or they mainly have historical networks (e.g. Norway, Finland). Thus, this overview does not claim to be exhaustive, and it is fairly certain that in many other countries, national or regional networks existed, or are still running. An evaluation of the World Meteorological Organization (WMO) RA VI agro-meteorological questionnaire on phenological observations and

Chapter 2.3: Europe

53

networks revealed that from 28 replying countries only six countries (Belgium, Bosnia and Herzegovina, Denmark, Luxembourg, Portugal and the United Kingdom) had no regular (agro-meteorological) phenological network, whereas 22 countries (Armenia, Austria, Croatia, Czech Republic, Estonia, France, Germany, Hungary, Ireland, Italy, Israel, Latvia, Lithuania, Macedonia, Moldavia, Romania, Russia, Slovakia, Slovenia, Spain, Switzerland, Syrian Arabic Republic) have regular phenological networks (WMO 2000). However, following this WMO evaluation, the phenological observations, the applied observations methods, the structure of the networks, the coding systems, and the practical usage of data are highly diverse. The French phenological network that started in 1880 under the care of Meteo France may represent observations “fallen into oblivion.” Phenological observations have been reported continuously up to 1960 for most stations adjacent to meteorological stations, but only three of them continued their observations after that date, with the last one stopping in 2002. The conception of this network was similar to other central European ones still in operation, as the observational program comprised perennial wild species including trees (25 species), crops (10), and fruit trees (eight), and an instruction booklet was provided to observers to standardize observations. Observations are contained in archives, but have not been digitized. These data have not been analyzed, except at the very beginning of the network by C.A. Angot from Meteo France (in a few Annales du Bureau Central Météorologique) who mapped isolines of the onset of phenophases for the decade 1881-1891.

5.

CONCLUSIONS

The scientific community has a long list of requirements for monitoring. Long-term continuous data records of high quality are needed with good documentation, many auxiliary data, and often much more. Thus, special characteristics of the networks may be of interest to them. In general, observations are made by volunteers interested in nature, in special networks (Slovakia) or special stations (Germany, IPGs), or by experts. In the IPGs, observations are made on 3 specimens of each clone, whereas national networks’ rules describe how the observing area is defined and how the specimens have to be chosen in the near surroundings of a phenological station. Constant specimen and locations are desired, however only in Slovenia forest trees, shrubs, fruit trees, and vines are they permanently marked. All networks possess paper forms to note observations (even the new established networks in the UK and in the Netherlands do not want to

54

Phenology: An Integrative Environmental Science

exclude “offline” recorders and developed forms). The requested frequency of submitting forms differs between once a year to weekly, and event-based information immediately. Data consistency and quality is difficult to evaluate. Most of the networks analyzed have monitoring guidelines, however they are very different, ranging from brand-new instructions, also available on the Internet (such as complete manual of ICP Forests, species and/or phenophase information of the Nature’s Calendar or the German Weather Service) or substantial printed manuals (Germany Weather Service), to descriptions used since the beginning of the network (IPG). Quality control of the data mostly consists of only simple plausibility control. Accompanying disaster information does not exist at the moment, but could be available in the future (e.g., ICP Forests). The general data release policy varies as well, and in most cases data are open on an individual decision basis only. Data formats are also quite different. In some countries older records still need to be digitized from paper, but most networks do have their data in ASCII files or databases. The Nature’s Calendar (with almost 50% online observers) offers quick data access for registered observers and allows different kinds of data comparisons. In the National Weather Services’ networks, observations on cultivated species are generally accompanied by information about varieties. However, associated data about the site (such as meteorology, soil, relief, and slope) are not available and (due to the coarse information about the station location) it is nearly impossible to gather exact auxiliary data. Two initiatives of the European Phenological Network (EPN project, see Chapter 2.7) may facilitate phenological research in the nearer future. The meta-database will hopefully provide a complete overview of all phenological networks in Europe including information about spatial and temporal extent, monitored species and classes, data consistency, data usage, and additional data. Coding of phenophases following the BBCH code (Chapter 4.4, Biologische Bundesanstalt 1997) may assist in understanding phenophase definitions in different languages and making observations comparable.

ACKNOWLEDGEMENTS I kindly thank my phenology colleagues Rein Ahas (Estonia), Olga Braslavska (Slovak Republic), Isabelle Chuine (France), Zoltan Dunkel (Hungary), Elisabeth Koch (Austria), Jiri Nekovar (Czech Republic), Tim Sparks (United Kingdom), Andreja Susnik (Slovenia), and Arnold vanVliet (The Netherlands) for their valuable support and information.

Chapter 2.3: Europe

55

REFERENCES CITED Ahas, R., Spatial and temporal variability of phenological phases in Estonia, Dissertationes Geographicae Universitatis Tartuensis 10, 1999. Ahas R. (Editor), Estonian phenological calendar, Publicationes Instituti Geographici Universitatis Tartuensis, 90, 206 pp., 2001. Baumgartner, A. and F. Schnelle, International Phenological Gardens (Purpose, results, and development), 16thh IUFRO World Congress Oslo Subject Group S1.03, 7 pp., 1976. Biologische Bundesanstalt für Land- und Forstwirtschaft (Editor), Growth Stages of Plants BBCH Monograph, Blackwell Wissenschafts-Verlag Berlin Wien, 622 pp., 1997. Chmielewski, F. M., The International Phenological Gardens across Europe. Present state and perspectives, Phenology and Seasonality, 1(1), 19-23, 1996. Chmielewski, F. M. and T. Rötzer, Response of tree phenology to climate change across Europe, Agricult. Forest Meteorol., 108, 101-112, 2001. Chmielewski F. M. and T. Rötzer, Annual and spatial variability of the beginning of growing season in Europe in relation to air temperature changes, Clim. Res., 19, 257-264, 2002. Davitaja, F. F., Agrometeorological problems, Moscow (in Russian with English Contents), Hydrometeorological Publishing House, 160 pp., 1958. Deutscher Wetterdienst (Ed.), Anleitung für die phänologischen Beobachter des Deutschen Wetterdienstes (BAPH), Offenbach am Main, 155 pp., 1991. Eilart, J., Phytophenological observation manual (in Estonian), Estonian Naturalists Society, Tartu, 16 pp., 1959. Eilart, J., Teaduse ajaloo ehekülgi Eestis, in Some aspects of history of phenology in Estonia, edited by Ü. Ü Lumiste (in Estonian with summary in German and Russian), Tallinn, Academy of Sciences, 1, 169-176, 1968. EMHI, Manual for Hydrometeorological Observation Stations and Points, Estonian Department of Hydrometeorology and Environmental Monitoring 11, I-II, Tallinn, 164 pp., 1987. Hydrometeorological Printing House, Natural calendars of North Western USSR (in Russian), Geographical Society of USSR, Hydrometeorological Printing House, Leningrad, 71 pp., 1965. Hydrometeorological Printing House, Methodology for hydrometeorological stations and observation points 11, Agri-meteorological observations in stations and observation points, 3rd edition (in Russian), Hydrometeorological Printing House, Leningrad, 186 pp., 1973. Kokorin A. O., A. V. Kozharinov, and A. A. Minin (Eds.), Climate change impact on Ecosystems, Nature protected areas in Russia, analyses of long-term observations, WWF Russia policy book No. 4 (in Russian with English Summaries), World Wildlife Foundation, 174 pp., 2001. Menzel, A., Phänologie von Waldbäumen unter sich ändernden Klimabedingungen, Dissertation at the Forest Faculty of the LMU Munich, Forstlicher Forschungsbericht 164, 1997. Menzel, A., Trends in phenological phases in Europe between 1951 and 1996. Int. J. Biometeorol., 44, 76-81, 2000. Menzel, A., Phenology, its importance to the global change community, Climatic Change, 54, 379-385, 2002. Menzel, A., and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999. MeteoSchweiz (Editor), Pflanzen im Wandel der Jahreszeiten – Anleitung für phänologische Beobachtungen, Geographica Bernensia, 287 pp., 2003

56

Phenology: An Integrative Environmental Science

Oettingen, A.J. von, Phänologie der Dorpater Lignosen, in Archiv Naturk. Liv-, Est- u. Kurlands, II Serie. Bd. 8, pp. 241-352, Dorpat, 1882. Schnelle, F., Pflanzen-Phänologie, Akademische Verlagsgesellschaft, Geest and Portig, Leipzig, 299 pp., 1955. Schultz, G. E., General phenology (in Russian), Nauka, Leningrad, 186 pp., 1981. Schultz, G. E., Geographische Phänologie in der USSR (in German), Wetter und Leben, 34, 160-168, 1982. þ Ģ fenologických staníc PoĐné Đ SHMÚ Bratislava (Editor), Metodický predpis 2 Návod na þinnos plodiny (Manual for special crop station), 120 pp., 1988a. þ Ģ fenologických staníc SHMÚ Bratislava (Editor), Metodický predpis 3 Návod na þinnos Ovocné plodiny (Manual for special fruit and grapevine station), 136 pp., 1988b. SHMÚ Bratislava (Editor), Fenologické pozorovanie všeobecnej fenológie, Metodický predpis. (Manual for general stations), 126 pp., 1996a. SHMÚ Bratislava (Editor), Fenologické pozorovanie lesných rastlín, Metodický predpis (Manual for forest stations), 16 pp., 1996b. Sokolowska, J., Przewodnik fenologiczny, Instytut Meteorologii i Gospodarki Wodnej, Wydawnictwa Komunikacji i Lacznosci, Warszawa, 163 pp., 1980. Sparks, T.H. and P. D. Carey, The responses of species to climate over two centuries: an analysis of the Marsham phenological record, 1936-1947, J. Ecology, 83, 321-329, 1995. Tomsazewska, T., and Z. Rutkowski, Fenologiczne pory roku u uch zmiennosc w wieloleciu 1951-1990, Materialy Badawcze, Seris Meteorlogia – 28, Instytut Meteorologii i Gospodarki Wodnej, Warszawa, 39 pp., 1999. World Meteorological Organization, Commission for Agricultural Meteorology (Editor), Report of the RA VI Working Group on Agricultural Meteorology, CAgM report No. 82, WMO/TD No. 1022, 274 pp., 2000. Zentralanstalt für Meteorologie und Geodynamik, Anleitung zur phänologischen Beobachtung in Österreich, Anleitungen und Betriebsunterlagen Nr 1 der ZAMG, Wien, 31 pp., 2000.

Chapter 2.4 NORTH AMERICA Mark D. Schwartz1 and Elisabeth G. Beaubien2 1

Department of Geography, University of Wisconsin-Milwaukee, Milwaukee, WI, USA; Devonian Botanic Garden, University of Alberta, Edmonton, Alberta, Canada

2

Key words:

Agricultural Experiment Station networks, Lilacs, Royal Society, Plantwatch, Environment Canada

1.

UNITED STATES

1.1

Early Observations and Research

Throughout the early history of the United States, extending into the first years of the 20th century, there were few attempts to create organized phenological networks. One of the most noteworthy in this period was started by the Smithsonian Institution in 1851, and included observations on eighty-six plant species, birds, and insects in thirty-three states, but only lasted till 1859 (Hough 1864; Hopp 1974). A few individuals who were part of this and subsequent weather/climate observation networks did record phenological data at selected sites during other periods. For example, Dr. Samuel D. Martin’s April 1865 report (from Pine Grove, Kentucky) contains the dates of numerous phenological events (Martin 1865). Thomas Mikesell at Wauseon, Ohio, compiled another important local record over the period 1873-1912 (Smith 1915). Later instructions to Weather Bureau observers included lists of phenological phenomena to record (concentrating on agricultural crops, but also including timing of leaf opening and fall in deciduous forests, Weather Bureau 1899). However, there is little evidence to suggest that large numbers of observations were taken based on these Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 57-73 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

58

Phenology: An Integrative Environmental Science

instructions. Hopp (1974) notes that the Weather Bureau made a final limited attempt to start a phenological network at twenty cooperative sites in the state of Indiana, during 1904-1908, and lists several other extensive local records taken in Indiana, Kansas, and Minnesota. An important phenological research contribution from the United States during the first half of the 20thh century was Hopkins’ (1938) “Bioclimatic Law.” The most well-known part of this law states that (other conditions being equal) the south to north progression of spring phenological events in temperate portions of North America is delayed by four days for each degree of latitude northward, for each five degrees of longitude eastward, and for each 400-foot increase in elevation. This model was developed from data available around the northern hemisphere at the time. Hopp (1974) observed that the law is highly generalized, has geographical limitations, and is difficult to apply to individual plant species in any one season. Despite these limitations, Hopkins’ Bioclimatic Law became one of the best-understood concepts of phenology for other scientists and the public. Paradoxically, its simplicity could have made phenology seem too easily predictable, which may have hindered and delayed efforts to develop new data collection networks, especially in the United States.

1.2

Agricultural Experiment Station Regional Networks

The first extensive U.S. phenological observation networks began in the 1950s with a series of regional agricultural experiment station projects, designed to employ phenology to characterize seasonal weather patterns and improve predictions of crop yield (Schwartz 1994). J. M. Caprio at Montana State University began the first of these projects, W-48 “Climate and Phenological Patterns for Agriculture in the Western Region” in 1957. This network contained up to 2500 volunteer observers distributed throughout 12 Western states (Caprio 1957, Figure 1). Common purple lilac plants (Syringa vulgaris) were observed initially, with two honeysuckle cultivars ((Lonicera tatarica ‘Arnold Red’ and L. korolkowiii ‘Zabeli’ added later. Observations ended in 1994, however, a few observers have again reported data since the later 1990s (Cayan et al. 2001). Encouraged by Caprio’s program, similar projects were started in the central U.S. (NC-26 “Weather Information for Agriculture”) by W. L. Colville at the University of Nebraska in 1961, and in the northeastern U.S. with the renewal of NE-35 (“Climate of the Northeast—Analysis and Relationships to Crop Response”) by R. Hopp at the University of Vermont in 1965. Both of these networks observed cloned plants of the lilac cultivar Syringa chinensis ‘Red Rothomagensis’ and the two honeysuckle cultivars from W-48. In 1970, NC-26 and NE-35 were combined as part of a new

Chapter 2.4: North America

59

regional project, NE-69 “Atmospheric Influences on Ecosystems and Satellite Sensing.” The program expanded to about 300 observation sites (Figure 1), with three individuals as unofficial leaders: B. Blair (Purdue University), R. Hopp (University off Vermont), and P. Dubé (Laval University). In 1975 NE-69 was replaced by another new project, NE-95, “Phenology, Weather and Crop Yields,” which was replaced by still another, NE-135, “Impacts of Climatic Variability t on Agriculture” in 1980. This project was coordinated by M. T. Vittum (Cornell University) until 1985, when responsibility was turned over to R. C. Wakefield (University of Rhode Island). The phenology portion of NE-135 was briefly supervised by W. Kennard (University of Connecticut) until the eastern U.S. network lost funding and was terminated at the end of 1986.

Figure 2.4-1. Locations in North America with five years or more of lilac phenology data, 1956-2001 (Québec stations not included).

As an extension of the lilac/honeysuckle regional networks, a statewide phenological garden system of 12 stations operated in Indiana during the 1960s and 1970s. Numerous protocols were developed, and observations were taken on up to 14 species at each site (Blair et al. 1974). Also, an extensive phenology network observing redbud (Cercis canadensis),

60

Phenology: An Integrative Environmental Science

dogwood (Cornus florida), and red maple ((Acer rubrum) operated in North Carolina during the 1970s (Reader et al. 1974).

1.3

Network-Related Phenological Research

A large number of horticultural and physiological studies have reported results regarding site-specific phenological characteristics of (commercially important) fruit tree species, and general phenological responses of woody plants. These are adequately summarized elsewhere, and will not be addressed here (e.g., Flint 1974; Schwartz 1985; Schwartz et al. 1997). Relatively few researchers have taken advantage of the Agricultural Experiment Station Regional Network data to examine phenological relationships on the continental scale. Caprio (1974) was the first, developing the Solar Thermal Unit Concept from lilac phenological data recorded in the western U. S. These data were also used in a recent study to examine the relationship between lilac-honeysuckle phenology and the timing of spring snowmelt-runoff pulses, in the context of global change. Earlier spring onsets since the late 1970s are reported throughout most of the region (Cayan et al. 2001). Schwartz (1985) began an extensive phenological research program in the mid-1980s that has made intense use of lilac-honeysuckle network data from the eastern U. S. Areas explored include modeling, resulting in the Spring Indices (e.g., Schwartz 1998; Schwartz and Reiter 2000, and detailed discussion in Chapter 4.8), spring plant growth impacts on the lower atmosphere (e.g., Schwartz 1992; Schwartz and Crawford 2001), and analyses and comparisons to remote sensing measurements (discussed in Chapter 5.1)

1.4

Current Networks

After the “decommissioning” of eastern U. S. lilac-honeysuckle phenology network operations by the Agricultural Experiment Stations in 1986, M. D. Schwartz corresponded with the most recent network supervisors (Schwartz 1994). They granted him permission to contact the observers and invite them to continue participating in an “interim” network, pending new funding. Approximately 75 observers responded to a renewed survey form sent out in March 1988, returning data for 1988 and in many cases 1987 as well. From that time to the present, Schwartz has continued to operate this interim “Eastern North American Phenology Network” with approximately 50 observers reporting lilac or honeysuckle event dates each year (http://www.uwm.edu/~mds/enanet.html). Plans are underway to expand this network across the entire country, and add observation of events

Chapter 2.4: North America

61

from a small number of appropriate native species at each site, following the example of Plantwatch Canada (see section 2.1.5). Another operational network of note is run by the Wisconsin Phenological Society (http://www.naturenet.com/alnc/wps/). Data for a large number of native and cultivated flowers and shrubs extend back to the early 1960s. Unfortunately, there is considerable variation in the number and types of plants observed at each site (selected from a standard form), and most individual station records are less than 10 years long. However, these data are now largely in digital form, and have contributed to an innovative methodological study exploring ways to “fill-in” the gaps in such incomplete Lastly, the GLOBE program records (Zhao and Schwartz 2003). (http://www.globe.gov), which works to get primary and secondary students involved in taking measurements and interacting with scientists has developed a number of phenology protocols (including a lilac phenology “special measurement” based on the Agricultural Experiment Station event descriptions).

2.

CANADA

Canada has a long and rich history of phenological observations. Since deglaciation some eight to ten thousand years ago, First Nations and Inuit have perfected their oral knowledge of “nature’s calendar” to maximize their survival and find resources efficiently across a wide landscape. The earliest recorded observations will likely be found in the journals of fur traders and missionaries. Phenological data “serve as a check of season against season, and region against region” (Minshall 1947, p.56). Phenological studies vary in the size of area surveyed and in the duration of observations, but they can basically be divided into three types: – the “snapshot” study, in which many observers survey phenology over a large area at one point in time. – the intensive study, in which one or a small number of people survey a small area over a period of one or more growing seasons. – the extensive study, in which a network of observers surveys a large area over a period of years. This article concentrates mainly on involvement of Canadians in such networks. The studies described here focus mostly on plant phenology and are divided into two major sections, national and regional networks. Firstly, the national networks described include extensive studies such as the Royal Society of Canada survey launched in 1881, participation by eastern Canadian observers in the NE Agricultural Project in the United States in the 1950s, and more recently Plantwatch, a program engaging Canadians with

62

Phenology: An Integrative Environmental Science

coordinators in each province and territory. Secondly, the regional and localized networks and research are described by region, from east to west, and in the north.

2.1

National Networks in Canada

2.1.1

First Nations and Inuit: traditional phenological knowledge

For Canada’s First Nations, phenology was a well-honed tool. The Blackfoot in Alberta used the flowering time of Thermopsis rhombifolia (golden bean or buffalo bean) to indicate the best time in spring to hunt bison bulls (Johnston 1987). In British Columbia more than twenty cultural/linguistic groups used over 140 indicators (Lantz and Turner, in press). These authors note that phenological indicators permitted the most efficient use of human resources in acquiring food or materials from the land. One example is that the west coast Nuu-Chah-Nulth peoples use the ripening of salmonberries (Rubus ( spectabilis) to indicate that adult sockeye salmon are starting to run in freshwater streams. Also the interior Stl’atl’imx peoples used the blooming of wild rose (Rosa spp.) to indicate the best time to collect cedar roots and basket grass. A third: the west coast Comox peoples used the bloom time of oceanspray (Holodiscus ( discolor) to alert them to dig for butter clams. In the Okanagan area, First Peoples observed that female black bears generally headed to dens when the western larch needles turned gold in the fall. Later denning often meant these bears would miscarry and not produce cubs (perhaps due to poor berry crops and thus insufficient weight gain). Across North America accurate timing was the key to survival, and phenology was “common sense” to those who lived so close to the land. 2.1.2

Royal Society of Canada survey

Only twenty-three years after Confederation in 1867, a countrywide phenology survey was initiated. In 1890 the Royal Society of Canada passed a resolution requesting affiliated natural history and scientific societies to: “obtain accurate records in their individual localities of meteorological phenomena, dates of the first appearance of birds, of the leafing and flowering of certain plants, and of any events of scientific interest for collation and publication in the ‘Transactions of the Society’” (Proceedings 1893, p. 54).

Chapter 2.4: North America

63

In the following year, 1891, the Botanical Club of Canada was formed, affiliated with the Royal Society (MacKay 1899). One of its departments was responsible for promoting the nation-wide phenology survey, which initially included observations of 67 events. In 1897 this number increased to 100 events. These included recording the first bloom dates of many species of native plants, as well as the timing of migratory bird arrivals, thunderstorms, and ice melt on rivers. Blank schedules were distributed annually to past and potential observers with instructions to submit them by the end of January for compilation. In the first year, 1892, nine observers from Nova Scotia reported. In 1895, 25 reports were received from nine provinces. In general, the numbers of events and the diversity of locations increased over the duration of this extensive survey. The secretary of the Botanical Club, Dr. A. H. MacKay, coordinated the survey up until 1910 when the club was dissolved, and F.F. Payne of the Meteorological Service then coordinated the survey until 1922. With this change of coordination the number of events observed dropped to 50. Observations for 1892-1922 were published annually in the Proceedings and Transactions of the Royal Society of Canada. 2.1.3

Participation of Atlantic and Central provinces in U.S. Agricultural Experiment networks

The U.S. regional agricultural experiment station northeast project, NE69, added stations in several Canadian provinces in 1970 (see also section 1.2). The next year Quebec initiated a large observer network to track these plant species. Observations were made until 1977 at over 300 locations, of which 51 were adjacent to meteorological stations (Dubé and Chevrette 1978). Three indices (earliness index, summer index, and growing season index) were derived from the data, which were used to define bioclimatic zones. By 1977 observers in the six eastern provinces were involved. Quebec had the largest number of observers by far, with 268 active observation sites in 1977 versus New York State, the next largest, at 84 sites (Vittum and Hopp 1978). Pierre André Dubé of Laval University coordinated the Quebec participants and also computerized results for the whole project. Analysis of the phenological and meteorological data confirmed the existence of significant differences between phenological zones in Quebec (Castonguay and Dubé 1985). The resulting maps were used to modify agricultural taxation zones.

64 2.1.4

Phenology: An Integrative Environmental Science Canadian Forest Service (Natural Resources Canada): intensive studies

Several Forestry Centers have done intensive studies of the phenology of trees and insects, such as Parry et. al. (1997) but large networks of observations have not been established. Several historic data series of the phenology of three conifer species exist for New Brunswick and Quebec for 1975-2001. The phenology of the spring development of spruce budworm (Choristoneura fumiferana) in New Brunswick for 1950-1985 has been studied, by the Canadian Forest Service, the Quebec Ministry of Natural Resources, and the New Brunswick Department of Natural Resources and Energy. The data were being modeled and published at the time of writing (J. Regnière, Canadian Forest Service, Québec, personal communication, 2002). 2.1.5

Plantwatch: national network

Plantwatch began in 1995 based at the University of Alberta’s Devonian Botanic Garden. It has enlisted volunteers in North America and internationally to track spring bloom times of indicator plant species useful as key indicators for phenology (Beaubien 1997; Beaubien and Freeland 2000). Reporting has been via the Internet: (http://www.devonian.ualberta.ca/pwatch) where tables and maps of data were updated regularly, up to and including 2001 (see “archives” at this website). Initially the focus of this extensive survey was on students (ages eight to eleven years), reporting bloom dates for three plant species across the Prairie Provinces. By 1997 the survey had expanded to a Canada-wide program for both adults and youth, with seven indicator plants. For more information see the article posted at: http:// eqb-dqe.cciw.ca/eman/reports/publications/nm97_abstracts/part-31.htm (Beaubien 1997). International data were also gathered for one species Syringa vulgaris: common purple lilac. Beginning in the early 1990s Elisabeth Beaubien gave talks across Canada to encourage the formation of provincial plant phenology programs. A teacher guide was posted on the University of Alberta website (above) in 2001, providing curriculum applications in science, mathematics, social studies, etc. for students from elementary to high school level. In 2000-2002 Plantwatch expanded with assistance from Environment Canada’s Ecological Monitoring and Assessment Network Coordinating Office (EMAN CO). Elisabeth Beaubien, as national coordinator, made contacts to find coordinators for each of the provinces and territories. The coordinators met in Ottawa in May 2000 and in Winnipeg in November

Chapter 2.4: North America

65

2001. The national coordinator developed 14 criteria to select plant species best suited to a cross-Canada spring phenology survey by the public, and applied these to select 15 species. Each province and territory selected species from this list, with some coordinators adding more species suitable for their particular ecozones. The national coordinator researched phenology protocols used by the Deutscher Wetterdienst (weather service in Germany), and other European networks, to simplify and standardize the phenophase descriptions for first, full bloom, and leafing. First bloom was defined as “in at least three places on the observed tree or shrub, or patch of smaller plants, the first flowers have just opened.” Mid bloom (called full bloom in Germany) was when “50% of the flower buds were now open”. Leafing was defined as “when, in at least three places on the tree or shrub, the first leaves have emerged and unfurled completely.” In the spring of 2002 a booklet “Plantwatch: Canada in Bloom” was produced through the Canadian Nature Federation (CNF) and a webpage (http://www.plantwatch.ca) posted. Plantwatch is one of several environmental monitoring or “NatureWatch” programs currently sponsored by EMAN CO and CNF. Others are FrogWatch, WormWatch, and IceWatch.

2.2

Atlantic Region

2.2.1

Nova Scotia

2.2.1.1 Student network 1897 to 1923 Dr. A.H. MacKay (see section 2.1.2) was not only secretary for the Botanical Club of Canada, but also Superintendent of Education for Nova Scotia. He promoted phenology very successfully. The popularity of the program was such that in 1898, 800 sets of observations on up to 100 events were submitted by school classes (MacKay 1899). Observations by Nova Scotia schools continued as part of the Nature Studies curriculum until at least 1923 (MacKay 1927). 2.2.1.2 Recent networks This interest in Nova Scotia for tracking nature’s calendar has been rekindled through a number of programs. A “Peeper Program” started in 1994 based at the Nova Scotia Museum of Natural History, for the public to report calling dates of spring peeper frogs. Nova Scotia Plantwatch began in the spring of 1996, tracking bloom times for 12 plant species at about 200 sites. Liette Vasseur, Peta Mudie, Bob Guscott and others formed the initial team to promote and coordinate Plantwatch. They produced an observer’s guide, a webpage on the EMAN CO website, and a colorful newsprint poster of the 12 plant species. In 2001 Ed Reekie of Acadia University took over

66

Phenology: An Integrative Environmental Science

Liette Vasseur’s functions, summarizing the data and sending out annual newsletters. Fifteen plant species are now tracked. The results of the first three years (1996-1998) were compared with the MacKay data for the same species. For most species significant differences in bloom times were not found. (Vasseur et. al. 2001). It is interesting to note that climate records for Atlantic Canada show a cooling trend for 1948 to 1995. The “Thousand Eyes” project started in 2000 based at the Nova Scotia Museum of Natural History (http://www.thousandeyes.ca), first coordinated by Elizabeth Kilvert and later by Chris Majka. It gathers records via the Internet from students in Nova Scotia on timing of 50 events selected from the MacKay program. These include first blooms, bird migrations and weather events. Results are compared with student observations from a century ago. 2.2.2

New Brunswick

Dr. Liette Vasseur moved to the University of New Brunswick in 2001 and started a New Brunswick Plantwatch that tracks 12 species. Liette Vasseur and Diane Pruneau have worked with Atlantic region teachers to encourage participation by school classes in the NatureWatch programs. 2.2.3

Prince Edward Island

The Bedeque Bay Environmental Management Association started a PEI Plantwatch in 2000 with Ilana Kunelius promoting the project to students and volunteers. In 2002 Charmaine Noonan took Ilana Kunelius’ place as Plantwatch coordinator, promoting the tracking of 12 plant species. In 2003 a Plantwatch video as well as a PEI Plantwatch Guide were being produced. Three school classes and several members of the public were observers in 2002. The PEI Natural History Society gathers phenology data from members and publishes it annually in the Island Naturalistt newsletter. 2.2.4

Newfoundland and Labrador

Luise Hermanutz and Madonna Bishop of Memorial University started Newfoundland and Labrador Plantwatch in 1998. The number of plants observed has grown from five species in 1998 to 13 species in 2002. They produced an observer’s guide, a webpage (http://www.mun.ca/biology/plantwatch), and an annual newsletter for observers.

Chapter 2.4: North America

2.3

Central Canada

2.3.1

Quebec

67

Dr. P. A. Dubé coordinated a large phenology network in the 1970s (see section 2.1.3). Intensive studies of tree phenology have been done by Dr. Martin Lechowicz and his students at McGill University (Hunter and Lechowicz 1992; Lechowicz 1995), and he continues to promote phenology as an essential method to monitor ecosystems for the effects of climate change (Lechowicz 2001). Since 2001 “Opération floraison” or Quebec Plantwatch has been based at the Montréal Botanical Garden, coordinated by botanist Stéphane Bailleul. Fifteen species of plants are tracked. Detailed observations of bloom times were made at the botanical garden in 20012002. Promotion to the public in Quebec will begin in 2003, once the French translation of the booklet “Plantwatch: Canada in Bloom” becomes available. 2.3.2

Ontario

Starting in 1932 dates for the Ottawa district of flowering and fruiting for weeds and native plants were gathered by the Division of Botany and Plant Pathology of the (federal) Department of Agriculture (Minshall 1947). Minshall provides a brief review of early Canadian phenology research, and notes that in 1939-1940 an interdepartmental federal committee presented recommendations to coordinate all federal projects in phenology. Unfortunately the subsequent war prevented action on these recommendations. Bassett et al. (1961) of the federal Department of Agriculture, analyzed this data for selected trees, shrubs, herbs and grasses for 1936-1960 and calculated the effective base temperatures for spring development of 10 early-blooming tree species. The Royal Botanical Garden in Hamilton gathered lilac data as part of the NE Network described in section 1.3 in the 1970s. This is the base for the Plantwatch program for Ontario that began in 2002 and now tracks 14 plant species. Carl Rothfels replaced initial coordinator Bruce Peart r in 2003. No promotional products have been produced as yet. In 1995 Don MacIver of the Headwaters Coalition (Grand River Valley Conservation Authority) launched a program called Green Wave Ontario. Observers tracked phenology of existing lilacs, honeysuckles, and wildflowers for a few years.

68

2.4

Phenology: An Integrative Environmental Science

Western Canada

Research on first bloom dates for 50 native plants was carried out by Agriculture Canada’s Laboratory of Plant Pathology in three prairie cities, Winnipeg, Saskatoon, and Edmonton, from 1936 to 1961. Wheat phenology (time of seeding, emergence, heading, maturity) was also recorded for comparison, and results of this intensive study were published annually (Russell 1962). Erskine (1985) presents data for first bloom of native plants and bird arrival for five boreal sites across Canada, at a different site for each year from 1971 to 1975. He also notes the appeal of phenology observations: “such visible events as the flowering of plants and the arrival of birds have an appeal that is lacking in the cold statistics of the meteorological record.” (p. 188). He lists many phenology references previously published in the Canadian Field Naturalist, commenting that most phenology studies at that time were done as extra studies, which were related to but were not the main task of field workers. 2.4.1

Manitoba

Criddle (1927) observed 400 prairie species over 20 years in southern Manitoba and published flowering and seed-ripening times valuable for present-day use in reclamation work. Mitchener (1948) presented data on flowering times and pollen availability from beekeepers. In 2001 Kim Monson became coordinator of Manitoba Plantwatch, which is cosponsored by the University of Winnipeg Geography Department and the Manitoba Naturalists’ Society. Sixteen plants are tracked and products include a promotional pamphlet, an observer guide, a teacher guide, and annual newsletters. Observers receive a certificate with space for an annual sticker to thank them for data submitted. 2.4.2

Saskatchewan

Budd and Campbell (1959) report first bloom dates for 145 prairie native plants recorded at Agriculture Canada’s Experimental Farm in Swift Current. They recommend using the bloom of Wood’s rose (Rosa woodsii) as the best indicator of “range readiness” (i.e., pasture plants can now withstand grazing). Saskatchewan Plantwatch began in 2001, coordinated by Kerry Hecker and affiliated with Nature Saskatchewan. Eleven plant species are tracked, and promotional articles have been produced.

Chapter 2.4: North America 2.4.3

69

Alberta

Moss (1960) recorded "height of bloom" dates for 25 spring-blooming shrubs and trees near the University in Edmonton, Alberta from 1926 to1958. These flowering data were correlated with degree-days to determine the average amount of warmth the plants were exposed to before flowering. Dr. Charles Bird, Professor of Botany at the University of Calgary, established a volunteer network to record the flowering of native plants, and results were published annually in Alberta Naturalist, the journal of the Federation of Alberta Naturalists, from 1973-1982 (Bird 1974). This survey was revived and revised in 1987 (Beaubien 1991) as the Alberta Wildflower Survey, based at the University of Alberta. Between 150-200 volunteers per year reported dates of first (10%), mid (50%), and full (90%) flowering for 15 native plant species from 1987-2001. Training was provided using printed program information with tips on site selection, protocols, and species identification, including color photographs and sketches. These data have been used in preliminary determinations of growing degree-days required for flowering (Beaubien and Johnson 1994). In 2002 the program was renamed Alberta Plantwatch. Some later-blooming species were dropped and others added for a total of 21 species, and phenophase definitions were changed to match the updated national program. Newsletters have been sent out each fall and spring since 1987 by mail and more recently, also by email. The May Species Count is a “snapshot phenology” study, which has been coordinated by the Federation of Alberta Naturalists on the last weekend of May every year since 1976. It includes a count of wildlife including plants (species in bloom), birds, mammals, butterflies, etc. Numbers of plant species found in bloom indicate the relative earliness or lateness of the spring season. 2.4.4

British Columbia

The entomologist R. Glendenning (1943) summarized some methods, history and uses of phenology and recommended certain species as suitable across Canada. Based on his own observations over 34 years, he suggested phenological events to observe in each month of the year for the British Columbia coast. In 1984 Bill Merrilees of the Federation of British Columbia Naturalists launched a study of the flowering of vascular plants, requesting 16 phenophases for up to 50 native species of the observer's choice. In 2000 he modified the survey, requesting bloom times for a shorter list of plants found in southeastern Vancouver Island. In 2001 Dave Williams of the University College of the Cariboo in Kamloops took on the

70

Phenology: An Integrative Environmental Science

task of coordinating Plantwatch in British Columbia, tracking the phenology of 15 plant species.

2.5

Northern Canada

Erskine (1985) describes articles in northern phenology research published in the Canadian Field Naturalist. Climate warming is predicted to show the biggest effects in arctic regions. In 1990 the International Tundra Experiment (ITEX) began, linking arctic and alpine scientists to study the effects of climate change on northern ecosystems (Henry and Molau 1997). In 1999 the Canadian researchers in this group started CANTTEX, the Canadian Tundra and Taiga Experiment (http://www.emannorth.ca/canttex) to develop a strategy for studying climatic and ecological change in Canada’s north. Presently there are 13 research and monitoring sites which track phenology using ITEX protocols, and/or carry out experimental manipulations such as warming or fertilizing. The three northern territories, Yukon, Northwest Territories and Nunavut, all started Plantwatch programs in 2001. In the Yukon Lori Schroeder is coordinator through the Yukon Conservation Society, promoting tracking of 16 species. In the Northwest Territories Jen Morin coordinates through Ecology North, tracking 15 species. In Nunavut, Paula Hughson of Parks Canada is assisted by Guy d’Argencourt and Jamal Shirley of the Nunavut Research Institute in promoting the tracking of seven plant species (see http://pooka.nunanet.com/~research/plantwatch.htm). The coordinators from the three territories and northern Manitoba (Kim Monson) with assistance from E. Beaubien and Leslie Wakelyn of Environment Canada’s Ecological Monitoring and Assessment Network North (EMAN-North) applied as “Plantwatch North” and received funds from Environment Canada’s Northern Ecosystems Initiative in 2001 and 2002. As a result each region now has an observer’s guide, and use the “Plantwatch North” poster produced in 2002 (versions in English, French and Inuktitut), as well as recognition pins for observers. The funding also permitted several workshop meetings. In 2003 a full color booklet describing the Plantwatch North program was in preparation.

2.6

Conclusions

Canada enjoys a wealth of phenological studies, starting with early applications by First Nations and continued today by naturalists, gardeners, students and scientists. Environment Canada has now embraced phenology via its “NatureWatch” programs, to involve the public in finding out what is changing in the environment and why. Plantwatch has a bright future in

Chapter 2.4: North America

71

engaging the public as “eyes of science”, tracking nature’s plant calendar and boosting awareness of the biotic response to climate change. Historically, many Canadians have enjoyed noting down the timing of seasonal events such as the appearance of wildflower blooms, spring birds, calling of frogs, and melt or freeze-up of lakes and rivers. If the reader is aware of records of such “closet phenologists”, or of other phenology studies in Canada not mentioned in this article, the author would be delighted to learn of these (see Beaubien’s address). These data are important as a baseline against which to compare current timing in our increasingly variable climate.

ACKNOWLEDGEMENTS United States section: Thanks to Glen Conner for information on Dr. Samuel D. Martin and other early phenological observers and networks. Canada section: Thanks to these regional coordinators for review of the text: S. Bailleul, M. Bishop, L. Hermanutz, C. Noonan, C. Rothfels, L. Shroeder, and L. Vasseur. L. Seale kindly edited the article. Thanks to EMANCO and the Canadian Nature Federation for their help in promoting the Plantwatch program. Dr. Geoffrey Holroyd, Canadian Wildlife Service, has financially supported much work in phenology.

REFERENCES CITED Bassett, I. J., R. M. Holmes, and K. H. MacKay, Phenology of several plant species at Ottawa, Ontario, and an examination of the influence of air temperatures, Can. J. Plant Sci., 41, 643-652, 1961. Beaubien, E. G., Phenology of vascular plant flowering in Edmonton and across Alberta, MSc. thesis, Dept. of Botany, University of Alberta, Edmonton, 1991. Beaubien, E. G., Plantwatch: Tracking the biotic effects of climate change using students and volunteers. Is spring arriving earlier on the prairies?, in Ecological Monitoring and Assessment Network Report on the 3rd National Science Meeting, pp. 66-68, Environment Canada, Saskatoon, January 1997. Beaubien, E. G., and H. J. Freeland, Spring phenology trends in Alberta, Canada: links to ocean temperature, Int. J. Biometeorol., 44, 53-59, 2000. Beaubien, E. G., and D. L. Johnson, Flowering plant phenology and weather in Alberta, Canada, Int. J. Biometeorol., 38, 23-27, 1994. Blair, R. J., J. E. Newman, and J. R. Fenwick, Phenology Gardens in Indiana, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 45-54, Springer-Verlag, New York, 1974. Bird, C. D., 1973 flowering dates, Alta. Nat., 4(1), 7-14, 1974.

72

Phenology: An Integrative Environmental Science

Budd, A. C., and J. B. Campbell, Flowering sequence of a local flora, J. Range Manage., 12, 127-132, 1959. Caprio, J. M., Phenology of lilac bloom in Montana, Science, 126, 1344-1345, 1957. Caprio, J. M., The Solar Thermal Unit Concept in Problems Related to Plant Development and Potential Evapotranspiration, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 353-364, Springer-Verlag, New York, 1974. Castonguay, Y., and P. A. Dubé, Climatic analysis of a phenological zonation: a mutivariate approach, Agric. and Forest Met., 35, 31-45, 1985. Cayan, D. R., S. A. Kammerdiener, M. D. Dettinger, J. M. Caprio, and D. H. Peterson, Changes in the Onset of Spring in the Western United States, Bull. Amer. Met. Soc., 82, 399-415, 2001. Criddle, N., A calendar of flowers, Can. Field Nat., 41, 48-55, 1927. Dubé, P. A., and J. E Chevrette, Phenology applied to bioclimatic zonation in Québec, in Phenology: an aid to agricultural technology, Vt. Agric. Exper. Sta. Bull., 684, edited by R.J. Hopp, pp. 33-43, Vermont Agricultural Experiment Station, Burlington, 1978. Erskine, A. J., Some phenological observations across Canada's boreal regions, Can. Field Nat., 99(2), 185-195, 1985. Flint, H. L., Phenology and Genecology of Woody Plants, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 83-97, Springer-Verlag, New York, 1974. Glendenning, R., Phenology, the most natural of sciences, Can. Field Nat., 57, 75-78, 1943. Henry, G. H. R., and U. Molau, Tundra plants and climate change: The International Tundra Experiment (ITEX), Global Change Biol., 3(Suppl. 1), 1-9, 1997. Hopkins, A. D., Bioclimatics–A science of life and climate relations, U.S. Dept. Agr. Misc. Publ. 280, 1938. Hopp, R. J., Plant Phenology Observation Networks, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 25-43, Springer-Verlag, New York, 1974. Hopp, R. J., ed., Phenology: an aid to agricultural technology, Vt. Agric. Exper. Sta. Bull., 684, Vermont Agricultural Experiment Station, Burlington, 51 pp., 1978. Hough, F. B., Observations upon periodical phenomena in plants and animals from 1851 to 1859, with tables of the dates of opening and closing of lakes, rivers, harbors, etc., in Results of Meteorological Observations, Made Under the Direction of the United States Patent Office and the Smithsonian Institution, from the year 1854 to 1859, inclusive, Rept. of the Commissioner of Patents, Vol. 2, Part 1, Exec. Doc. 55, 36th Congress, 1st Session, U.S. Government Printing Office, Washington, 1864. Hunter, A. F., and M.J. Lechowicz, Predicting the timing of budburst in temperate trees, J. Appl. Ecol., 29, 597-604, 1992. Johnston, A., Plants and the Blackfoot, Occas. Paper No. 15, Lethbridge Historical Society, Historical Society of Alberta, Lethbridge, 1987. Lantz, T. C., and N. J. Turner, Traditional phenological knowledge (TPK) of aboriginal peoples in British Columbia, Jour. of Ethnobiol., (in press), 2003. Lechowicz, M. J., Seasonality of flowering and fruiting in temperate forest trees, Can. J. Bot., 73, 175-182, 1995. Lechowicz, M. J., Phenology, in Encyclopedia of Global Environmental Change, vol. 2, Biological and ecological dimensions of global environmental change, edited by J. G. Canadell, pp. 461-465, Wiley, London, 2001. MacKay, A. H., Phenological observations in Canada, Can. Record of Sci., 8(2), 71-84, 1899. MacKay, A. H., The phenology of Nova Scotia, 1923, Trans. Nova Scotia Inst. Sci., 16(Pt. 2), 104-113, 1927.

Chapter 2.4: North America

73

Martin, S. D., Register of Meteorological Observations, Pine Grove, Kentucky, Smithsonian Institution, Washington, April 1865. Minshall, W. H., First dates of anthesis for four trees at Ottawa, Ontario, for the period of 1936 to 1945, Can. Field Nat., 61, 56-59, 1947. Mitchener, A.V., Nectar & pollen producing plants of Manitoba, Sci. Agric., 28, 475-480, 1948. Moss, E. H., Spring phenological records at Edmonton, Alberta, Can. Field Nat., 74, 13-118, 1960. Parry, D., W. J. A. Volney, and C. R. Currie, The relationship between trembling aspen phenology and larval development of the large aspen tortrix, Information report NOR-X30, Canadian Forest Service, Natural Resources Canada, Edmonton, 1997. Proceedings, Proc. and Trans. of the Royal Soc. of Canada, 10 (Session 3), 53-55, 1893. Reader, R., J. S. Radford, and H. L. Lieth, Modeling Important Phytophenological Events in Eastern North America, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 329-342, Springer-Verlag, New York, 1974. Russell, R. C., Phenological records of the prairie flora, Can. Plant Disease Survey, 42(3), 162-166, 1962. Schwartz, M. D., The Advance of Phenological Spring Across Eastern and Central North America, Ph.D. dissertation, Dept. of Geography, University of Kansas, Lawrence, 1985. Schwartz, M. D., Phenology and Springtime Surface Layer Change, Mon. Wea. Review, 120(11), 2570-2578, 1992. Schwartz, M. D., Monitoring global change with phenology: the case of the spring green wave, Int. J. Biometeorol., 38, 18-22, 1994. Schwartz, M. D., Green-wave phenology, Nature, 394(6696), 839-840, 1998. Schwartz, M. D., and T. M. Crawford, Detecting Energy-Balance Modifications at the Onset of Spring, Phys. Geography, 21(5), 394-409, 2001. Schwartz, M. D., and B. E. Reiter, Changes in North American Spring, Int. J. Climatology, 20(8), 929-932, 2000. Schwartz, M. D., G. J. Carbone, G. L. Reighard, and W. R. Okie, Models to Predict Peach Phenology from Meteorological Variables, HortScience, 32(2), 213-216, 1997. Smith, J. W., Phenological dates and meteorological data recorded by Thomas Mikesell between 1873 and 1912 at Wauseon, Ohio, Mon. Wea. Review Suppl., 2, 23-93, 1915. Vasseur, L., R. L. Guscott, and P. J. Mudie, Monitoring of spring flower phenology in Nova Scotia: comparison over the last century, Northeast. Nat., 8(4), 393-402, 2001. Vittum, M. T., and R. J. Hopp, The N.E.- 95 lilac phenology network, in Phenology, an aid to agricultural technology, Vt. Agric. Exper. Sta. Bull., 684, edited by R.J. Hopp, pp. 1-5, Vermont Agricultural Station, Burlington, 1978. Weather Bureau, Instructions for Voluntary Observers, U. S. Department of Agriculture, Washington, 1899. Zhao, T., and M. D. Schwartz, Examining the Onset of Spring in Wisconsin, Clim. Res., 24(1), 59-70, 2003.

Chapter 2.5 SOUTH AMERICA L. Patrícia C. Morellato Departamento de Botânica, Plant Phenology and Seed Dispersal Research Group, Universidade Estadual Paulista, São Paulo, Brasil

Key words:

1.

Phenological patterns, Flowering, Fruiting, Tropical, Climate zones

INTRODUCTION

Comprising about one eighth of the earth’s land surface, the South American continent is situated between 12°N-55°S latitude and 80°-35°W longitude. It covers an area of about 17,500,000 km2 divided between 13 countries. Eighty percent of its land is within the tropical zone, yet it extends into the subantarctic (Davis et al. 1997). Essentially, all life zones and vegetation formations are represented. The principal vegetation types are tropical evergreen and semi-evergreen moist forest, dry forest to woodland (cerrado or woody savanna), open grassy savanna, desert and arid steppe, Mediterranean-climate communities, temperate evergreen forest, and several montane formations (e.g. páramo, stone fields or campos rupestres, puna). The large array of vegetation types comprises some of the most diverse in the world. This includes the upper Amazon forest and Atlantic forest, as well as vegetation types with great concentrations of local endemism, the Andean montane forests and the Mediterranean-climate region of central Chile (Davis et al. 1997). At least 46 sites distributed over eight large regions have been recognized as centers of plant diversity (Davis et al. 1997), and several are considered biodiversity hotspots for conservation priorities (Myers et al. 2000).

Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 75-92 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

76

Phenology: An Integrative Environmental Science

All this diversity of species and vegetation types has not been completely studied in respect to its floristic diversity. Consequently, just a small percentage of its species and vegetation has been examined from the point of view of their seasonal changes. Phenological studies are uncommon, although the number of papers a published has increased over the last 20 years. However, long-term phenological data are rare and just a few longterm monitoring systems are known. I searched the major electronic databases (Web of Science, Current Contents, Scielo, Biological Abstracts), for phenological studies of South American native species or vegetation. Phenological data from studies on agricultural or introduced species or those from studies on herbivory, pollination, frugivory and seed dispersal were excluded when the main focus of the paper was the examination of animal feeding behavior. I also searched, using mainly some electronic and internet search tools and the database of South American main research agencies and institutions, for groups, universities, institutions or researchers doing phenological work. Finally, I looked for historical phenological data. However, this information was not easily found in the databases searched. The information is biased towards Brazil for two reasons. First, my 15 years of phenological research in Brazil, and second, the Brazilian Federal Agency for Science CNPq (National Counsel for Scientific and Technological Research) maintains a very well organized database of all active researchers and research groups in Brazil (http://www.cnpq.br). The main goal of this chapter is provide an overview of phenological work carried out on South American vegetation. I describe the phenological patterns of the main vegetation types studied up to the present, and point out areas where there is a lack of phenological information. To make descriptions more comparable, the overview is based largely on community studies that include information on flowering and fruiting patterns. I have tried to map the actual ongoing research and research groups and discuss the actual data collection to explore the future of phenological research on South America. As far as I am aware the phenology of South American vegetation has not been compared and analyzed from this perspective.

2.

BRIEF HISTORY OF PHENOLOGICAL DATA COLLECTION IN SOUTH AMERICA

The oldest phenological information surveyed was a description of the annual cycle of plants and animals in two Atlantic forests at Rio de Janeiro, Southeast Brazil (Davis 1945). Otherwise, the phenological data found, out of the electronic databases surveyed, were information included in

Chapter 2.5: South America

77

comprehensive papers describing the plant community (Veloso 1945; Andrade 1967), or just phenological notes, where the authors recorded the dates of flowering or fruiting for several tropical species from a particular site or botanical garden (Silveira 1935; Lima 1957; Santos 1979). Alvin (1964) was one of the first researchers to describe and analyze the phenology of native tropical forest trees from Bahia, Brazil (Alvin 1964; Alvim and Alvim 1976), although most of his work focused on the flowering of coffee and cocoa trees. The work of Araujo (1970) on the phenology of 36 species of Amazon lowland forest trees marks the beginning of the contemporary studies on phenology on South America. This paper is especially important because it represents the primary report of the first and oldest as well as possibly unique long-term phenological data collection for South America tropical forest trees. The INPA (Instituto Nacional de Pesquisas da Amazônia) phenological work started in 1965, at Reserva Florestal Ducke (Manaus, Amazonas State, Brazil). Trees were systematically selected with up to a total number of 300 trees (three per species) marked over an area of 140.5 hectares of native Amazon lowland tropical forest. In 1970 the sample size was extended to 500 trees (five per species), which are still monitored today. In 1974, INPA researchers replicated the phenology study. They marked 500 trees of another 100 species from an Amazon lowland forest at Estação Experimental de Silvicultura Tropical, about 30 km from Reserva Florestal Ducke. Both studies performed monthly observations for changes on reproductive and vegetative phenology of 10 defined phenophases (Araújo 1970; Alencar et al. 1979). A similar program of phenological data collection was established by Companhia Vale do Rio Doce (CVRD) at Reserva Natural de Linhares (Northeast Espírito Santo State, Brazil), a lowland evergreen forest reserve locally known as “Tabuleiro forest.” They employed the same methodology proposed by Araújo (1970), selecting 41 species and marking 205 trees (five per species). The project started in 1982 and seems to be active up to the present. Another interesting South American long-term database, although not active, is the one analyzed by Ter Steege and Persaud (1991). The authors compiled data on the flowering and fruiting of Guyanese forest trees collected over about 100 nonconsecutive years. I did not find any other long-term data set or project undertaking systematic phenological observations on South American vegetation. All information obtained refers to the collection of phenological data during a defined time span, usually between one to three years. Several Institutions and Universities conducting agronomic-related research have phenology programs for crops and some economically valuable species. For example, in Argentina, many of the National Universities have a discipline on

78

Phenology: An Integrative Environmental Science

climatology and agricultural phenology, and several plant species are investigated for phenological changes. Brazilian research institutions such as INPA, EMBRABA (Empresa Brasileira de Agropecuária) and IAC (Instituto Agronomico de Campinas) perform phenological observation of crop plants and native or exotic fruit trees. However, it was not possible to define the type of phenological data collected or their duration, and if some could be a long-term data set. Finally, I was unable to find any kind of phenology network, besides the GLOBE educational project, involving nine South American countries (for more information, see http://www.globe.gov). Therefore, there is a need for long-term phenological studies in South America. After Araujo’s (1970) work, just a few more papers were published in the 1970s (eight), even though the number increased during the 1980s (15 papers). The great production of phenological information for South American vegetation was in the 1990s (1990-1999), with at least 70 papers in the electronic databases surveyed. The papers are not evenly distributed over the 13 South American countries. Almost half of the production are from Brazil (30), followed by Venezuela (12) and Chile (12), and with some papers from Argentina (seven), Colombia (five), French Guyana (two) and Guyana (one). The number of published papers stays elevated from the year 2000 until the present, with about 36 papers surveyed in just three years, 22 of those from Brazil. Surprisingly, there are no recent papers from Chile, and Bolivia has two papers. Therefore, there is an increasing trend in publication of phenology studies in South America. If we considered other sources of information (books, journals not indexed, etc), the number of papers would be much higher (about twice) but the countries producing papers are nearly the same. A complete list of all papers surveyed can be obtained from the author.

2.1

Actual state of phenological research

Among the more than 150 papers researched, approximately half of them are phenology studies at the community level, describing the phenological patterns of several species belonging to one or different life-forms, from a defined geographic region and vegetation type. These papers are a good source of information for a large number of native species from different South American vegetation (Table 1). The other papers are studies focused on one or few species, or on a group of species belonging to the same plant family. The tendency is the same if just the papers produced recently, from year 2000 to date, are analyzed, with almost half focused on community studies. However, some of these recent studies start to address climatic

Chapter 2.5: South America

79

changes, relating satellite observation, “El Niño” effects and CO2 assimilation to plant phenology (Asner et al. 2000; Parolin 2000). Specifically for Brazil, I was able to locate at least 40 universities or institutions carrying out phenological research on almost every type of native vegetation across the country, from the Amazon forest around the equator to the Araucaria forest and subtropical vegetation of South Brazil. Some vegetation types have been the focus of researchers, such as the Amazon lowland forest or terra-firme forest, Atlantic forest, semideciduous forest and cerrado. However, in the last five years I found papers and ongoing projects focused on seasonal changes of plant species from caatinga, gallery forest, swamp forest, Amazon-inundated and seasonallyinundated forest, coastal plain vegetation and dunes, among others. Another interesting point is that almost any paper, project or ongoing research is concerned with the effects of climatic change and its evaluation using plant phenology.

3.

OVERVIEW OF PHENOLOGICAL PATTERNS FOR SOUTH AMERICAN VEGETATION

Summarizing the phenology of South American vegetation is difficult, due to the diversity of vegetation and species. To see how the phenological information is distributed over the different vegetation and to compare the phenological patterns observed, the land vegetation of South America was subdivided into six large vegetation groups (Table 1), plus the more complex montane formations occurring along the Andean cordillera, the Tepuis and in the coastal cordillera of Brazil (Davis et al. 1997). I only considered community studies including information on flowering and fruiting patterns, preferably for a large number of species. Studies that focused on just one plant family were excluded. Patterns are described based on the number of species flowering and fruiting per month unless otherwise noted. I. Tropical moist forest - In spite of the more or less non-seasonal climate and being an evergreen or semi-evergreen (semideciduous) forest, a marked seasonal pattern of flowering was observed for the majority of moist forests surveyed (Table 2). Amazon lowland evergreen forest species tend to flower during the dry season (Araújo 1970; Alencar et al. 1979; Sabatier 1985; Ter Steege and Persaud 1991; Peres 1994a; Wallace and Painter 2002, Figure 1). The same pattern is observed for Amazon “Campina” forest (Alencar 1990). Most species of semideciduous forest are flowering by the end of the dry season and beginning of wet season (Morellato et al. 1989; Stranghetti and Ranga 1997; Mikich and Silva 2001, Figure 1); Venezuelan semideciduous

80

Phenology: An Integrative Environmental Science

forest have species flowering in both seasons (Monasterio and Sarmiento 1976), the same is observed for climbers from Brazilian semideciduous Table 2.5-1. Number of published phenology studies at community level in the six large lowland vegetation groups of South America (adapted from Davis et al. 1997), plus montane formations. Vegetation group Plant formations Number of papers included I. Tropical moist forest Amazon Forest, 20 (evergreen or Brazilian Atlantic semi-evergreen Rain Forest, rain forest) Semideciduous forest, Gallery Forest and Swamp forest, Premontane/montane Pacific rain forest, Chocó and lower Magdalena Valley 9 II. Dry forest Chaco, Cerrado, (integrating into Caatinga, Woody woodland) Savanna Northern Colombia and Venezuela, coastal Equator and Peru, and the Deciduous (dry) forests III. Open grassy savanna Pampas region, the 2 Llanos, Cerrado grassland, Pantanal and Gran Sabana and Sipaliwini savanna in the Guyana region IV. Desert and arid step Sechura and 4 Atacama regions along west coast and in the Monte and Patagonian Steppes in the Southern Cone of South America V. MediterraneanCentral Chile 1 climate region

Chapter 2.5: South America

81

Vegetation group

Plant formations included

Number of papers

VI.

Temperate evergreen forest

Chile and Argentina

1

VII.

Montane formations

Complex montane formations along the Andean Cordillera, the Tepuis, and in the coastal cordillera of Brazil

5

forest (Morellato and Leitao 1996). Brazilian evergreen Atlantic rain forest and montane forest trees flower mainly in the wet season (Davis 1945; Morellato et al. 2000; Talora and Morellato 2000). Gallery forest shows a flowering pattern similar to Atlantic forest, flower peak occurring in the wet season (Funch et al. 2002). Two papers addressing swampy forest phenology were surveyed (Ramirez and Brito 1987; Spina et al. 2001). The swamp forest show flowering patterns similar to those observed for semideciduous forests (Spina et al. 2001). The Bolivian Sartanejal forest, a vegetation type influenced by forest streams, has a flower peak in the dry season (Wallace and Painter 2002). Fruiting patterns were more variable than flowering patterns across forest types and locations (Table 2, Figure 1). Some Amazon forests show seasonal fruiting patterns, peaking during the wet season (Alencarr et al. 1979; Sabatier 1985; Peres 1994a), while other present a bimodal pattern, with both peaks occurring in the dry season (Ter Steege and Persaud 1991). For Colombian lowland forests and Campina forest fruiting is non-seasonal (Alencar 1990; Wallace and Painter 2002). The fruiting patterns for semideciduous forests are not as seasonal as the flowering patterns, even though most species bear ripe fruits in the dry season or in the transition from dry to wet season (Morellato et al. 1989; Stranghetti and Ranga 1997; Mikich and Silva 2001). Fruiting is not seasonal for most of the Atlantic forest (Morellato et al. 2000), but a seasonal pattern is observed for Coastal Plain forest and montane forest, peaking in the less wet season (Davis 1945; Morellato et al. 2000; Talora and Morellato 2000). Gallery and Sartanejal forest present a fruiting peak in the wet season (Funch et al. 2002; Wallace and Painter 2002), and swamp forest fruiting was nearly non-seasonal (Spina et al. 2001). For the only Amazon floodplain forest surveyed the peak of flowering and fruiting occurs during the aquatic (flooded) phase (Schongart et al. 2002). Finally, Colombian Premontane rain forest presents a relatively constant number of species in flower or fruit through the year (Hilty 1980).

82

Phenology: An Integrative Environmental Science

Figure 2.5-1. Flowering (black line) and fruiting (gray line) patterns of Brazilian forest types: Amazon (AM, n = 36 species), data source Araujo (1970); Semideciduous (SD, n = 103), data source Morellato (1995); and Atlantic (AT, n = 214), data source Morellato et al. (2000). The dry season period is shown in gray.

Chapter 2.5: South America

83

II. Dry forest (integrating into woodland) – Savanna-forest mosaic on the Venezuelan Central Plain shows a marked seasonality of flower and fruit production for all habitats except forest (Table 2, Ramirez 2002). Flowering peaked during the rainy season, and fruiting peaked toward the end of the rainy season (Ramirez 2002). Woody savannas or cerrado vegetation comprises trees, shrubs and herbs. It includes at least two structural vegetation types: cerrado sensu stricto or woody savanna, an open landscape dominated by shrubs and trees at different densities, and Cerradão, a welldeveloped forest form of cerrado with canopy trees up to 20m tall (Ribeiro and Tabarelli 2002). Generally, the woody flora of cerrado present a seasonal phenology, although flowering and fruiting are not restricted to any particular season (Figure 2). Flowering is clearly associated with the end of the dry season and beginning of the rainy season (Monasterio and Sarmiento 1976; Batalha and Mantovani 2000; Oliveira and Gibbs 2000; Silberbauer-Gottsberger 2001; Ribeiro and Tabarelli 2002; Wallace and Painter 2002). Fruiting is more widespread over the year (Oliveira and Gibbs 2000), however, in some studies a middle-to-late rainy or dry season fruit peak is detected (Batalha and Mantovani 2000; Ribeiro and Tabarelli 2002; Wallace and Painter 2002). Herbaceous plants show a different pattern, flowering peak occurring at the end of the rainy season and fruiting peak in the dry season (Batalha and Mantovani 2000). Flowering peak in the Bolivian dry forest occurs at the transition between the dry and the rainy seasons. There is a major peak of fruiting in the dry season and a minor one during the rainy season (Justiniano and Fredericksen 2000). Caatinga, a deciduous tree-shrub vegetation from Northeastern Brazil, has a low rainfall climate, which is very seasonal and variable between years (Machado et al. 1997). Reproductive events are concentrated in the rainy season. The flowering peak occurs early and the fruiting peak late in the rainy period (Machado et al. 1997). Table 2.5-2. Main phenological patterns and peak season for South American vegetation. Patterns are ranked as seasonal or non-seasonal. The time of flowering and fruiting peak is indicated as dry season and wet season for tropical climates, and as spring, fall, winter or summer for temperate climates. Main phenological pattern Vegetation Type*

Group*

Flowering Pattern/peak season

Amazon lowland forest

I

Seasonal/dry

Amazon floodplain forest

I

Seasonal/flooding

Fruiting Pattern/peak season Seasonal to nonseasonal Seasonal/flooding

84

Phenology: An Integrative Environmental Science Main phenological pattern Vegetation Type*

Semideciduous forest Atlantic rain forest Gallery forest Swamp forest Sartanejal forest Pre-montane Rain forest Savanna-forest mosaic Cerrado (woody flora)

Group*

Flowering Pattern/peak season

I

Seasonal/dry-to-wet

I I I I I II II

Seasonal/wet Seasonal/wet Seasonal/dry-to-wet Seasonal/dry Non-seasonal Seasonal/wet Seasonal/dry-to-wet

Fruiting Pattern/peak season Seasonal/Dry or dryto-wet Non-seasonal Seasonal/wet Non-seasonal Seasonal/wet Non-seasonal Seasonal/wet Seasonal to nonseasonal Seasonal/dry Seasonal/rain Seasonal/dry Seasonal/after rain Seasonal/after rain

Dry forest II Seasonal/dry-to-wet Caatinga II Seasonal/rain III Seasonal/rain Campo cerrado Desert IV Seasonal/after rain Mediterranean-climate V Seasonal/after rain region Temperate Valdivian rain VI Seasonal/early summer Seasonal/late summer forest (dry) Chile Andean Zone (2000VII Seasonal/summer Seasonal/late 3600m altitude) summer-to-fall Montane grassland VII Seasonal/summer Seasonal/summer Pre-montane sub-tropical VII Seasonal/dry Seasonal/wet forest * See Table 1 and text for more detailed description of vegetation and source data.

III. Open grassy savanna – Few studies describe the phenology of open savanna. Dominant perennial grasses from Brazilian campo cerrado (open savanna) show a flower peak during the rainy season and fruit peak in the dry season (De Almeida 1995). A similar pattern is described for the Venezuelan Central Plain with fruiting more widespread over the year (Ramirez 2002). IV. Desert and arid step – Phenology studies were undertaken in four different places: arid Patagonia (Bertiller et al. 1991) and Monte Phytogeographical Province (Giorgetti et al. 2000) in Argentina, and the Chilean Coastal Desert and Southern Atacama Desert in Chile (Squeo et al. 1988; Vidiella et al. 1999). Phenology is highly constrained by rainfall, which determines the onset of reproduction for most of the species.

Chapter 2.5: South America

85

V. Mediterranean-climate region – The phenology of shrub species from the coastal desert in North-Central Chile indicates the existence of at least two groups of species, with phenological patterns more or less dependent on precipitation (Olivares and Squeo 1999).

Figure 2.5-2. Flowering (black line) and fruiting (gray line) phenology of woody flora in a Central Brazil cerrado community (n = 54 species), data source Oliveira and Gibbs (2000). The dry season period is shown in gray.

VI. Temperate evergreen forest – The timing of reproductive events and their ecological and climatic constraints in a Valdivian rain forest of Chiloé, Chile, one of the most widespread and species-rich forest types in austral South America, is discussed by Smith-Ramirez and Armesto (1994). Peak flowering for most species occurs in the dry season (late spring to early summer, Figure 3). Ripe fruits are available all year round, but the number of species is lowest in early spring with the maximum in late summer (Smith-Ramirez and Armesto 1994). VII. Montane formations – Includes a wide range of austral vegetation types, from distinct vegetation belts in the Andean zone to pre-montane subtropical forest. Phenological changes in flowering and fruiting of distinct vegetation belts in the Chile Andean zone (Andean scrub, cushion communities and fell-fiel) ranging from 2000 to 3600 m, have been described (Arroyo et al. 1979; Arroyo et al. 1981; Riveros 1983). The growing season lasts five to eight months, peak flowering coinciding with the period of maximum temperature at lower altitudes and after this period at higher altitudes, while fruiting peak takes place in the late summer and fall (Figure 4, Arroyo et al. 1981). Pre-montane subtropical forests in Argentina have a seasonal reproductive phenology (Malizia 2001). Flowers

86

Phenology: An Integrative Environmental Science

are present throughout the dry season and the number of species with ripe fruits peaks during the wet season (Malizia 2001). In Argentina montane grassland flowering and fruiting are concentrated in a short period during the summer (Diaz et al. 1994).

Figure 2.5-3. Percent of species (trees, shrubs, and miscellaneous - epiphytes, vines, and hemi parasites) flowering (black line) and fruiting (gray line) in the temperate rain forest of Chiloé, Chile, data source Smith-Ramírez and Armesto (1994). Summer is shown in gray.

Figure 2.5-4. Percent of species flowering (black lines) and fruiting (gray lines) at Andean scrub (AS), and Cushion communities (CC) sites (altitudinal Andean vegetation), data source Arroyo et al. (1981). Summer is shown in gray.

Additional studies - One paper from the Caribbean region describes flowering and fruiting phenology in tropical semi-arid vegetation of Northeastern Venezuela (Delampe et al. 1992). Venezuelan thorn woodland and thorn scrub desert formation show seasonality in their flowering and fruiting phenology (Delampe et al. 1992). Flowering activity is concentrated in the rainy season. Mature fruit index peaks in the dry season for trees and

Chapter 2.5: South America

87

tall shrubs, and is concentrated in the wet season for low shrubs and herbs (Delampe et al. 1992). Phenological patterns of Brazilian coastal dune vegetation were described by Cordazzo and Seeliger (1988). Most species of Southern Brazil coastal dune vegetation, under a warm temperate climate, flower during spring, summer and fall, while fruiting is concentrated in fall and winter (Cordazzo and Seeliger 1988).

4.

FUTURE DEVELOPMENTS AND CONCLUDING REMARKS

The distribution of phenological studies on South American vegetation is very uneven over the different vegetation types and life forms. Tropical forest is by far the better studied ecosystem and the tree is the life-form observed in almost all papers surveyed. By contrast a single study surveyed temperate forest (Smith-Ramirez and Armesto 1994). Studies focused on climbers, epiphytes and specially understory herbs are uncommon (Seres and Ramirez 1993; Putz et al. 1995; Morellato and Leitao 1996), although some studies include different life forms besides trees (Peres 1994a). If studies concentrated on just one phenophase, (those not included in Table 1) are considered, the number of papers on tropical forests is even higher. Several papers focused on fruiting patterns (Zhang and Wang 1995; Stevenson et al. 1998; Develey and Peres 2000; Grombone-Guaratini and Rodrigues 2002), and some on leafing behavior (Jackson 1978; Loubry 1994) or flowering (van Dulmen 2001). A number of studies on tropical and temperate forests are focused on just one family (Sist 1989; Peres 1994b; Riveros et al. 1995; Smith-Ramirez et al. 1998; Listabarth 1999; Henderson et al. 2000; Ruiz et al. 2000). Dry forests, savannas and cerrados are the second most studied vegetation group. Most of the studies were developed in savannas, and again the phenology of woody lifeforms dominates, but more studies include other lifeforms if compared to tropical forests. Very few studies were undertaken on open savannas, deserts, Mediterranean-climate region, and montane formations. Although a number of studies focused on one phenophase (Rozzi et al. 1989), or one or few species were surveyed (Silva and Ataroff 1985; Rusch 1993; Fedorenko et al. 1996; Löewe et al. 1996; Jaramillo and Cavelier 1998; Velez et al. 1998; Rosello and Belmonte 1999; Rossi et al. 1999; Damascos and Prado 2001). Therefore, there is a need for phenological studies on South America’s vegetation. The condition is quite critical if we consider that some vegetation types or regions have a high species diversity and an elevated number of endemic species. For example, the Mediterranean climatic region and Andean montane forest are basically unknown with respect to their seasonal

88

Phenology: An Integrative Environmental Science

patterns. Community studies should be undertaken, making it possible to understand the seasonal changes in those vegetation types. For the betterstudied tropical forests, investigations exploring different lifeforms are necessary. Long-term phenological observations are required in order to better comprehend the effects of climatic changes on plant phenology. Finally, there is a growing number of scientists interested in plant phenology. Building some phenology networks is the great challenge for South American phenologists, demanding an effort from and cooperation among universities, research institutions, governmental, and nongovernmental agencies.

ACKNOWLEDGMENTS I am grateful to Eliana Gressler for help in many ways during the literature and data survey, and V. B. Zipparro and A. Mantovani for assistance with the figures and literature. The author was supported by a research grant from FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo; # 95/09626-6), and received a research fellowship from the Brazilian Research Council (CNPq). The manuscript was improved by suggestions from E. Beaubien, M.R. Keatley, and M. A. Pizo.

REFERENCES CITED Alencar, J. d. C., Interpretação fenológica de espécies lenhosas de campina na Reserva Biológica de Campina do Inpa ao norte de Manaus, Acta Amazonica, 20, 145-183, 1990. Alencar, J. d. C., R. A. Almeida, and N. P. Fernandes, Fenologia de espécies arbóreas em floresta tropical úmida de terra-firme na Amazônia Central, Acta Amazonica, 9, 163-198, 1979. Alvim, P. T., and R. Alvim, Relation of climate to growth periodicity in tropical trees. in Tropical trees as living sistems, edited by Zimmermann, M. H., pp. 445-464, Cambridge University Press-London, 1976. Alvin, P. T., Periodicidade do crescimento das árvores em climas tropicais, Anais do XV Congresso da Sociedade Botânica do Brasil, 405-422, 1964. Andrade, M. A. B., Contribuição ao conhecimento da ecologia de plantas do litoral do estado de São Paulo, Boletim da Faculdade de Filosfia, Ciencias e Letras - Botânica, 22, 1-169, 1967. Araújo, V. C. d., Fenologia de essências florestais amazônicas I, Boletim do INPA - Série Pesquisas florestais, 1-25, 1970. Arroyo, M. T. K., J. Armesto, C. Villagran, and P. Uslar, High Andean Plant Phenology in Central Chile, Archivos De Biologia Y Medicina Experimentales, 12, 497-497, 1979. Arroyo, M. T. K., J. J. Armesto, and C. Villagran, Plant Phenological Patterns in the High Andean Cordillera of Central Chile, J. Ecol., 69, 205-223, 1981.

Chapter 2.5: South America

89

Asner, G. P., A. R. Townsend, and B. H. Braswell, Satellite observation of El Niño effects on Amazon forest phenology and productivity, Geophys. Res. Lett., 27, 981-984, 2000. Batalha, M. A., and W. Mantovani, Reproductive phenological patterns of cerrado plant species at the Pe-de-Gigante reserve (Santa Rita do Passa Quatro, SP, Brazil): A comparison between the herbaceous and woody floras, Revista Brasil. Biol., 60, 129-145, 2000. Bertiller, M. B., A. M. Beeskow, and F. Coronato, Seasonal Environmental Variation and Plant Phenology in Arid Patagonia (Argentina), J. Arid. Environ., 21, 1-11, 1991. Cordazzo, C. V., and U. Seeliger, Phenological and Biogeographical Aspects of Coastal Dune Plant Communities in Southern Brazil, Vegetatio, 75, 169-173, 1988. Damascos, M. A., and C. Prado, Leaf phenology and its associated traits in the wintergreen species Aristotelia chilensis (Mol.) Stuntz (Elaeocarpaceae), Rev. Chil. Hist. Nat., 74, 805815, 2001. Davis, D. E., The Annual Cycle of Plants, Mosquitoes, Birds, and Mammals in 2 Brazilian Forests, Ecol. Monogr., 15, 243-295, 1945. Davis, S. D., V. H. Heywood, O. Herrera-MacBride, J. Villa-Lobos, and A. C. Hamilton, Centres of Plant Diversity: a guide and strategy for their conservation, vol. 3, The Americas, IUCN Publications Unit, Cambridge, 562 pp., 1997. De Almeida, S. P., Phenological groups of perennial grass community on “campo-cerrado” area in the Federal District of Brazil, Pesqui. Agropecu. Bras., 30, 1067-1073, 1995. Delampe, M. G., Y. Bergeron, R. McNeil, and A. Leduc, Seasonal Flowering and Fruiting Patterns in Tropical Semiarid Vegetation of Northeastern Venezuela, Biotropica, 24, 6476, 1992. Develey, P. F., and C. A. Peres, Resource seasonality and the structure of mixed species bird flocks in a coastal Atlantic forest of southeastern Brazil, J. Trop. Ecol., 16, 33-53, 2000. Diaz, S., A. Acosta, and M. Cabido, Grazing and the Phenology of Flowering and Fruiting in a Montane Grassland in Argentina - a Niche Approach, Oikos, 70, 287-295, 1994. Fedorenko, D. E. F., O. A. Fernandez, C. A. Busso, and I. E. Elia, Phenology of Medicago minima and Erodium cicutarium in semi- arid Argentina, J. Arid. Environ., 33, 409-416, 1996. Funch, L. S., R. Funch, and G. M. Barroso, Phenology of gallery and montane forest in the Chapada Diamantina, Bahia, Brazil, Biotropica, 34, 40-50, 2002. Giorgetti, H. D., Z. Manuel, O. A. Montenegro, G. D. Rodriguez, and C. A. Busso, Phenology of some herbaceous and woody species in central, semiarid Argentina, Phyton-Int. J. Exp. Bot., 69, 91-108, 2000. Grombone-Guaratini, M. T., and R. R. Rodrigues, Seed bank and seed rain in a seasonal semi-deciduous forest in south-eastern Brazil, J. Trop. Ecol., 18, 759-774, 2002. Henderson, A., B. Fischer, A. Scariot, M. A. W. Pacheco, and R. Pardini, Flowering phenology of a palm community in a central Amazon forest, Brittonia, 52, 149-159, 2000. Hilty, S. L., Flowering and Fruiting Periodicity in a Premontane Rain-Forest in Pacific Colombia, Biotropica, 12, 292-306, 1980. Jackson, J. F., Seasonality of Flowering and Leaf-Fall in a Brazilian Sub-Tropical Lower Montane Moist Forest, Biotropica, 10, 38-42, 1978. Jaramillo, M. A., and J. Cavelier, Fenologia de dos especies de Tillandsia (Bromeliaceae) en un bosque montano alto de la Cordillera Oriental Colombiana, Selbyana, 19, 44-51, 1998. Justiniano, M. J., and T. S. Fredericksen, Phenology of tree species in Bolivian dry forests, Biotropica, 32, 276-281, 2000. Lima, D. A., Notas para fenologia da zona da Mata de Pernambuco, Revista de Biologia Lisboa, 1, 125-135, 1957.

90

Phenology: An Integrative Environmental Science

Listabarth, C., The palms of the surumoni area (Amazonas, Venezuela). II. Phenology and pollination of two flooded forest palms, Mauritiella aculeata and Leopoldinia pulchra, Acta Botanica Venezuelica, 22, 153-165, 1999. Löewe, V., C. Alvear, and F. Salinas, Fenología de E. globulus, E. nitens y E. camaldulensis en la Zona Central de Chile: Estudio preliminar., Ciencia e Investigación Forestal, 10, 7398, 1996. Loubry, D., Phenology of Deciduous Trees in a French-Guianan Forest (5 Degrees Latitude North) - Case of a Determinism with Endogenous and Exogenous Components, Can. J. Bot.-Rev. Can. Bot., 72, 1843-1857, 1994. Machado, I. C. S., L. M. Barros, and E. Sampaio, Phenology of caatinga species at Serra Talhada, PE, northeastern Brazil, Biotropica, 29, 57-68, 1997. Malizia, L. R., Seasonal fluctuations of birds, fruits, and flowers in a subtropical forest of Argentina, Condor, 103, 45-61, 2001. Mikich, S. B., and S. M. Silva, Composição florística e fenologia das espécies zoocóricas de remanescentes de floresta estacional semidecidual no centro-oeste do Paraná, Brasil., Acta Botanica Brasilica, 15, 89-113, 2001. Monasterio, M., and G. Sarmiento, Phenological strategies of plants species in the tropical savanna and semi-deciduous forest of the Venezuelan Lianos., J. Biogeography, 3, 325356, 1976. Morellato, L. P. C., As estações do ano na floresta. in Ecologia e preservação de uma floresta tropical urbana, edited by Morellato, L. P. C. and H. F. Leitão-Filho, pp. 37-41, Editora da Unicamp, Campinas, 1995. Morellato, L. P. C., R. R. Rodrigues, H. F. Leitão-Filho, and A. C. Joly, Estudo comparativo da fenologia de espécies arbóreas de floresta de altitude e floresta mesófila semidecídua na Serra do Japi, Jundiaí, São Paulo., Revista Brasileira de Botânica, 12, 85-98, 1989. Morellato, L. P. C., D. C. Talora, A. Takahasi, C. C. Bencke, E. C. Romera, and V. B. Zipparro, Phenology of Atlantic rain forest trees: A comparative study, Biotropica, 32, 811-823, 2000. Morellato, P. C., and H. F. Leitao, Reproductive phenology of climbers in a Southeastern Brazilian forest, Biotropica, 28, 180-191, 1996. Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. Fonseca, and J. Kent, Biodiversity hotspots for conservation priorities, Nature, 403, 853-858, 2000. Olivares, S. P., and F. A. Squeo, Phenological patterns in shrubs species from coastal desert in north-central Chile, Rev. Chil. Hist. Nat., 72, 353-370, 1999. Oliveira, P. E., and P. E. Gibbs, Reproductive biology of woody plants in a cerrado community of Central Brazil, Flora, 195, 311-329, 2000. Parolin, P., Phenology and CO2-assimilation of trees in Central Amazonian floodplains, J. Trop. Ecol., 16, 465-473, 2000. Peres, C. A., Primate Responses to Phenological Changes in an Amazonian Terra-Firme Forest, Biotropica, 26, 98-112, 1994a. Peres, C. A., Composition, Density, and Fruiting Phenology of Arborescent Palms in an Amazonian Terra-Firme Forest, Biotropica, 26, 285-294, 1994b. Putz, F. E., G. B. Romano, and N. M. Holbrook, Comparative Phenology of Epiphytic and Tree-Phase Strangler Figs in a Venezuelan Palm Savanna, Biotropica, 27, 183-189, 1995. Ramirez, N., Reproductive phenology, life-forms and habitats of the Venezuelan Central Plain, Amer. J. Botany, 89, 836-842, 2002. Ramirez, N., and Y. Brito, Patterns of Flowering and Fructification in a Swampy Community, Morichal Type (Calabozo, Edo Guarico, Venezuela), Acta Cientifica Venezolana, 38, 376381, 1987.

Chapter 2.5: South America

91

Ribeiro, L. F., and M. Tabarelli, A structural gradient in cerrado vegetation of Brazil: changes in woody plant density, species richness, life history and plant composition, Journal of Tropical Ecology, 18, 775-794, 2002. Riveros, M., Andean Plant Phenology, Volcan Casablanca, 40-Degrees-S, X- Region, Chile, Archivos De Biologia Y Medicina Experimentales, 16, R181-R181, 1983. Riveros, M., B. Palma, S. Erazo, and S. Oreilly, Phenology and Pollination in Species of the Genus Nothofagus, Phyton-Int. J. Exp. Bot., 57, 45-54, 1995. Rosello, N. E., and S. E. Belmonte, Fenologia de Browningia candelaris (Meyen) Britt. et Rose en la Quebrada de Cardones, Norte de Chile, Idesia, 47-55, 1999. Rossi, B. E., G. O. Debandi, I. E. Peralta, and E. M. Palle, Comparative phenology and floral patterns in Larrea species (Zygophyllaceae) in the Monte desert (Mendoza, Argentina), J. Arid. Environ., 43, 213-226, 1999. Rozzi, R., J. D. Molina, and P. Miranda, Microclimate and Flowering Periods on Equatorial and Polar- Facing Slopes in the Central Chilean Andes, Rev. Chil. Hist. Nat., 62, 75-84, 1989. Ruiz, A., M. Santos, J. Cavelier, and P. J. Soriano, Phenological study of Cactaceae in the dry enclave of Tatacoa, Colombia, Biotropica, 32, 397-407, 2000. Rusch, V. E., Altitudinal Variation in the Phenology of Nothofagus-Pumilio in Argentina, Rev. Chil. Hist. Nat., 66, 131-141, 1993. Sabatier, D., Fruiting Periodicity and Its Determinants in a Lowland Rain- Forest of FrenchGuyana, Rev. Ecol.-Terre Vie, 40, 289-320, 1985. Santos, N., Fenologia, Rodriguesia, 31, 223-226, 1979. Schongart, J., M. T. F. Piedade, S. Ludwigshausen, V. Horna, and M. Worbes, Phenology and stem-growth periodicity of tree species in Amazonian floodplain forests, J. Trop. Ecol., 18, 581-597, 2002. Seres, A., and N. Ramirez, Flowering and Fructification of Monocotyledons in a Venezuelan Cloud Forest, Rev. Biol. Trop., 41, 27-36, 1993. Silberbauer-Gottsberger, I., A hectare of cerrado. II. Flowering and fruiting of thick- stemmed woody species, Phyton-Ann. REI Bot., 41, 129-158, 2001. Silva, J. F., and M. Ataroff, Phenology, Seed Crop and Germination of Coexisting Grass Species from a Tropical Savanna in Western Venezuela, Acta Oecologica-Oecologia Plantarum, 6, 41-51, 1985. Silveira, F. R., Queda de Folhas, Rodriguesia, 1, 1-6, 1935. Sist, P., Structure and Phenology of the Palm Community of a French Guiana Rain-Forest (Piste-De-St-Elie), Rev. Ecol.-Terre Vie, 44, 113-151, 1989. Smith-Ramirez, C., and J. J. Armesto, Flowering and Fruiting Patterns in the Temperate RainForest of Chiloe, Chile - Ecologies and Climatic Constraints, J. Ecol., 82, 353-365, 1994. Smith-Ramirez, C., J. J. Armesto, and J. Figueroa, Flowering, fruiting and seed germination in Chilean rain forest myrtaceae: ecological and phylogenetic constraints, Plant Ecol., 136, 119-131, 1998. Spina, A. P., W. M. Ferreira, and H. d. F. Leitão Filho, Floração, frutificação e síndromes de dispersão de ma comunidade de floresta de brejo na região de Campinas (SP). Acta Botanica Brasilica, 15, 349-368, 2001. Squeo, F. A., N. Olivares, and F. Espinoza, Studies of Plant Phenology in the Chilean Coastal Desert, Iv Region, Archivos De Biologia Y Medicina Experimentales, 21, R334-R334, 1988. Stevenson, P. R., M. J. Quinones, and J. A. Ahumada, Annual variation in fruiting pattern using two different methods in a lowland tropical forest, Tinigua National Park, Colombia, Biotropica, 30, 129-134, 1998.

92

Phenology: An Integrative Environmental Science

Stranghetti, V., and N. T. Ranga, Phenological aspects of flowering and fruiting at the Ecological Station of Paulo de Faria-SP-Brazil, Tropical Ecology, 38, 323-327, 1997. Talora, D. C., and L. P. C. Morellato, Fenologia de especies arboreas em floresta de planicie litoranea do sudeste do Brasil, Revista Brasil. Bot., 23, 13-26, 2000. Ter Steege, H., and C. A. Persaud, The Phenology of Guyanese Timber Species - a Compilation of a Century of Observations, Vegetatio, 95, 177-198, 1991. van Dulmen, A., Pollination and phenology of flowers in the canopy of two contrasting rain forest types in Amazonia, Colombia, Plant Ecol., 153, 73-85, 2001. Velez, V., J. Cavelier, and B. Devia, Ecological traits of the tropical treeline species Polylepis quadrijuga (Rosaceae) in the Andes of Colombia, J. Trop. Ecol., 14, 771-787, 1998. Veloso, H. P., As comunidades e as estações botânicas de Teresópolis, RJ., Boletim do Museu Nacional do Rio de Janeiro, 3, 3-95, 1945. Vidiella, P. E., J. J. Armesto, and J. R. Gutierrez, Vegetation changes and sequential flowering after rain in the southern Atacama Desert, J. Arid. Environ., 43, 449-458, 1999. Wallace, R. B., and R. L. E. Painter, Phenological patterns in a southern Amazonian tropical forest: implications for sustainable management, For. Ecol. Manage., 160, 19-33, 2002. Zhang, S. Y., and L. X. Wang, Comparison of 3 Fruit Census Methods in French-Guiana, J. Trop. Ecol., 11, 281-294, 1995.

Chapter 2.6 THE GLOBAL PHENOLOGICAL MONITORING CONCEPT Towards International Standardization of Phenological Networks Ekko Bruns1, Frank-M. Chmielewski2, and Arnold J. H. vanVliet3 1

Department of Networks and Data, German Meteorological Service, Offenbach, Germany; Subdivision of Agricultural Meteorology, Institute of Crop Sciences, Faculty of Agriculture and Horticulture, Humboldt-University, Berlin, Germany; 3Environmental Systems Analysis Group, Department of Environmental Sciences, Wageningen University, Wageningen, The Netherlands 2

Key words:

1.

Global, Monitoring, Network, Observation program, Fruit trees

BACKGROUND AND OBJECTIVES OF GPM

In the 1990s interest in phenological research and thus demand for phenological observations has increased substantially. Mainly, rising air temperatures in recent decades and the clear phenological response of plants and animals to this increase have caused the growing interest. Many studies have shown that the timing of life cycle events is able to provide a good indicator for climate change impacts (Schwartz 1994; Menzel and Fabian 1999; Chmielewski and Rötzer 2001, 2002). Furthermore, the potential use of these data in other fields like remote sensing (to calibrate and evaluate NDVI satellite information) has added value to phenological data (Reed et al. 1994; Carleton and O'Neal 1995; Schwartz 1999; Schwartz and Reed 1999; Tucker et al. 1999; Chen et al. 2000). So, climate researchers have accepted the values of phenological data, and this renewed interest has increased demand for international cooperation in this area. In 1991, this Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 93-104 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

94

Phenology: An Integrative Environmental Science

demand was illustrated by a quote in the Proceedings of the International Conference on Climatic Impacts onn the Environment and Society: “ is necessary for all of us to consider an establishment of a global “It phenological observation network for monitoring of changing climate and its impact to ecosystem” (University of Tsukuba, Ibaraki, Japan January 27–February 1, 1991). The plans for establishing a new global phenological monitoring network were started by the “Phenology “ Study Group” of the International Society for Biometeorology (ISB) at a 1993 meeting in Canada. The objectives of the Phenology Study group were: – To promote a global dialogue among phenologists, by compiling information on phenological research and databanks, – To use this global forum to encourage establishment and expansion of phenological networks, data exchange, and international cooperation, – To encourage research that correlates phenological trends with climatic trends, especially in the context of global change monitoring, – To explore methods of using phenology to stimulate public interest in science, especially among pupils and students. At a second meeting in May 1995 (hosted by the German Meteorological Service in Offenbach), the Phenology Study Group drew up concrete benchmarks that facilitated network implementation. In 1996, the preparations of a Global Phenological Monitoring program (GPM) were completed at the 14th ISB Congress in Ljubljana, Slovenia. Phenologists from all over the world discussed the set-up of GPM. They agreed that the establishment of a Global Phenological Monitoring program was an important tool to meet the objectives of the ISB Phenology Study Group. A main objective of GPM is to form a global standard phenological backbone that can link “local” phenological networks and encourage establishment and expansion of phenological networks throughout the world. GPM can actively increase cooperation. Furthermore, data generated by GPM provide a basis for communication, research, and public relations.

2.

CONSTRUCTION AND SET UP OF GPM

During the design of the GPM program a number of details had to be considered, including the following issues: – What climate-biosphere relations should GPM address? – Which areas of the earth should be covered by GPM? – What species should be included in the monitoring program?

Chapter 2.6: The Global Phenological Monitoring Concept

95

– How should the GPM gardens obtain their plants? – What specific site conditions could be tolerated? – How should the observations instructions be formulated? Each of these questions is examined in more detail in the following sections.

2.1

Climate-Biosphere Relations and Geographical Focus in GPM

The timing of phenological phases depends on numerous environmental conditions: temperature, precipitation, soil type, soil moisture, and insolation. However, in mid- and high latitudes, with vegetation-rest (dormancy) in winter and an active growing period in summer, air temperature has the greatest influence on phenology (Fitter et al. 1995; Sparks et al. 2000; Chmielewski 2002). This is especially true for spring phenological phases (Figure 1).

Figure 2.6-1. Beginning of Forsythia suspensa flowering (right axis, yearly dates: dashed gray line with dot symbols, ten year running mean: solid black line, 59-year mean: dashed black line) at Hamburg (Lombardsbrücke 53º33’N, 10º00’E, 10m elev.), 1945-2003, and mean temperature over the three months before the beginning of flowering date (left axis, ten year running mean: solid gray line) at Hamburg-Fuhlsbüttel (53º38’N, 09º59’E, 16m elev.).

96

Phenology: An Integrative Environmental Science

Therefore, GPM focused mainly on temperature impacts on the timing of life cycle events. The influence of temperature is not quite so pronounced for autumnal phases (Estrella 2000). In the arid and semiarid tropics and subtropics, phenology is mainly driven by precipitation, because in these regions air temperature is never a limiting factor. Thus the global network will be mainly restricted to mid-latitudes (about 35° north to the Arctic Circle, and the Tropic of Capricorn to 50° south).

2.2

Selection of Species

The selection of plants is an important factor in determining the success of the monitoring program. A number of criteria were used to choose species: – Plants should have phenological phases that are easy to recognize and observe; – The start of the phases should be sensitive to air temperature; – Plants should be economically important; – Plants should have a broad geographic distribution and/or ecological amplitude; – Plants should be easy to propagate and vegetative propagation of these plants should be common practice; – The whole set of phenophases from the selected plants should cover with flowering stages as many months as possible during the growing season. Based on these criteria 14 species were selected for the GPM-program (Tables 1 and 2). These species consist mainly of fruit trees (specified varieties), some park bushes, and spring flowers. The fruit species represent the so-called “Standard Program”, which is required for each GPM-garden that will be established. The “Standard Program” can be supplemented by the “Flowering Phase Program” (ornamental shrubs and snowdrops) to obtain the “Maximum Program”. Due to different environmental conditions it is not possible to have all plants in the program at all stations in mid- and high latitudes.

2.3

Supply of Plant Species: GPM-Parent Gardens

A global network for plant observations depends, among other things, on the quality of observation objects. Unhealthy plants will disturb the measurements. Furthermore, since genetic differences can have a profound influence on the timing of life cycle events, a mechanism must be in place to guarantee the plant’s genetic identity. The best option is to work with one or several so called “parent gardens”, which are specialized in growing plants, and which are able to distribute the plant material. In 1996, the “Müller“

Chapter 2.6: The Global Phenological Monitoring Concept

97

Platz” nursery in Germany was engaged for this task, and now acts as parent garden for Europe. In the future, parentt gardens need to be established in other regions of the world like Asia and North America.

2.4

Site and Planting Conditions

Although temperature is the main forcing factor affecting plant development, other environmental factors also play a role. Therefore, to improve data analysis, a number of requirements for phenological garden site conditions were specified to standardize the monitoring program. With the focus on temperature, precipitation impacts were excluded by allowing irrigation in case of extreme water shortage. Another requirement was that the location should be characteristic of the larger region around the observation area. Sites are to be avoided which, due to specific sun exposure (e.g., southern slope), shady side, topographical conditions, (e.g., frost hollow), or urban development, are known to have climatic anomalies, or where deviations from characteristic conditions can be expected. The plants should be planted on level ground (slopes of up to 3 degrees in all directions are still acceptable). The trees and shrubs do not have to be planted in a specified order. The optimum growing site is open ground without obstacles, traffic routes, detrimental (for example, herbivory) or favorable influences (for example, artificial light), or other factors affecting the plants Table 2.6-1. Standard GPM-Observation Program and minimum distances between plants. Species Variety Rootstock Minimum Tree support distance required? Almond Perle der St. Julien A 3.0 while taking Weinstrasse root Red currant Werdavia own-rooted 1.5 none Sweet cherry Hedelfinger GiSelA 5 3.0 while taking root Morello Vladimirskaja own-rooted 3.0 while taking root Pear Doyenne de OHF 333 3.0 while taking Merode root Malus Apple Yellow 2.5 permanent Transparent transitoria Apple Golden M26 3.0 permanent Delicious European Dore de Lyon seedling detached while taking chestnut root

98

Phenology: An Integrative Environmental Science

(shading). As such conditions are certainly not always met; minimum standards were defined (Tables 1 and 2). The minimum distances (Tables 1 and 2) are only valid when plants have been placed taking into account the direction in which the different species of the GPM program are set relative to each other. Larger distances are desirable and consequently not an issue. Table 2.6-2. Flowering Phase GPM-Observation Program and minimum distances between plants. Species Variety Rootstock Minimum Tree support distance required? Witch hazel Snowdrop Forsythia Lilac Mock-orange Heather Heather Witch hazel

Jelena (genuine) Fortunei Red Rothomagensis (genuine) Allegro Long White (genuine)

own-rooted own-rooted

2.5 1.5 2.5

no no no

own-rooted own-rooted own-rooted own-rooted

3.0 0.5 0.5 2.5

no no

If the observed plants are located near obstacles the following issues apply. The minimum distance from the base of any obstacle (building, tree, wall, etc.) should be at least 1.5 times the height of the obstacle (more, two times, from the edge of forested areas). The distance from a two-lane road should be at least 8 m, and from any larger (eight-lane) highway, at least 25 m. All plants must be protected against herbivory (consumption by wild or domestic animals) by a wire-netting fence or individually by an anti-game protective agent. So-called “plant protection covers” (e.g., tube protection and growth covers) are unsuitable, as they can accelerate growth considerably (heat congestion). Thus, preference should be given to wirenetting systems.

2.5

Observation Instructions

Clear and understandable observation instructions help observers accurately monitor the plants and improve the quality of observations. GPM observers are asked to monitor the different phases of each species variety on only one plant. The other plants of the same variety serve to check the observation results, as well as being a reserve in case of loss. Thus, if a plant fails, another is ready to be used without any loss in the data quality. During the main growing season when temperatures are favorable, plants

Chapter 2.6: The Global Phenological Monitoring Concept

99

may develop at a tremendous rate. In order to obtain the exact date of the beginning of a phase, observations should be made at least 3-4 hours after the sun has passed zenith (midday). This helps to eliminate the possibility that phase onsets were missed during previous plant development. 2.5.1

Definition of phases

Phenological phases are recorded according to a BBCH1 code, which classifies plant growth phases of a lot of species according to a standardized system. The BBCH scale is an internationally recognized standard in the agricultural sector, and is thus an excellent source of standardized guidelines (the BBCH system is explained more in detail in Chapter 4.4). BBCH codes are available for all cultivated plants with economic importance. Consequently, the phases for apples, pears, cherries, and currants can be compared directly with their appropriate scales. For species that are not explicitly considered in a specific BBCH scale, the general BBCH scale can be used, which allows determination of phenological phases for all plants according to the standardized BBCH code. The following descriptions of the phenological phases are a complement to the BBCH definitions. They are somewhat more “traditional” than the short BBCH definitions, giving more detailed descriptions (illustrations of the phases by means of either photos or sketches are included in GPM2 web pages and literature). The descriptions here and definitions in the BBCH monograph3 should in no way contradict each other. Ultimately, the BBCH definitions are to be used. SL = Sprouting of leaves ((bud break: BBCH 07, bud burst: BBCH 53): The buds begin to open in at least 3 places on the object under observation. In the case of flower buds (bud burst) the green leaf tips enclosing flowers are visible; in the case of leaf buds (bud break) the first green is visible. UL = Beginning of the unfolding of leaves (BBCH 11): In at least 3 places on the object under observation, first leaves have pushed themselves completely out of the bud or leaf sheath and have unfolded completely, so that the leaf stalk or leaf base is visible. This phase is sometimes only recognizable by bending back the young leaf. The individual leaf has taken on its ultimate form, but has not yet reached its ultimate size. BF = First flowers open, Beginning of flowering/blossom (BBCH 60): In at least 3 places on the object under observation the first flowers have opened completely. Exceptions:

100

Phenology: An Integrative Environmental Science

Snowdrops (Galanthus nivalis): the first 3 flowers have opened at the plantation. The flower is considered open only when the outer leaves have spread and the stamens are visible. Heather (Calluna vulgaris): on 3 places of the plantation the first flowers have opened completely. FF = Full flowering, General flowering, Full blossom (BBCH 65): Approximately 50% of the flowers are open. EF = End of flowering/blossom (BBCH 69): This phase occurs when the flowers have faded. In some existing networks “flowers have faded” is equated with “approximately 95% of the total petals have fallen”. This rule is somewhat different in formulation to the BBCH69 definition but in practice “de facto” identical. RP = Fruit ripe for picking (for apple, pear, sweet cherry, morello, red currant, BBCH 87): The fruits show the coloring characteristic for their variety and can be removed easily from the fruiting lateral. Exception: Premature ripening should not be reported. RP = First ripe fruits (for almond, European chestnut, BBCH 864): The first ripe fruits fall from the tree naturally. Exception: Premature ripening should not be reported. CL = Coloring of leaves (BBCH 944): Approximately 50% of the leaves have taken on the colors of autumn. Coloring of leaves, caused by drought, should nott be reported. FL = Leaf falll (BBCH 95): Approximately 50% of the leaves have fallen off. Falling of leaves, caused by drought, should not be reported.

3.

ESTABLISHMENT OF THE GPM NETWORK

There are two ways to establish the Global Phenological Monitoring network: setting up new gardens, or adapting existing networks to the proposed standardization. In recent years, both approaches have been pursued.

3.1

Setting Up New GPM Gardens

The first GPM network garden was started in 1998 at Deuselbach (Germany). It is located at a measuring station of the Federal Environmental Agency. Further gardens quickly followed (at the beginning only in Germany), but now also in other countries of the northern hemisphere. The current network includes 15 gardens located in Asia,

Chapter 2.6: The Global Phenological Monitoring Concept

101

Europe, and North America (Table 3). More gardens are required before the data can be effectively analyzed. At the moment the data from the German GPM-stations are gathered at the Humboldt-University of Berlin. In the near future decisions need to be made on how the network will be administrated in terms of data storage and access. Two existing networks give an idea of the number of stations required for acquiring observational data from genetically homogeneous plants, i.e. the European IPG network5 and the lilac/honeysuckle network6 in the USA. Both networks currently consist of approximately 50 sites and both networks do not cover all of their respective continents. Based on the experience of these networks (and other factors like region, climate, and altitude gradation), we propose at least 75 stations for Europe. It will be an especially effective network if the stations are optimally distributed between Gibraltar and the Ural Mountains. Numbers of needed stations for other continents have yet to be assessed. Table 2.6-3 Established GPM gardens Number of sites Country China Estonia Germany

1 1 9

The Netherlands Slovakia USA

2 1 1

3.2

Locations Beijing Jögeva Blumberg, Brunswick, Deuselbach, Erbeskopf, Geisenheim, Linden Schleswig, Tharandt, Zingst Amsterdam, Wageningen Banska Bystrica Milwaukee

Adaptation of Existing Networks: Linking Networks

The second way of establishing the GPM network is by adapting existing networks into the new network. In the last years, the GPM program expanded because several existing networks added some GPM plants to their own program. In 2001, the “Red Rothomagensis” lilac variety (also used in the USA) and the “Fortunei” forsythia variety were incorporated into the International Phenological Gardens program (from the GPM program). At the same time the first “link gardens” were laid out in Schleswig, Deuselbach and Tharandt (Germany). These are gardens in which both the IPG and the GPM assortments are planted. The link between IPG and GPM will continue, and the present three combined IPG/GPM gardens (as of 2002) will be followed by others. In autumn 2000, the Wageningen Agricultural University distributed bulbs of the snowdrop clone, which is

102

Phenology: An Integrative Environmental Science

also contained in the GPM program, to 700 observers in the Netherlands. In 2002, a standardized phenological garden plan was introduced into the GLOBE program for schools (http://www.globe.gov). The suggested garden consists of the “Flowering Phase Program” of GPM, as these stages are easy for students to observe. Thus, schools around the world can now help to extend the GPM program, fulfilling one of the aims of the original ISB Phenology Study Group: “To stimulate public interest in science, especially among pupils and students.” Finally, some countries have considered using concepts from the GPM program to set up their own national networks. Scientists from Peking University (Beijing) would like to build up a phenological observation network in botanical gardens across China, in which GPM will play a central role. At present the GPM assortment is being propagated at the Beijing Botanical Garden. If these network plans are successful, it will be the first time that a national organization has adopted the full GPM program. By standardizing observations, it becomes possible to link the different networks. Standardization can be applied to the species included in the programs, to the stand of the observation objects (for example, solitary plants or a stand of forest/woodland) as well as to the observation area (for example, maximum distance from the reference point), even to the object (for example, the same individual year to year) and to the definitions used for phenological stages. In recent years, progress has been made in standardizing definitions for phenological stages in Europe, based on European Phenology Network (EPN) activities. EPN has applied BBCHmethodology to the definitions used in twelve phenological networks in Europe so far. This analysis has made it possible to identify to what extent the existing networks are compatible among each other and with the GPM program. In addition to the EPN standardization activities, several other developments have contributed to these efforts. For example, the Meteorological and Hydrological Service of Croatia modified their phenological guidelines in 1996, orienting them more towards the German program, which necessarily meant higher compatibility with BBCH. In 2000, the Central Institute for Meteorology, Austria, proceeded in the same way, and with the same effect. The Dutch phenological network (revived in February 2001) also modeled itself around the German program, so that Dutch plant phases are in complete agreement with those of Germany, and (due to the phase selection) are almost completely in agreement with BBCH. The Canada Plantwatch program was expanded in 2002 and an instruction booklet was published. More plant species were added and phenophases modified to better match European protocols and the BBCH system. MeteoSwiss also has new instructions, which are nearly identical to the

Chapter 2.6: The Global Phenological Monitoring Concept

103

German program. In all, six European networks now work de facto according to compatible rules, where phenological phases overlap. In the Swiss guide the phenological phases are compared to the corresponding BBCH codes for the first time in national instructions (Brügger and Vassella 2003), and that is at least the intention of the Slovak Hydrometeorological Institute in 2003. This list is not complete, but documents the tendency toward greater standardization, which is not limited to Europe, but also Before the applies to the North American continent, and China. development of the BBCH scales (in the 1990s) and prior to GPM, there were no internationally recognized standards, apart from Zadoks et al.’s (1974) cereal grain scales.

4.

CONCLUSIONS

In recent years, the Global Phenological Monitoring network has steadily increased in size. Set-up issues have been thoroughly explored, and sites successfully implemented in different parts of the world. GPM has demonstrated that it can play a significant role in standardization of phenological networks, as the BBCH-coding system is being adopted by other phenological networks. The first phase of the GPM network also improved cooperation between groups all over the world, and formed the basis for several successful initiatives, such as reviving the Dutch phenological network and the European Phenology Network. GPM will continue to contribute to the further expansion of existing networks, and the establishment of new networks, both to improve the use of phenological information, and improve cooperation and communication between the many actors involved in phenology. The program is now poised for future expansion into other parts of the world. Hopefully, GPM will be just as successful in gaining acceptance from phenologists internationally, as BBCH has been in worldwide agricultural experiments.

NOTES 1

BBCH = Biologische Bundesanstalt, Bundessortenamt, Chemische Industrie (Federal Biological Research Centre for Agriculture and Forestry, Federal Office of Plant Varieties, Chemical Industry) 2 http://www.dow.wau.nl/msa/gpm/ 3 BBCH-Monograph, Blackwell Science, 622 pp., 1997. 4 The reference numbers BBCH 86 and BBCH 94 were defined for this purpose. They fit into the context and do not violate the BBCH principle.

104 5 6

Phenology: An Integrative Environmental Science

http://www.agrar.hu-berlin.de/pflanzenbau/agrarmet/ipg.html http://www.uwm.edu/~mds/enanet.html

REFERENCES CITED Brügger, R., and A. Vassella, Pflanzen im Wandel der Jahreszeiten, Geographica Bernensia, Bern, Switzerland, 287 pp., 2003. Carleton, A. M., and M. O'Neal, Satellite-derived land surface climate “signal” for the Midwest U.S.A., Int. J. Remote Sensing, 16, 3195-3202, 1995. Chen, X., Z. Tan, M. D. Schwartz, and C. Xu, Determining the growing season of land vegetation on the basis of plant phenology and satellite data in northern China, Int. J. Biometeorol., 44, 97–101, 2000. Chmielewski, F.-M., Trends in the seasons, Bull. Amer. Met. Soc., 10, 1464-1465, 2002. Chmielewski, F.-M., and T. Rötzer, Response of tree phenology to climate change across Europe, Agricultural and Forest Meteorology, 108, 101-112, 2001. Chmielewski, F.-M., and T. Rötzer, Annual and spatial variability of the beginning of growing season in Europe in relation to air temperature changes, Clim. Res., 19(1), 257264, 2002. Estrella, N., On modeling of phenological autumn phases, in Progress in phenology: Monitoring, data analysis, and global change impacts, edited by A. Menzel, p. 49, Conference abstract booklet, 2000. Fitter, A. H., R. S. R. Fitter, I. T. B. Harris, and M. H. Williamson, Relationships between first flowering date and temperature in the flora of a locality in central England, Funct. Ecol., 9, 55, 1995. Menzel, A., and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999 Reed, B. C., J. F. Brown, D. Vander Zee, T. R. Loveland, J. W. Merchant, and D. O. Ohlen, Variability of land cover phenology in the United States, J. Veg. Sci., 5, 703-714, 1994. Sparks, T. H., E. P. Jeffree, and C. E. Jeffree, f An examination of relationships between flowering times and temperature at the national scale using long-term phenological record from the UK, Int. J. Biometeorol., 44, 82–87, 2000. Schwartz, M. D., Monitoring global change with phenology: the case of spring green wave, Int. J. Biometeorol., 38, 18–22, 1994. Schwartz, M. D., Advancing to full bloom: planning phenological research for the 21st century, Int. J. Biometeorol., 42, 113–118, 1999. Schwartz, M. D., and B. C. Reed, Surface phenology and satellite sensor-derived onset of greenness: an initial comparison, Int. J. Remote Sensing, 20, 3451–3457, 1999. Tucker, C. J., D. A. Slayback, J. E. Pinzon, S. O. Los, R. B. Myneni, and M. G. Taylor, Higher northern latitude normalized difference vegetation index and growing season trends from 1982 to 1999, Int. J. Biometeorol., 45, 191-195, 1999. Zadoks, J. C., T. T. Chang, and C. F. Konzak, A Decimal code for the growth stages of cereals, Weed Res., 14, 415-421, 1974.

Chapter 2.7 TOWARD A MULTIFUNCTIONAL EUROPEAN PHENOLOGY NETWORK Arnold J. H. vanVliet and Rudolf S. deGroot Environmental Systems Analysis Group, Department of Environmental Sciences, Wageningen University, Wageningen, The Netherlands

Key words:

1.

Network, European, Users, Communication, Cooperation

WHY A EUROPEAN PHENOLOGY NETWORK?

Phenology as a scientific discipline has a very long history. Many local, regional, and national networks exist (see Chapters 2.1-2.6), and the number of disciplines that deal with phenological processes in their own profession is large and diverse (see, for example, the diversity of topics in this book). The phenological community, however, faces a number of problems: – There is insufficient cooperation and communication between the existing regional and national phenological monitoring networks in Europe. – There is a lack of access to and integration of data. This is partly caused by the lack of information on what datasets a are available, the different definitions and techniques used, and the quality of the data. – There is inefficient use and exchange of existing knowledge within and between the different scientific disciplines on tools and techniques already available for monitoring, data storage, and data analysis. – There is insufficient insight in and awareness of the potential practical uses of phenological data. These problems lead to sub-optimal production of data (both in quantity and quality) and inefficient use of phenological information. The many possibilities for data use and production techniques are not fully exploited. Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 105-117 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

106

Phenology: An Integrative Environmental Science

More importantly, the low recognition of the multifunctional uses and socioeconomic values of phenological observations has led to a lack of financial support for existing phenological networks. At the same time, there is a strong increase in demand for phenological observations in the past decade, caused by the observed changes in climate and possible consequences for biodiversity and society (see Chapter 2.6). Integration of phenological networks can provide valuable input in large European initiatives like the Global Change and Ecosystems program of the European Commission Sixth Framework Program1, which aims to strengthen the capacity to understand, detect and predict global change. Other initiatives like the Global Monitoring of Environment and Security program (GMES) will also benefit from the input of phenological networks. To address the above-mentioned problems, a European Phenology Network (EPN) was set up in 2001. This chapter presents how this network is helping to improve communication and cooperation between the many disciplines involved, and in the production and use of phenological data. EPN is a Thematic Network (2001-2003), in which 13 partners2 from seven countries participate. The European Commission3 finances the network. EPN aims to “improve monitoring, assessment and prediction of climate induced phenological changes and their effects in Europe.” Its overall objective is to increase the efficiency, added value, and use of phenological monitoring and research, and to promote the practical use of phenological data in European member states in assessing the impact of global (climate) change, and possible adaptation measures. It realizes this objective by focusing on three different areas, which this chapter addresses. First, it demonstrates the variety of possible applications of phenological research and its benefits for biodiversity conservation and society (section 2). Second, it identifies the many different user groups involved (section 3). Third, it provides a number of tools that directly facilitated communication and cooperation (section 4).

2.

APPLICATIONS OF PHENOLOGY

The timing of life cycle events is a fundamental ecological process. In order to identify phenological data and actors that work on phenological issues, it is important to determine for which natural and socio-economic processes phenological information has relevance. Only then is it possible to improve the collection and use of phenological observations. The following sections give an overview of the range of applications of phenological data, demonstrating the diversity of subjects involved and the need and

Chapter 2.7: Toward a Multifunc. European Phenology Network

107

opportunities for cooperation and communication between the different actors.

2.1

Biodiversity Science

Biodiversity is defined by the Convention on Biological Diversity (CBD) as: “the variability among living organisms from all sources including, among other things, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diversity within species, between species and of ecosystems” (CBD 2001). To maintain biodiversity, plants and animals have to survive until they have reproduced. Thereby, the variability off the timing of life cycle events plays an important role in determining the reproductive success of plants and animals. During their life cycle, they face many abiotic and biotic lifethreatening factors that often only occur during certain times of the year. One of the best ways to cope with these threats is by not being there when they occur. Therefore, for many species, timing of migration, hiding, or transformation in another, less vulnerable, phase is essential. Just as it is important to avoid life-threatening situations, species should also be able to be active when there is enough food and water available for growth. In many cases, life-supporting resources of sufficient quality are not constantly available during the whole year and species have to adjust their timing to optimal periods. Especially in extreme abiotic conditions (like a very cold year), the total productivity and the quality of food (for example plant material) might be low. This directly affects the ability of other species to find enough resources. Furthermore, the ability of plants and animals to find enough resources also depends on the timing of life cycle events of competitors of individuals of the same species or of other species. When organisms have been able to reach the reproductive phase, the success of reproduction still depends to a large extent on timing. First of all, if reproduction takes place by mating of males and females, the timing of reproductive ability of both sexes should be synchronized. Furthermore, the timing of the appearance of juveniles should coincide with the presence of nutrients and water. Migratory birds, for example, have to produce young early enough to give the young sufficient time to gain enough strength to migrate to other areas. If they leave too late, the young will not survive. The above-mentioned examples illustrate the importance of understanding the causes and consequences of variation in timing of life cycle events for biodiversity science. Phenological studies significantly contribute to studies on productivity, reproduction and survival of individual organisms and contribute to studies on the consequences for a whole range

108

Phenology: An Integrative Environmental Science

of species interactions like competition, predation, parasitism, and mutualism.

2.2

Agriculture/Forestry/Fisheries

From the previous section, it is clearr that growth and reproduction of plants and animals are closely linked to abiotic and biotic factors that are strongly controlled and determined by the timing of life cycle events. Because agriculture, forestry, and fishery strongly depend on the productivity and reproduction of certain plant and animal species, these economic sectors have to clearly take due account of the timing of a large number of life cycle events. The length of the growing season, for example, determines the growth potential for crops and potentially the number of rotations within a year. It, however, also determines the amount of production loss caused by “extreme” weather events or by pests and diseases. An early start of the growing season in combination with a late spring frost event can, for example, cause significant damage to crops. Given the importance of the variation in timing of life cycle events for productivity, the agricultural, forestry, and fisheries sectors try to continuously adjust their management activities and techniques (like sowing/planting, harvesting, nutrient supply, pests and diseases control measures, and water supply) to the timing of life cycle events. Because of the long-term experience with the timing of life cycle events, the sectors mentioned in this section have the potential to significantly contribute to other disciplines that have to deal withh the study of phenological processes.

2.3

Human Health

The timing of life cycle events of plants and animals also directly influence many aspects of human health such as allergies, diseases, pests, and water quality. Many people are allergic to allergens that are connected to particles in the air, especially pollen. The timing of pollen release by plants determines the start and length of the pollen season, and thus affects the timing of taking precautions or the occurrence of illness (Huynen et al. 2003). The occurrence and timing of many diseases is also closely related to the timing of the appearance of other organisms, especially insects. The clearest examples are vector borne diseases like malaria. The distribution and occurrence of the insects (vectors) that are able to distribute these diseases strongly depend on environmental variables. In addition to diseases, many organisms and especially insects are considered to be pests that cause distress to people. For example, mosquitoes and ants cause many problems during the time they are active. In addition to cases where plants

Chapter 2.7: Toward a Multifunc. European Phenology Network

109

and animals cause direct impacts on human health, there are also many examples where plants and insects cause indirect impacts to human health by affecting agricultural production (quality and quantity), as mentioned in the previous section, leading to health problems.

2.4

Transportation

A good understanding of the variation of timing of life cycle events is also of relevance for transportation. For example, timing of bird migration directly influences the frequency of collisions between birds and airplanes. Another issue is the timing of leaf fall and thus the maintenance activities to remove leaves from roads and rail tracks. If the timing of the maintenance activity is not adjusted to the timing of leaf fall, significant delays in (public) transport can be the result.

2.5

Tourism and Recreation

Many people spend significant amounts of their free time in natural areas. In the Netherlands, for example, the yearly number of day excursions of more than two hours per day (without visits to family, or visits from holiday destinations) is 935 million (http://www.cbs.nl). More and more people adjust their vacation or short visits to nature reserves to the timing of certain life cycle events like coloring of leaves, appearance of migratory birds, or flowering of specific plants. By having a better understanding of the variation in and changes of the timing of life cycle events, the tourism sector should be able to better inform the public on when certain events take place. Based on phenological information, the sector will also be better able to prepare themselves for busy and/or quiet periods.

3.

PHENOLOGY USER GROUPS

As we have seen in the previous section, a large number of ecological and socio-economic sectors are affected by phenological events. In order to improve practical use of phenological data, communication, and cooperation, it is important to realize that within each sector a large number of different user groups are, or potentially can be, involved. In this section we briefly describe the main user groups (see Table 1). These user groups interact by exchanging money, knowledge, data, tools, public relation, and consumables to improve and “sell” their products (in this case phenological information and their applications).

110

Phenology: An Integrative Environmental Science

Effective communication and cooperation between user groups can only take place if they are aware of each other’s roles and objectives. In Table 1 an overview is given of how the different user groups interact through the main “products” they provide. Table 2.7-1. Interaction between user groups: Each cell contains an overview of the product that is provided by the user groups listed in the top row to a specific user group in the first column.

Data

Data provider

Research

Policy

NGOs

Commerce

Media

Public

Kno

Kno,

Fin

Fin, PR

Fin, Too

PR, Kno

Fin, Data

Fin, PR

Fin, PR,

Fin, Kno,

PR, Kno

Fin

Kno

Too

Kno, PR

Fin, Kno,

PR, Kno

Fin

Too, Fin

provider Research

Data

Kno, Too, Fin

Policy

Data

Kno, Too

Kno

Too NGOs

Data

Kno, Too

Fin

Kno, PR

Fin, Too

PR, Kno

Fin

Commerce

Data

Kno, Too

Fin

Kno, PR

Fin, Kno,

PR, Kno

Fin

Kno

Fin

Kno

Data, Kno

Too Media

Data

Kno, Too

Fin,

Kno

Kno Public

Data

Kno, Too

Kno

Fin, Kno, Too

Kno

Kno, Too, Con

Fin = Financial support; Kno = Knowledge: information about processes, problems, and solutions; Data: quantitative and qualitative representation of phenological events and applications; Too = Tools that can be used to find, analyze, and exchange data, or knowledge or to facilitate communication; PR = Public Relations: information on activities of the network of people and organizations included in phenological data collection, analyses and application; Con = Consumables (e.g., food, medicines): the production of which is influenced in quality and/or quantity by phenological information.

3.1

Data Providers

Data providers record at what date and time a predefined (phenological) activity of a specific species takes place in a specific year at a specific location. Data providers should constantly be aware of the specific demands of the users of the data and assess whether they still provide the data that the users need and in what format the users want to receive the data. Researchers will undoubtedly have other demands than the media. Phenological monitoring networks should also be aware of the fact that the same information can be used in many different ways. For almost all user groups the quality of the observations is important and increases if the number of observations and the geographical coverage is increased. Phenological networks have demonstrated that involvement of the public is

Chapter 2.7: Toward a Multifunc. European Phenology Network

111

an important help in realizing this objective, provided that the observation technologies applied are standardized. The usefulness of the data also increases if the observations are easy and quickly accessible (see for example the Nature’s Calendar network in the United Kingdom and the Dutch network De Natuurkalenderr4).

3.2

Research

Research is essential to develop and improve tools for monitoring, data analysis, modeling, forecasting, and decision support. Research is also important to interpret the outcome and application possibilities of phenological observations, for example: – What (a)biotic factors determine the timing of life cycle events; – What is the impact of (changes in) the timing of life cycle events on ecological and socio-cultural processes; – How can we control or influence the timing in such a way that we can mitigate adverse effects or better benefit from it. The knowledge and tools provided by the research community have relevance for all user groups because many of the underlying processes related to timing are universal. The research community adds content and relevance to the original phenological observations and makes the data easier to use. In many cases, researchers can apply the tools developed for one sector to other sectors. For example, ecological models developed to assess the start of flowering of tree species are of direct relevance for those users that assess the start of pollen release in the context of human health. Communication between scientific communities is required to improve efficient use and exchange of knowledge and tools.

3.3

Policy Makers

Policy makers are usually specialized by theme or sector, each with their own objectives: for example, environmental policy (to protect the environment); nature policy (to protect nature); and human health (to provide good health care). To monitor the effects of policy decisions and the degree to which certain objectives are achieved, phenological information is essential. In addition to the thematic specialization, it is also important to realize that policy makers address issues at different scales: local, regional, national, or international level. This has implications for the type of information they require and the actions that they will undertake to obtain this information, and the willingness to provide financial support for data collection, analyses, and dissemination.

112

3.4

Phenology: An Integrative Environmental Science

Non Governmental Organizations

The objective of Non-Governmental Organizations (NGOs) is to inform society on issues for which they (i.e., their constituency) think that the government or industry is not paying enough attention. In order to do this, NGOs gather information on the issues and disseminate this knowledge to user groups that can undertake action to prevent or solve the problems. NGOs often gather knowledge themselves or they support individuals or organizations in their activities. This support can be financial but also through public relations to raise attention for the issue.

3.5

Commerce

Commercial products can be consumables (e.g., food or wood produced by agriculture, fisheries, and forestry). Phenological data and knowledge can significantly improve the quality and/or efficiency of the production process of consumables, as explained in section 2. Phenological information helps to understand and deal with processes like crop growth, frost damage, and pests and diseases, and also carbon sequestration. Another commercial sector of relevance is the Information and Communication Technology (ICT) sector that develops tools and techniques (e.g., database and communication technologies). Commercial companies can provide the tools and techniques that can be used for gathering, storage, analysis, and exchange of phenological data and information. The different commercial sectors mentioned can also use phenology as a public relation tool since many people and organizations are interested in phenology. By working together with phenological networks and by applying phenological information they can show that they care about the environment by improving their production process (for example, use of phenological information to reduce pesticide use) or by supporting activities that are seen as a general public interest.

3.6

Media

The aim of the media is to provide the public with information and news in which the public (or specific target groups) is interested, so that people are willing to pay for their news, overviews, or listen/look at their programs. Only then, they can attract advertisements or government support and thus money to continue to exist. Phenology has proven to be an interesting subject for the media as it can closely meet the requirements of different media (newspapers, television, radio). As competition for attention in the

Chapter 2.7: Toward a Multifunc. European Phenology Network

113

media is strong, providers of phenological information should deliver the information in a way that it can easily be incorporated into news items.

3.7

General Public

Phenology can provide non-commercial information that the public can use to improve their welfare. With access to phenological information, people will become better aware of their own surroundings, both the things that happen and the actions taken. With phenological information, people will become better aware of the many interactions between changes in, for example, weather and climate and the effects on nature and their own welfare (for example human health). With this information, people will be able to better understand causes and effects of their personal behavior and the consequences of corporate and government decisions.

4.

TOOLS FOR IMPROVING COOPERATION AND COMMUNICATION

An important aim of the European Phenology Network is to provide the phenological community, with its many different sectors, with a common platform for better cooperation and communication. A variety of activities contribute to achieving the goals: international meetings, databases, standardization, and ICT-technologies and media.

4.1

International Meetings

During the EPN project (2001-2003), international conferences and workshops played an important role in bringing people together from many different networks, sectors, and user groups who normally would not meet each other. These meetings facilitated direct exchange of knowledge, data, and tools and techniques. These meetings discussed the changes in timing that are going on as a result of the observed change in climate, and focused on demonstrating the many application possibilities of phenological information. In addition to two conferences, several specialist workshops, attended by up to 30 people, provided the opportunity to discuss several main subjects in more detail. Six themes were selected: bird migration, agriculture, human health, modeling, remote sensing, and communication, dissemination, and capacity building (reports are available at the EPN website, http://www.dow.wau.nl/msa/epn/).

114

4.2

Phenology: An Integrative Environmental Science

Databases

Well-structured databases that are easily accessible and up to date are essential to facilitate the exchange of information on what organizations and products are available. The European Phenology Network developed two on-line freely available databases: a meta-database and a bibliographical database. The meta-database aims to provide easy access to information on what phenological networks exist or have existed in the past. Formerly, it was unclear what networks existed or who was the contact person for a specific network. This, of course, hampered cooperation, communication and effective use of the data. The bibliographical database aims to provide an overview of the publications that exist related to phenology. References in this database contain publications on: botany, zoology, ornothology, entomology, geography and history, and also agronomy, forestry, environment and medical sciences (Jeanneret 1997). Each of these disciplines has its own publication forum, which makes publications difficult to find. Furthermore, much fundamental work was done a long time ago and is no longer accessible. By bringing together these scientific disciplines in one database, scientists are able to find existing expertise that they might need more quickly and more efficiently.

4.3

Standardization

Use of data from different networks for making large-scale analyses of, for example, climate change impacts, is only possible if the data are in the right format. Because the history of the various networks is often quite different, there are contrasting standards and procedures used. For example, definitions of how to determine the start of leaf unfolding could differ from network to network. Standardization has to be applied in order to make cooperation and exchange of data possible. The European Phenology Network provided a standardization key to harmonize and interpret the different definitions used for phenological phases (see Chapter 4.4).

4.4

ICT Technologies and Media

Very promising tools for the phenological community are the recent developments in ICT and cooperation with the media. A few phenological networks have started to develop on-line information systems4. These systems have been very successful as they provide observers the ability to enter their phenological observations and to directly provide a detailed

Chapter 2.7: Toward a Multifunc. European Phenology Network

115

visualization of all observations made within the network. The system provides “live” maps, tables, and graphs of the changes that occur in natural systems. Although EPN has not been directly involved in the development of these tools, its partners2 are. The information systems have proven to be a very (cost) efficient way to run phenological networks with thousands of active observers. These tools provide new ways to meet the information and data requirements of many users.

5.

CONCLUSION AND DISCUSSION

The objective of the EPN project was to increase the efficiency, value, and use of phenological monitoring and research, and to promote the practical use of phenological data in Europe. To achieve these objectives, EPN provided a systematic overview of the large number of application possibilities of phenological information, made contact with the many data providers and user groups involved, and improved communication tools. The two International Conferences held in 2001 and 2003 and the six workshops significantly increased the potential for cooperation within the phenology-community in Europe and beyond. In this chapter, we have shown that phenological information is valuable to a large number of environmental and socio-economic sectors. Data, knowledge, and techniques gathered for one sector often has high relevance for other sectors. Models that assess the date of insect appearance, for example, have relevance for agriculture, forestry, ecology, and human health. Therefore, when people involved in one sector expand their network, they can increase the importance of, and interest in their work, and provide their product to more customers. This chapter also highlighted that the groups of people that are (potentially) interested in phenological data, information, and technologies is very diverse, even within one sector. We identified many different user groups and demonstrated that each of these groups has its own objectives and role in the phenological community. Often, these different groups can benefit from each other since they all have their own specialization. There is a large potential for increased cooperation, both within and between different user groups. To improve collaboration, it is very important that actors (stakeholders) specify and communicate their needs (for example, data, information, tools) as well as their own products. Ideally, each actor group should carry out a stakeholder analysis that identifies potentially interested collaborators and products that are available. Based on the stakeholder analysis in this chapter we would like to emphasize the important role of data providers in the phenological

116

Phenology: An Integrative Environmental Science

community. Without data, there will be no research that can show the importance of phenology, which will reduce the interest in phenology, both from the public and (commercial) users and which will make it difficult to maintain existing and develop new networks. In addition, communication tools provide the “lubricant” of the network. People and organizations must be able to easily and quickly exchange their data, knowledge, and techniques. From our analyses we are convinced that the phenological community has an enormous potential to grow in the future and contribute significantly to environmental monitoring programs and applications. However, this can only be realized if the actors involved increase their internal and external communication and cooperation through better use of existing facilities and investment in the further development of new instruments and active participation in networks.

NOTES 1

Sixth Framework Program website: http://www.cordis.lu/fp6/ EPN Partners are: Environmental Systems Analysis Group, Wageningen University, the Netherlands (coordinator); German Weather Service, Germany; SME-Milieuadviseurs (GLOBE-the Netherlands), the Netherlands; Department Data & Computation, Potsdam Institute for Climate Impact Research, Germany; Institute of Geography, University of Berne, Switzerland; Lehrstuhl für Bioklimatologie, Technical University Munich, Germany; Institute for Environment and Sustainability, Unit Land Management, Joint Research Center, Italy; Centre for Geoinformation, Wageningen University, the Netherlands; International Center for Integrative Studies, the Netherlands; World Health Organization, Italy; Department of Agricultural Sciences, Royal Veterinary and Agricultural University, Denmark; Centre for Ecology and Hydrology Monks Wood, United Kingdom, International Center for Environmental Assessment, Foundation for Sustainable Development, the Netherlands. 3 Energy Environment and Sustainable Development Program, subsection “Better Exploitation of existing data and adaptation of existing observing systems.” 4 Nature’s Calendar in the UK (http://www.phenology.org.uk/) and Natuurkalender in The Netherlands (http://www.natuurkalender.nl). 2

REFERENCES CITED Convention on biological diversity (CBD), Convention text, Secretariat of the Convention on Biological Diversity, United Nations Environmental Program, 2001.

Chapter 2.7: Toward a Multifunc. European Phenology Network

117

Huynen, M., B. Menne, H. Behrendt, R. Bertollini, S. Bonini, R. Brandao, C. BrownFahrlaender, B. Clot, C. D'Ambrosio, P. de Nuntiis, K.L. Ebi, J. Emberlin, E. Erdei Orbanne, C. Galan, S. Jaeger, S. Kovats, P. Mandioli, P. Martens, A. Menzel, B. Nyenzi, A. Rantio-Lehtimaeki, J. Ring, O. Rybnicek, C. Traidl-Hoffmann, A. vanVliet, T. Voigt, S. Weiland, and M. Wickman, Phenology and human health: allergic disorders, Health and Global Environmental Change Series No.1., EUR/o3/5036791, Rome, Italy, 55 pp., 2003. Jeanneret, F., International bibliography of phenology, Institute of Geography, University of Berne, Switzerland, 68 pp., 1997.

PART 3

PHENOLOGY OF SELECTED BIOCLIMATIC ZONES

Chapter 3.1 TROPICAL DRY CLIMATES Arturo Sanchez-Azofeifa1, Margaret E. Kalacska1, Mauricio Quesada2, Kathryn E. Stoner2, Jorge A. Lobo3, and Pablo Arroyo-Mora4 1

Earth and Atmospheric Sciences Department, University of Alberta, Edmonton, Alberta, Canada; 2Centro de Investigaciones en Ecosistemas, Universidad Nacional Autónoma de México, Morelia, México; 3Biology Department, Universidad de Costa Rica, San Jose, Costa Rica; 4Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT, USA

Key words:

1.

Tropical, Remote Sensing, Dry Forest, Conservation, Land cover change

INTRODUCTION

Based on the Holdridge life zone system (Holdridge 1967) approximately 111,599,269 km2 around the world have a climate favorable for dry forest (Leemans 1992, Figure 1). Of that area, 94% is located in the tropics. Tropical dry forests are found between the two parallels of latitude, the Tropics of Cancer and Capricorn (23º27’ N and S) where there are several months of little or no precipitation (Holdridge 1967). In general, three to seven month’s dry season duration has been quoted for seasonally dry forests (Janzen 1983; Murphy and Lugo 1986; Luttge 1997; Piperno and Pearsall 2000). The tropical dry forest ecosystem is one of the most fragile and least protected ecosystems in the world. In general, Neotropical dry forests are less species rich than moist or wet forests. For example, 430 species of woody plants have been recorded in the wet forest of La Selva Biological Station, Costa Rica (Hartshorn and Hammel 1994), while in the dry forest of the Santa Rosa National Park, Costa Rica, 160 species (51 families) have been inventoried (Kalácska and Sánchez-Azofeifa, unpublished data). In addition, Gentry (1995) reports a Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 121-137 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

122

Phenology: An Integrative Environmental Science

range of 21–121 species (nine and 41 families) from various 0.1ha plots around the Neotropics. However, there is more structural (e.g., wood specific gravity) and physiological (e.g., growth seasonality) diversity in the plant life forms of dry forests than in wet forests (Medina 1995).

Figure 3.1-1. Areas around the world with a climate favorable for supporting a dry forest ecosystem. Spatial resolution: 0.5 degrees latitude by 0.5 degrees longitude (modified after Leemans 1992).

Tropical forests that once formed a continuous habitat across Mesoamerica and some regions of the Pacific and Atlantic regions of South America, are now found in fragmented patches (Whitmore and Sayer 1992; Heywood et al. 1994; Trejo and Dirzo 2000). Tropical deforestation is likely to affect both biotic and abiotic factors that control the phenological expression of plant communities with severe consequences to plant populations and the communities that interact or depend on them (Cascante et al. 2002; Fuchs et al. 2003). However, fortunately in certain regions of the Neotropics such as in Costa Rica, the secondary forests are in a state of regeneration through which the dry forests are also starting to recuperate (Arroyo-Mora 2002). Both savannahs and dry forests (T-df) can co-occur in areas with the same climate, but the dry deciduous forests have a a tendency to be found in areas with greater soil fertility (Ratter et al. 1973; Mooney et al. 1995). In many areas however, the occurrence of either savannah or dry forest is principally controlled by human disturbance (Maass 1995; Menaut et al. 1995). Due to the favorable climatic conditions in which they are found, tropical dry forests have been heavily exploited for agriculture (Ewel 1999). Piperno and Pearsall (2000) argue that historically, tropical wet and dry forests had completely different associations with human activities. They state that the

Chapter 3.1: Tropical Dry Climates

123

deciduous and semi-evergreen forests (especially the T-df) were the locations of the majority of the early human settlements in addition to being the home of the wild ancestors of many crop plants as well as the origin of animal husbandry. Even today, in most tropical countries, the majority of the agriculture and pasturelands are located in areas that used to be dry and moist forest. This pattern of both higher population density as well as higher intensity food production in the T-df as compared to wetter life zones may be a reflection of the historical tendency for humans to settle these areas (Piperno and Pearsall 2000). It appears that the tuberous plants that are rich in starch for human consumption seem to be more common in the seasonal forests, since the tuber is developed in part for energy storage during the dry season. The long dry season aided in burning of the vegetative cover in order to prepare the fields for agriculture (Piperno and Pearsall 2000). In addition, weeds and pests are less aggressive in the drier environments (Murphy and Lugo 1986). Tropical dry forest phenology is an area that is still in its early stages of academic discovery, since historically more emphasis has been placed on tropical evergreen forests, especially the Amazon Basin (Luttge 1997). Therefore, there is a need for continuous and systematic efforts to understand its phenological patterns and integrate its phenological mechanisms at two basic levels: 1) In the context of conservation biology and 2) the context of land use and land cover change that are taking place on this rich agricultural frontier. In this chapter we document different aspects related to leaf phenology in the tropical dry forest ecosystem and its implications for satellite remote sensing. Emphasis is placed on presenting a description of the causes of leaf phenological change in this threatened ecosystem, and how these can be linked with conservation biology and land use/land cover change at the regional level.

2.

CAUSES OF PHENOLOGICAL CHANGE

Several studies have indicated that the phenological expression of leaves, flowers and fruits are affected by biotic and abiotic factors. Abiotic factors such as changes in water level stored by plants (Reich and Borchert 1984; Borchert 1994, but also see Wrightt and Cornejo 1990; Wright 1991), seasonal variations in rainfall (Opler et al. 1976), changes in temperature (Ashton et al. 1988; Williams-Linera 1997) photoperiod (Leopold 1951; Tallak et al. 1981; van Schaik 1986), irradiance (Wright and van Schaik 1994) or sporadic climatic events (Sakai et al.1999), have been proposed as the main causes of leaf production or leaf abscission in tropical dry forest plants. In contrast, biotic factors, such as competition for pollinators or

124

Phenology: An Integrative Environmental Science

pollinator attraction (Robertson 1895; Janzen 1967; Gentry 1974; Stiles 1975; Appanah 1985; Murray et al. 1987; Sakai et al. 1999; Lobo et al., in press), competition for seed dispersers, and avoidance of herbivory (Marquis 1988; Aide 1993; van Schaik et al. 1993) have been considered as the factors regulating the intensity and duration of leaf and flower production. The abiotic and biotic factors are not mutually exclusive, and it is likely that several are interacting to regulate the expression of each phenological phase. In tropical dry forests, apart from foliage seasonality, relationships between water availability and the structural and physiological characteristics such as hydraulic architecture or sensitivity to water stress produce a variety of phenological behaviors (Murphy and Lugo 1986; Bullock 1995; Holbrook et al. 1995; Luttge 1997, among others). However, due to the remote sensing component of this chapter, the discussion in this section will be largely limited to phenological changes in leaf cover, even though most studies that evaluate the effect of abiotic and biotic factors on the phenological expression of tropical plants (both dry and wet forests) have mainly studied leaf, but not flower or fruit phenology. One of the major causes of the leaf phenological patterns (as mentioned above) in all tropical dry forest is the length of the dry season. This difference may be partly responsible for the differences in physical characteristics such as canopy height or biomass. Apart from leaf phenology, the length of the dry season and the seasonality of precipitation are also important for evolutionary adaptations of gene and seed dispersal, which are distinct in dry forests from the wet forests. In general, in dry forests most trees have conspicuous flowers and wind-dispersed seeds. Dry forests also have a lower biomass and a smaller stature than wet forests (Gentry 1995). Two other main factors that influence leaf phenological patterns are edaphic associations and topography since they determine the spatial heterogeneity of the available water (Murphy and Lugo 1995). Water stress can vary at both regional and local scales. This variability induces a multitude of tree life forms with different leaf phenological patterns (Mooney et al. 1995). At the regional scale, the structure of the forest is greatly affected. It has been shown that as water availability decreases, so does the number of canopy stories as well as the horizontal continuity of the canopy (Murphy and Lugo 1995). Figure 2 compares the climate diagrams (Walter 1971) for fourteen sites from different life zones, ranging from wet to dry forests in Costa Rica as well as the dry forest in Chamela, Mexico. The mean monthly temperature (ºC) and the monthly precipitation (mm) are scaled to represent the potential evapotranspiration. Dry months are represented by dotted areas, humid months by the vertical lines, and months with rain in excess of 100 mm are in solid black. Differences in the severity of the dry season as well as the pattern of the

Chapter 3.1: Tropical Dry Climates

125

rainfall can cause the different leaf phenological patterns observed at various sites. For example, in Guanacaste, Costa Rica, Gentry (1995) estimates that 40-60% of the tree species are deciduous whereas over 70% are deciduous in Chamela, Mexico, where the severity of the dry season is more pronounced (Figure 2).

Figure 3.1-2. Climate diagrams for fourteen representative sites in Costa Rica and Chamela, Mexico.

The general leaf phenological response to the dry season is drought deciduousness where the woody plants lose their leaves in the dry season, but there are exceptions (wet season deciduous, Fanjul and Barradas 1987) as well as dry evergreen forests and evergreen succulent plants in dry forests (Gentry 1995; Holbrook et al. 1995). Occasional anomalous rains in the dry season and drought spells in the wet season complicate this variation in resource availability in the rainy season. The growing periods are thus affected by the variability in flushing as it occurs in response to anomalous rains in the dry season or variation in the drying out process (Murphy and Lugo 1995). In a comparison between wet (La Selva) and dry (Comelco) sites in Costa Rica, Frankie et al. (1974) found that the forest at La Selva maintained its evergreen appearance throughout the year. However, even this wet forest experienced increased leaf flushing with the onset of the wet

126

Phenology: An Integrative Environmental Science

season. In Comelco they found that while leaf fall began as early as October, the majority of the trees lost their leaves in the dry season, with the peak in leaf fall occurring in March. Of the 113 species they inventoried at Comelco, 83 partially or completely lost their leaves and 19 were evergreen (ex. Clusia rosea, Styrax argenteus). The trees in the Riparian zones lost their leaves, but were simultaneously replaced. One species, Lysiloma seemannii had an unusual leaf-flushing pattern in that after it lost its leaves in the dry season, the new leaves did not appear until one month after the rainy season began. Certain species also brought new leaves in January and March but most of these species (for example, Anacardium excelsum, Coccoloba padiformis) were from the Riparian zones. In total, Frankie et al. (1974) found that 75% of the species are affected by the seasonality of the precipitation in the dry forest, compared to 17% in the wet forest. The timing of leaf flushing was also found to be very different: in the wet forest, most of the leaves were produced in the dry season, whereas in the dry forest the leaves were produced at the beginning of the wet season.

3.

PHENOLOGY AND CONSERVATION BIOLOGY

In this section, we review the literature and present some of the main consequences that change or disruption of plant phenology may have on the viability of plant populations and animal communities that interact with them. Biotic factors, such as competition for f pollinators or pollinator attraction have been interpreted as important adaptive forces responsible for phenological patterns in tropical plants (Robertson 1895; Janzen 1967; Gentry 1974; Stiles 1975; Appanah 1985; Murray et al. 1987; Zimmerman et al. 1989; Sakai et al. 1999; Lobo et al., in press). A disruption of the flowering phenological patterns caused by disturbance or fragmentation is likely to affect the behavior and visitation rate of pollinators. Fragmented landscapes reduce the amount of resources available, as well as appropriate areas for roosting and perching for nectarivorous bats and birds that serve as important pollinators for many tropical plant species (Feinsinger et al. 1982; Andren and Angelstran 1988; Bierregaard and Lovejoy 1989; Rolstad 1991; Saunders et al. 1991; Helverson 1993; Quesada et al. 2003). If the flowering pattern of plants that share pollinators of the same guild is displaced over time (Frankie et al. 1974; Stiles 1975; Fleming 1988), competition for the same pollinators will occur, resulting in negative consequences for the reproductive success of the plants and the ability of the pollinators to obtain resources over time. For example, in the tropical dry forest of the ChamelaCuixmala Biosphere Reserve, Mexico, trees of the family Bombacaceae

Chapter 3.1: Tropical Dry Climates

127

provided the main resource to the nectarivorous bats Leptonycteris curasoae during eight months and Glossophaga soricina during six months. Both bat species concentrated on one bombacaceous species each month (Stoner et al. 2003). The sequential use of bombacaceous species by these bats coincided with the flowering phenology of the tree species. These data suggest that changes in the flower phenology caused by habitat disruption may result in competition between these bat species and ultimately may result in local extinction, especially of endemic species that are common in this dry forest. A rare endemic nectarivorous bat that is only found in four states in Mexico, Musonycteris harrisoni, foraged on the bombacaceous tree Ceiba grandiflora during three months of the year (Stoner et al. 2002). Since this species has such a restricted distribution and is a specialist nectarivore, changes in flower phenology could be catastrophic for populations of this bat. Timing of leaf flushing directly affects insect herbivores that depend upon flushing species to complete part a of their life cycle (Janzen 1970, 1983; Dirzo and Dominguez 1995). Phenological changes caused by habitat loss will also disrupt the pollination patterns of many long-distance pollinators and trap-liners such as some large bees, hawkmoths, nectarivorous bats, and hummingbirds that have been shown to follow the flowering phenology of plants (Stiles 1977; Haber and Frankie 1989; Frankie et al. 1998; Fleming et al. 1993; Haber and Stevenson 2003). For example, in Costa Rica, hawkmoths regularly move from the lowland tropical dry forest to surrounding areas at higher elevations, following patterns of flowering resources (Haber and Stevenson 2003). Similarly, in México and the southwestern U.S. some nectarivorous bats have been shown to migrate following the availability of flower resources, mainly from the family Cactaceae and Agavaceae (Fleming et al. 1993). Intra-specific variation in the frequency, duration, amplitude and synchrony of flowering by individuals also has been proposed as an important factor that affects the reproduction and the genetic structure of tropical plant populations in disturbed r habitats (Murawski et al. 1990; Murawski and Hamrick 1992; Newstrom et al. 1994; Doligez and Joly 1997; Nason and Hamrick 1997). Flowering phenology directly determines the effective number of pollen donors and the density of flowering individuals, both of which affect the patterns of pollen flow between trees (Stephenson 1982; Murawski and Hamrick 1992). Plants with asynchronous flowering may experience a decrease in reproductive output, the amount of pollen, the number of pollen donors and the levels of outcrossing compared to individuals blooming during the same period. Fuchs et al. (2003) suggested that pollinator behavior is likely to change the mating patterns of P. quinata. This study showed that in disturbed fragmented habitats or in trees with

128

Phenology: An Integrative Environmental Science

early or late peak flowering, bat pollinators are more likely to promote selfing within trees (i.e., geitonogamy) and they have a tendency to produce singly sired fruits, whereas in undisturbed natural forests outcrossing is higher and multiple paternity is more common. The long-tongued bat (Glossophaga soricina), one of the main pollinators of P. quinata, has been shown to adopt a territorial behavior within a single plant in disturbed isolated environments with limited resources (Lemke 1984, 1985). The timing of fruiting during the year, which may be altered as a result of environmental changes associated with habitat disturbance, may affect potential vertebrate seed dispersers that, in turn, may affect the reproductive success of the plants they disperse (Flemming and Sosa 1994). Frugivorous Old World and New World bats are known to migrate or change habitats depending on the availability of fruit resources (Eby 1991; Stoner 2001). Similarly, the abundance of temperate and altitudinal migrant birds in tropical forests is closely associatedd with fruit abundance (Levey et al. 1994). Furthermore, displacement of fruiting phenology of tree species that are keystone resources because they provide fruits when resources are relatively scarce, could have negative consequences on populations of birds and mammals that disperse their seeds and ultimately negative effects on recruitment of the species they disperse (Howe 1984). Seed dispersal by animals is negatively affected by deforestation and results in lower recruitment in forest fragments. Another factor affected by forest fragmentation is seed predation. In a tropical dry forest seed predation by bruchid beetles on the tree Samanea saman was higher in populations of trees found in continuous forest and found to be much less in isolated trees (Janzen 1978). The bruchid beetles, Merobruchus columbinus and Stator limbatus (Bruchidae) are specific seed predators of S. saman. It is likely that the populations of these bruchid species are affected by density dependent factors related to the availability and fluctuation of food resources within fragments, including seeds and flowers. Another explanation is that adult bruchids have to fly greater distances to find isolated trees than trees in continuous populations. This pattern of higher seed predation in populations from continuous forest also has been observed in the dry forest tree, Bahuinia pauletia. Finally, the ultimate consequences of habitat reduction and phenological disruption is a decrease in reproductive plants, increasing the negative effects of endogamy, reducing the quantity and quality of pollen, and lowering the genetic variability of the progeny (Cascante et al. 2002). This likely will affect the viability and establishment of plant populations over time.

Chapter 3.1: Tropical Dry Climates

4.

129

PHENOLOGY AND LAND USE / COVER CHANGE

Remote sensing data provides the possibility for an instantaneous look at a large area with the opportunity of acquiring frequent repeat imagery for the same area. This is important for phenological studies because the temporal variability of the ecosystem can be captured at large scales. In particular, it is essential to consider leaf phenology in order to correctly characterize areas of deciduous forest. Since they measure surface reflectance, optical sensors have been widely used for land cover classification and characterization. However, it must be taken into consideration that one of the greatest limitations to optical sensors is cloud cover. And in the tropics, cloud cover is especially prevalent in the wet season. Cloud free imagery is more easily acquired during the dry season, where in deciduous forests, the majority of the trees are leafless (Arroyo-Mora 2002). In addition, vegetation studies using reflectance data have generally focused on green leaves, with both dry vegetation and non-green components being neglected in comparison (van der Meer 1999). However, in areas of deciduous forest green leaves will not always dominate the spectral signature of the forest. In the dry season, only a small fraction of the spectra will be representative of green foliage. The majority of the pixels will be representing leaf litter, bark, branches and soil in various combinations. Therefore, this temporal variability of the spectral signatures that can be extracted from imagery must be taken into consideration in such environments. As an example, two false color composite images of the same area of dry forest surrounding the Chamela Biological Station, Mexico, were acquired during the dry (March) and wet (August) seasons from the Landsat 7 ETM+ sensor (not shown). While the two images visually look completely different, more importantly, the spectral signature of the forest also changes with the seasons. This is key because many algorithms rely on spectral signatures to classify areas. If the same unsupervised classification algorithm (Isodata) is run on the two images, 180 km2 of forest cover is extracted from the wet season image, while only 26 km2 of land cover exhibits the spectral signature of forest in the dry season (Kalacska et al. 2001). In the dry season image, only the Riparian areas appear to have forest cover. In a similar case study from the Santa Rosa National Park, Costa Rica, two images (dry season – April and wet season – October) of Landsat TM 5 were classified using an unsupervised classification into forest and nonforest classes. From the wet season image, 61 km2 of forest were extracted, whereas from the dry season image only 18 km2 were classified as forest (Kalacska et al. 2001). The discrepancy in the amount of forest extracted

130

Phenology: An Integrative Environmental Science

from the images in the two seasons is because dry deciduous forests where trees lose their leaves, may seem to have the spectral signature of pasturelands or agricultural fields in the dry season (Figure 3).

Figure 3.1-3. Spectral signatures of the dry forest at the Santa Rosa National Park from Landsat TM 5 images. a) wet season (October) and b) dry season (April). Solid line deciduous forest, dashed line - evergreen forest.

In the wet season (Figure 3a), the spectra for both the evergreen and deciduous components of the forest are similar. However, in the dry season (Figure 3b), the spectral signature of the deciduous forest no longer resembles that of the evergreen forest. In fact, there is more than a 20% difference in the near infrared band (band 4) between the two forest classes in the dry season. While these results are important at a local scale, their implications become more profound if regional or global scales are considered. For example, Sader and Joyce (1988) reported the total forest cover for Costa Rica as 17%. If their map of forest distribution is examined, it can be seen that the province of Guanacaste and the Nicoya Peninsula, both with large extents of deciduous forest, are shown as almost completely non-forest. In a more recent classification of Guanacaste and the Nicoya Peninsula, using Landsat 7 ETM+ imagery, Arroyo-Mora (2002) shows that the forest cover is actually 45%. At the national scale, in the most recent remotely sensed forest cover inventory to date of the entire country of Costa Rica, Sanchez-Azofeifa et al. (2002) report a total forest extent 58% greater than the other previous studies (Castro-Salazar and Arias-Murillo 1998). Seasonal changes in leaf phenology in the deciduous forest are part of the reason for those differences. Even at the spatial resolution of most global monitoring systems (1km) significant areas of forest can be missed if only dry season images are used or if the phenological changes in leaf cover are

Chapter 3.1: Tropical Dry Climates

131

not taken into consideration. This forest, which has been ignored by previous remote sensing analysis, is not uniform and includes different stages of succession with different levels of deciduousness (Arroyo-Mora 2002). For example, in the recent global land cover classification from the MODIS Land Cover Classification Program, neither the area encompassing the Chamela Biological Station, México nor the Santa Rosa National Park, Costa Rica is classified as forest. These complications are important not only for classification purposes, but also in many cases outputs from such data sets are used in global models like CENTURY. The calculations from such models are then further used to calculate baselines and benefits of a given policy for carbon sequestration, for example.

5.

FINAL REMARKS

Since so many organisms depend upon phenological patterns in tropical forests, it is crucial to document how these phenological patterns may be changed by deforestation and the resulting habitat fragmentation. Future studies on phenological patterns of tropical plants should attempt to document intra-specific variation within distinct habitat types and under different levels of disturbance, in order to provide a clear understanding of ecosystem phenological response to different levels and types of disturbance. This information will be important in quantifying the effects of forest fragmentation on phenological patterns and ultimately on tropical ecosystems. A wealth of information is available on studies conducted with remotely sensed data in both the temperate and tropical regions. And while the image processing techniques may be similar, the ground validation techniques are very different in certain aspects. The complexity (structural and temporal) of the tropical deciduous forests also requires special consideration when field data are being collected. In certain cases, for example when collecting Leaf Area Index (LAI), new sampling techniques need to be developed to account for the spatial and temporal heterogeneity of the forest. This is also the case if there are certain specific phenological patterns of interest. Both the scale of the sampling, as well as the technique should be determined by the required data. For example, biophysical parameters of the canopy such as LAI, vegetation fraction (VF) and the fraction of photosynthetically active radiation (f (fPAR) have been successfully linked to remotely sensed data in many studies in conifer stands, temperate broad leaf forests and agricultural fields (Chen and Black 1991; Price and Bausch 1995; Chen and Cihlar 1996; Chen et al. 1997). However, similar techniques have not been as thoroughly explored in tropical dry forest environments, nor is there a clear

132

Phenology: An Integrative Environmental Science

understanding of the impact of phenology in these important biophysical variables. In addition, with the exception of a few studies such as ArroyoMora (2002) or Clark (2002), optical remote sensing studies in tropical environments have been predominantly conducted with either the Landsat (TM and ETM+) or AVHRR sensors. However, high spatial resolution multispectral sensors such as IKONOS (4 and 1 m spatial resolution and four spectral bands) and Quickbird (2 m and 60 cm spatial resolution, four spectral bands) have begun acquiring substantial worldwide archives and can play a key role in monitoring phenological processes in tropical dry forest environments. Also, with the introduction of ASTER (15 m spatial resolution, 14 spectral bands) data can be obtained quite economically. All three of these sensors may be used to capture detailed temporal changes in the dry deciduous forest. In addition, ALI (Advanced Land Imager), a new sensor from the EO-1 platform provides a more cost effective alternative for acquiring Landsat-type data. Increased spectral resolution may also be an option to characterize deciduous forests from a remote sensing point of view. Hyperspectral sensors such as Hyperion (30 m spatial resolution and 220 spectral bands) or the airborne sensor HYDICE (1 m spatial resolution and 220 bands) offer new possibilities for describing the phenological changes in the deciduous forest, but their application will be limited to the short life span of this sensor type. More small changes at the canopy level can be observed with these sensors than can be captured by multispectral sensors. These changes can be correlated to ground measurements such as chlorophyll concentrations as a function of age and complexity in order to begin modelling the seasonal changes in the ecosystem in greater detail. Hyperspectral data sets will provide a greater range of possibilities for deriving indices that may be more sensitive to the vegetation characteristics, as well as to phenological changes in dynamic environments.

REFERENCES CITED Aide, T. M., Patterns of leaf development and herbivory in a tropical understory community, Ecology, 74, 455-466, 1993. Andren, H., and P. Angelstarn, Elevated predation rates as an edge effect in habitat islands, Ecology, 69, 544-547, 1988. Appanah, S., General flowering in the climax rain forest of Southeast Asia, Journal of Tropical Ecology,1, 225–240, 1985. Arroyo-Mora, P., Forest cover assessment, Chorotega Region, Costa Rica, Unpublished M.Sc. Thesis, Department of Earth and Atmospheric Sciences, University of Alberta, Canada, 2002.

Chapter 3.1: Tropical Dry Climates

133

Ashton, P. S., T. J. Givnish, and S. Appanah, Staggered flowering in the Dipterocarpaceae: new insights into floral induction and the evolution of mast fruiting in the aseasonal tropics, The American Naturalist, 231, 44–66, 1988. Bierregaard, R. O., and T. E. Lovejoy, Effects of forest fragmentation on Amazonian understory bird communities, Acta Amazonica 19, 215-241, 1989. Borchert, R., Soil and stem water storage determine phenology and distribution of tropical dry forest trees, Ecology, 75, 1437–1449, 1994. Bullock, S. H., Plant reproduction in Neotropical dry forests, in Seasonally Tropical Forests, edited by H. A. Mooney, S. H. Bullock, and E. Medina, pp. 277-303, Cambridge University Press, Cambridge, UK, 1995. Cascante, A., M. Quesada, J. A. Lobo, and E. J. Fuchs, Effects of dry tropical forest fragmentation on the reproductive success and genetic structure of the tree, Samanea saman, Conserv. Biol., 16, 137-147, 2002. Castro-Salazar, R., and G. Arias-Murillo, Costa Rica: Toward the sustainability of its forest resources, Technical Report Fonda Nacional de Financiamiento Foresta (FONAFIFO), San Jose, Costa Rica, 1998. Chen, J. M., and T. A. Black, Measuring leaf area index of plant canopies with branch architecture, Agricult. Forest Met., 57, 1-12, 1991. Chen, J. M., and J. Chilar, Retrieving leaf area index of boreal forests using Landsat TM images, Remote Sens. Environ., 55, 153-162, 1996. Chen, J. M., P. M. Rich, S. T. Gower, J. M. Norman, and S. Plummer, Leaf area index of boreal forests: Theory, techniques and measurements, J. Geophys. Res. 102, 429-443, 1997. Clark, D., Applications of 1m and 4m resolution satellite imagery to studies of tropical forest ecology, management and secondary forest detection, in Tropical Forests: Past, Present, Future, p. 20, The Association of Tropical Biology Annual Meeting, Smithsonian Tropical Research Institute, 2002. Dirzo, R., and C. A. Dominguez, Plant-herbivore interactions in Mesoamerican tropical dry forests, in Seasonally dry tropical forests, edited by S. H. Bullock, H. A. Mooney, and E. Medina, pp. 277-303, Cambridge University Press, Cambridge, 1995. Doligez A., and H. Joly, Genetic diversity and spatial structure within a natural stand of a tropical forest tree species, Carapa procera (Meliaceae), in French Guiana, Heredity, 79,72-82, 1997. Eby, P., Seasonal movements of grey-headed flying-foxes, Pteropus poliocephalus (Chirpotera: Pteropodidae), from two maternity camps in northern New South Wales, Wildlife Research, 18, 547-559, 1991. Ewel, J., Natural systems as models for the design of sustainable systems of land use, Agroforestry Systems, 45, 1-21, 1999. Fanjul, L. and V.L. Barradas, Diurnal and seasonal variation in the water relations of some deciduous and evergreen trees of a deciduous dry forest of the western coast of Mexico, J. Appl. Ecol., 24, 289-303, 1987. Feinsinger, P., J. A. Wolfe, and L. A. Swarm, Island ecology: reduced hummingbird diversity and the pollination biology of plants on Trinidad and Tobago, West Indies, Ecology, 63, 494-506, 1982. Fleming, T. H., The short-tailed fruit bat: A study in plant-animal interactions, University of Chicago Press, Chicago, 365 pp., 1988. Fleming, T. H., R. A. Nuñez, and L. S. L. Sternberg, Seasonal changes in the diets of migrant and non-migrant nectarivorous bats as revealed by carbon stable isotope analysis, Oecologia, 94, 72-75, 1993.

134

Phenology: An Integrative Environmental Science

Flemming, T. H. and V. J. Sosa, Effects of nectarivorous and frugiviorous mammals on reproductive success of plants, J. Mammalogy, 75, 845-851, 1994. Frankie, G. W., H. G. Baker, and P. A. Opler, Comparative phenological studies of tress in a tropical lowland wet and dry forest sites of Costa Rica, J. Ecology, 62, 881-913, 1974. Frankie, G. W., S. B. Vinson, M. A. Rizzardi, T. L. Griswold, Diversity and abundance of bees visiting a mass flowering tree species in disturbed seasonal dry forest, Costa Rica, a Journal of the Kansas Entomological Society, 70, 281-296, 1998. Fuchs, E. J, J. A. Lobo, and M. Quesada, Effects of forest fragmentation and flowering phenology on the reproductive success and mating patterns on the tropical dry forest tree, Pachira quinata (Bombacaceae), Conserv. Biol., 17, 149-157, 2003. Gentry, A. H., Flowering phenology and diversity in tropical Bignoniaceae, Biotropica, 6, 6468, 1974. Gentry, A. H., Diversity of floristic composition of Neotropical dry forests, in Seasonally Tropical Forests, edited by H. A. Mooney, S. H. Bullock, and E. Medina, pp. 146-194, Cambridge University Press, Cambridge, UK, 1995. Haber, W. A. and G. W. Frankie, A tropical hawkmoth community: Costa Rican dry forest Sphingidae, Biotropica, 21, 155-172, 1989. Haber, W. A., and R. Stevenson, Diversity, migration, and conservation of butterflies in northern Costa Rica, in Biodiversity Conservation in Costa Rica: Learning the Lessons in a Seasonal Dry Forest, edited by G.W. Frankie, A. Mata, and S.B. Vinson, University of California Press, Berkeley, California, 2003. Hartshorn, G., and B. E. Hammel, Vegetation types and floristic patterns, in La Selva, Ecology and Natural History of a Neotropical Rain Forest, edited by L. McDade, K. S. Bawa, H. A. Hespenheide, and G. Hartshorn, pp. 73-89, University of Chicago Press, Chicago, 1994. Helversen, O. V., Adaptations of flowers to the pollination by Glossophaginae bats, Animalplant interactions in tropical environments, edited by W. Barthlott, pp. 41-49, Alexander Koenig Museum, Bonn, Germany, 1993. Heywood, V. H., G. M. Mace, R. M. May, and S. N. Stuart, Uncertainties in extinction rates, Nature, 368, 105, 1994. Holbrook, N. M., J. L. Witbeck, and H. A. Mooney, Drought responses of Neotropical dry forest trees, in Seasonally Tropical Forests, edited by H. A. Mooney, S. H. Bullock, and E. Medina, pp. 243-276, Cambridge University Press, Cambridge, UK, 1995. Holdridge L., Life Zone Ecology, Tropical Science Centre, San Jose, Costa Rica, 206 pp., 1967. Howe, H. F., Implications of seed dispersal by animals for tropical reserve management, Biological Conservation, 30, 261-281, 1984. Janzen, D. H., Synchronization of sexual reproduction of trees within dry season in Central America, Ecology, 21, 620–637, 1967. Janzen, D. H., Herbivores and the number of tree species in tropical forests, American Naturalist, 104, 501-528, 1970. Janzen, D. H., Reduction of seed predation on Bauhinia pauletia (Leguminosae) through habitat destruction in a Costa Rican deciduous forest, Brenesia 14/15, 325-336, 1978. Janzen, D. H., Costa Rica Natural History, University of Chicago Press, Chicago, 816 pp., 1983. Kalacska, M., G. A. Sánchez-Azofeifa, B. Rivard, P. Arroyo-Mora, R. Hall, J. Zhang, and K. Dutchak, Implications of phenological changes in the extraction of Mesoamerican Tropical Dry Forests through remote sensing: A case study from Costa Rica and Mexico, Proceedings of the International Conference, pp. 84-85, Wageningen University, The Netherlands, 2001.

Chapter 3.1: Tropical Dry Climates

135

Leemans, R., Global Holdridge Life Zone Classifications, Digital Raster Data, in Global Ecosystems Database Version 2.0, NOAA National Geophysical Data Center, Boulder, Colorado, 1992. Lemke, T. O., Foraging ecology of the long-nosed bat Glossophaga soricina with respect to food availability, Ecology, 65, 538-548, 1984. Lemke, T. O., Pollen carrying by the nectar-feeding bat Glossophaga soricina in a suburban environment, Biotropica, 17, 107-111, 1985. Levey, D. J., T. C. Moermond, and J. S. Denslow, Frugivory: an overview, in La Selva, Ecology and Natural History of a Neotropical Rain Forest, edited by L. McDade, K. S. Bawa, H. A. Hespenheide, and G. Hartshorn, pp. 282-294, University of Chicago Press, Chicago, 1994. Leopold, A. C., Photoperiodism in plants, Quarterly Review of Biology, 26, 247-263, 1951. Lobo, J. A., M. Quesada, K. E. Stoner, E. J. Fuchs, and Y. Herrerías-Diego, Latitudinal variation of phenological patterns of bombacaceous species in seasonal forests in Costa Rica and Mexico, Amer. J. Botany, (in press), 2003. Luttge, U., Physiological ecology of tropical plants, Springer-Verlag, Heidelberg, 384 pp., 1997. Marquis, R. J., Phenological variation in the Neotropical understory shrub Piper arieianum: causes and consequences, Ecology, 69, 1552-1565, 1988. Maass, J. M., Conversion of tropical dry forest to pasture and agriculture, in Seasonally Tropical Forests, edited by H. A. Mooney, S. H. Bullock, and E. Medina, pp. 399-422, Cambridge University Press, Cambridge, UK, 1995. Medina, E., Diversity of life forms of higher plants in Neotropical dry forests, in Seasonally Tropical Forests, edited by H. A. Mooney, S. H. Bullock, and E. Medina, pp. 221-242, Cambridge University Press, Cambridge, UK, 1995. Menaut, J. C., M. Lepage, and L. Abbadie L., Savannas, woodlands and dry forests in Africa, in Seasonally Tropical Forests, edited by H. A. Mooney, S. H. Bullock, and E. Medina, pp. 64-92, Cambridge University Press, Cambridge, UK, 1995. Mooney, H.A., S. H. Bullock, and E. Medina, Seasonally Tropical Forests, Cambridge University Press, Cambridge, UK, 450 pp., 1995. Murphy, P.G. and A. E. Lugo, Ecology of tropical dry forest, Annual Review of Ecology and Systematics, 17, 67-88, 1986. Murphy, P.G., and A. E. Lugo, Dry forests of Central America and the Caribbean, in Seasonally Tropical Forests, edited by H. A. Mooney, S. H. Bullock, and E. Medina, pp. 9-34, Cambridge University Press, Cambridge, UK, 1995. Murawski, D. A., J. L. Hamrick, S. P. Hubbell, and R. B. Foster, Mating systems of two bombacaceous trees of a Neotropical moist forest. Oecologia, 82, 501-506, 1990. Murawski, D. A., and J. L. Hamrick, Mating system and phenology of Ceiba pentandra (Bombacaceae) in Central Panama, J. Heredity, 83, 401-404, 1992. Murray, K. G., P. Feinsinger, W. H. Busby, Y. B. Linhart, J. H. Beach, P. Kinsman, Evaluation of character displacement among plants in two tropical pollination guilds, Ecology, 68, 1283–1293, 1987. Nason, J. D., and J. L. Hamrick, Reproductive and genetic consequences of forest fragmentation: two case studies of neotropical canopy trees, J. Heredity, 88, 264-276, 1997. Newstrom, L. E., G. W. Frankie, H. G. Baker, and R. K. Colwell, Diversity of long-term flowering patterns, in La Selva, Ecology and Natural History of a Neotropical Rain Forest, edited by L. McDade, K. S. Bawa, H. A. Hespenheide, and G. Hartshorn, pp. 142160, University of Chicago Press, Chicago, 1994.

136

Phenology: An Integrative Environmental Science

Opler, P. A., G. W. Frankie, and H. G. Baker, Rainfall as a factor in the release, timing, and sychronization of anthesis by tropical tress and shrubs, J. Biogeography, 3, 231-236, 1976. Piperno, D. R., and D. M. Pearsall, Origins of Agriculture in the Neotropics, Academic Press, New York, 400 pp., 2000. Price, J. C., and W. C. Bausch, Leaf area index estimation from visible and near-infrared reflectance data, Remote Sens. of Environ., 52, 55-65, 1995. Quesada, M., K. E. Stoner, V. Rosas-Guerrero, C. Palacios-Guevara, and J. Lobo, Effects of habitat disruption on the activity of nectarivorous bats (Chiroptera: Phyllostomidae) in a dry tropical forest: implications for the reproductive success of the neotropical tree Ceiba grandiflora, Oecología, 135, 400-406, 2003. Ratter, J.A., P. W. Richards, G. Argent, and D. R. Gifford, Observations of the vegetation of northeastern Mato Frosso. I. The woody vegetation types of the Xavantina-Cachimbo expedition area, Philosophical Transactions of the Royal Society of London, Series B, 266, 449-492, 1973. Reich, P. B. and R. Borchert, Water stress and tree phenology in a tropical dry forest in the tropical dry forest lowlands of Costa Rica, J. Ecology, 72, 61-74, 1984. Robertson, C., The philosophy of flower seasons, and the phenological relations of the entomophilous flora and the anthophilous insect fauna, American Naturalist, 29, 97-117, 1895. Rolstad, J., Consequences of forest fragmentation for the dynamics of bird populations: Concelptual issues and the evidence, Biological Journal of the Linnean Society, 42, 149163, 1991. Sader, S., and A. T. Joyce, Deforestation rates and trends in Costa Rica, 1940 to 1983, Biotropica, 20(1), 11-19, 1988. Sakai, S., K. Momose, T. Yumoto, T. Nagamitsu, H. Nagamasu, A. A. Hamid, and T. Nakasiiizuka, Plant reproductive phenology over four years including an episode of general flowering in a lowland Dipterocarp forest, Sarwak, Malaysia, Amer. J. Botany, 86, 1414–1436, 1999. Sanchez-Azofeifa, G.A. and Calvo J., Final report on the extent to Costa Rica forest cover: Year 2002, Costa Rican National Forest Financing Fund, 20 pp., 2002. Saunders, P. A., Jr., R. J. Hobbs and C. R. Margules, Biological consequences of ecosystem fragmentation: a review, Conservation Biology, 5, 18-32, 1991. Stephenson, A. G., The role of extrafloral nectaries of Catalpa speciosa in limiting herbivory and increasing fruit production, Ecology, 63, 663-669, 1982. Stiles, G. F., Ecology, flowering phenology, and hummingbird pollination of some Costa Rican Heliconia species, Ecology, 56, 285–301, 1975. Stiles, G. F., Coadapted competitors: the flowering seasons of hummingbird-pollinated plants in a tropical forest, Science, 198, 1177-1178, 1977. Stoner, K. E., Seasonal changes in abundance of tropical dry forest frugivorous and nectarivorous bats in northwestern Costa Rica, Canadian Journal of Zoology, 79, 16261633, 2001. Stoner, K. E., M. Quesada, V. Rosas-Guerrero, and J. A. Lobo, Effects of forest fragmentation on Colima long-nosed bat foraging in tropical dry forest in Jalisco, Mexico, Biotropica, 34, 462-467, 2002. Stoner, K. E., O. Salazar, K. A., R. C. Fernández,, and M. Quesada Population dynamics, reproduction, and diet of the lesser long-nosed bat ((Leptonycteris curasoae) in Jalisco, Mexico: implications for conservation, Biodiversity and Conservation, 12, 357-373, 2003.

Chapter 3.1: Tropical Dry Climates

137

Tallak Nilsen, E., and W. H. Muller, Phenology of the drought-deciduous shrub Lotus scoparius: climatic controls and adaptive significance, Ecological Monographs, 51, 323– 341, 1981. Trejo, I., and R. Dirzo, Deforestation of seasonally dry tropical forest: a national and local analysis in Mexico, Biological Conservation, 94, 133-142, 2000. van der Meer, F., Physical Principals of Optical Remote Sensing, in Spatial Statistics for Remote Sensing (Remote Sensing and Digital Image Processing Vol. 1), edited by A. Stein, F. van der Meer, and B. Gorte, pp. 27-40, Kluwer Academic Publishers, New York, 1999. van Schaik, C. P., Phenological changes in a Sumatran rain forest, Journal of Tropical Ecology, 2, 327–347, 1986. van Schaik, C. P., J. W. Terborgh, and S. J. Wright, The phenology of tropical forests: Adaptive significance and consequences for primary consumers, Annual Review of Ecology and Systematic, 24, 353–377, 1993. Walter, H., Ecology of tropical and subtropical vegetation, Oliver and Boyd, Edinburgh, 539 pp., 1971. Williams-Linera, G., Phenology of deciduous and broad leaf evergreen tree species in a Mexican tropical lower montane forest, Global Ecology and Biogeography Letters, 6, 115127, 1997. Whitmore, T. C., and J. A. Sayer, Deforestation and species extinction in tropical moist forests, in Tropical deforestation and species extinction, edited by T. C. Whitmore and J. A. Sayer, pp. 1-14, Chapman and Hall, London, 1992. Wright, S. J., Seasonal drought and the phenology of shrubs in a tropical moist forest, Ecology, 72, 1643-1657, 1991. Wright, S. J., and C. P. van Schaik, Light and the phenology of tropical trees, The American Naturalist, 143, 192–199, 1994. Wright, S. J., and F. H. Cornejo, Seasonal drought and leaf fall in a tropical forest, Ecology 71, 1165–1175, 1990. Zimmerman, J. K., D. W. Roubik, J. D. Ackerman, Asynchronous phenology of a neotropical orchid and its euglossine bee pollinators, Ecology, 70, 1192–1195, 1989.

Chapter 3.2 MEDITERRANEAN CLIMATES Donatella Spano1, Richard L. Snyder2, and Carla Cesaraccio3 1

Department of Economics and Woody Plant Ecosystems, University of Sassari, Sassari, Italy; 2Department of Land, Air, and Water Resources, University of California, Davis, CA, USA; 3Agroecosystem Monitoring Laboratory, Institute of Biometeorology, National Research Council, Sassari, Italy

Key words:

Mediterranean ecosystems, Drought, Temperature, Climate variability, Plant communities

1.

MEDITERRANEAN CHARACTERISTICS

1.1

Zones

Mediterranean-type ecosystems are found in the far west regions of continents between 30° and 40° north and south latitude (Figure 1). They cover about 2.73 million km2 (IUCN 1999), with the majority (i.e., 73%) of the ecosystem in the Mediterranean Basin including parts of Spain, Turkey, Morocco, and Italy (Rundel 1998). Areas are also found in California, Chile, Southwest and Southern Australia, and South Africa. In response to the climate, similar woody, shrubby plants, with evergreen sclerophyll leaves, have developed in communities of varying density. The names for the shrub vegetation vary by region because of language and plant structure. Common names for the vegetation include: maquis and garrigue in the Mediterranean Basin, chaparrall in California, matorrall in Chile, fynbos or renosterveldd in South Africa, and mallee (kwongan or heathlands) in Australia.

Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 139-156 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

140

1.2

Phenology: An Integrative Environmental Science

Climate

Mediterranean ecosystems formed as a result of the unique climate, which falls in a transition between dry, tropical and temperate zones. The

Figure 3.2-1. Geographical distribution of Mediterranean-type ecosystems.

main characteristics are (1) variable winter rainfall, (2) summer droughts of variable length, (3) intensive summer sunshine, (4) mild to hot summers, and (5) cool to cold winters. Commonly, there is a cold ocean current off the West coast of regions with a Mediterranean climate that strongly influences the weather. The range of summer and winter temperatures mainly depends on proximity to the ocean (or sea) with higher temperatures near the coast during cooler periods and higher temperatures inland during warmer periods. Temperatures also vary with elevation having consistently cooler temperature in the mountains. Excluding mountains, the annual precipitation range at lower elevations typically varies between 250-900 mm with most falling in the winter and spring (i.e., November – April in the Northern Hemisphere and May – October in the Southern Hemisphere). Westerly winds over cold ocean currents often lead to heavy marine fog that maintain low temperatures on the coast during summers. In the winter, the coastal areas tend to be fog free, whereas inland valleys that receive winter rainfall are prone to high-inversion, radiation fog. Differences in relative humidity are mainly related to temperature variations over the zone rather than absolute humidity. The Mediterranean climate is dominated by westerly winds over the ocean, so the water vapor pressure (or dew point temperature) tends to be similar over most of the zone.

Chapter 3.2: Mediterranean Climates

141

The five Mediterranean zones have similar characteristics, but there are important differences within each of the regions. Differences within a region are mainly related to the length of the summer drought period, which generally decreases as one moves poleward. For example, di Castri (1973, 1981) described a six-zone climate classification based on the length of drought period after Emberger (1962), as shown in Table 1. Table 3.2-1. Climate classification based on length of summer drought period. Classification Drought Period (months) Perarid Arid Semiarid Subhumid Humid Perhumid

1.3

11-12 9-10 7-8 5-6 3-4 1-2

Soil

Soil and climate both influence the development of natural vegetation, so a short discussion of soils is included here. More extensive discussions are presented by Thrower and Bradbury (1973), Zinke (1973), and di Castri et al. (1981). Most Mediterranean soils exhibit (1) considerable erosion, (2) alluvial deposition, (3) limited profile development, and (4) decreased soil development with increasing elevation. Because limestone is deficient in some areas, most soils often tend to have water infiltration problems. Due to the lower precipitation, parent materials weather slower in Mediterranean zones than in more humid regions. Because of seasonal drying some soils are dominated by invertization processes and produce Vertisols. The soils tend to vary from reddish to brownish with increasing elevation. Higher precipitation and cooler temperatures at higher elevations have led to the development of predominant brownish podzolic soils with higher organic matter and moderate lime accumulations at middle elevations (500-1000 m). At lower elevations (0-500 m) with less precipitation and higher temperature, the older terra rossa soils, having lower organic matter and a reddish color due to iron oxidation, developed from limestone. In the river valleys, alluvial soils are found as highly weathered soils in terraces, light and well-drained in alluvial fans, and heavy and poorly drained in the valley floors. In some valley basins, fine textured soils have greatly inhibited drainage. In many areas within Mediterranean zones, older paleosoils, which were formed under different climate conditions, are prevalent.

142

Phenology: An Integrative Environmental Science

2.

VEGETATION TYPES

2.1

Structure

Although the Mediterranean climate developed relatively recently over a small part of the Earth’s land surface, the distinctive flora have evolved with similar characteristics in widely different parts of the world. The climate is similar in each of the five zones, but, within a small area of a Mediterranean ecosystem, high heterogeneity in plant communities is common. This heterogeneity developed because of large variations in landforms, microclimate, soils, phyllogenetic origin, evolutionary strategy, ecological tolerance, and land use within the ecosystem. The appearance of natural vegetation and landscape forms is strikingly similar between the five Mediterranean zones. The shrubland plants are woody, shrubby, and evergreen. The plant leaves tend to be small, broad, stiff, thick, and waxy or oily. In some locations, there are small trees with or without an understory of annual and herbaceous perennials. The vegetation represents different successional stages in relation to climate, topographical features, and human impact (di Castri 1981), and it is prone to wildfires. di Castri (1981) presented a classification of the six Mediterranean (summer drought based) climate types (Table 1) and provided information on the structure of vegetation in each of the climate types. He noted that there were several overlapping clusters of characteristics between all five regions. However, the similarities between vegetation structures were most apparent between California and Chile, and between Australia and South Africa.

2.2

Floristic Composition

Mediterranean ecosystems have large species diversity including about 48,250 plant species, which is approximately 20% of the world total (Cowling et al. 1996). The Mediterranean Basin, South Africa, Southwestern Australia, and California have about 25,000; 8550; 8000; and 900 species, respectively (Archibold 1995; Rundel 1998). The Mediterranean Basin is mainly covered by scrub, sparse grass, or bare rock. However, there are scattered evergreen trees that suggest earlier presence of mixed forests. Several species of Quercus including the holm oak (Quercus ilex) prevail in the west with cork oak (Q. suber) dominant on non-calcareous soils. Arbutus unedo and other shrubs are found in the same plant communities. As the climate becomes more arid to the east, Kermes oak (Q. coccifera) becomes more prevalent than holm oak. Stone pine ((Pinus pinea), cluster pine (P. ( pinaster), and Aleppo pine (P. ( halepensis) are

Chapter 3.2: Mediterranean Climates

143

common at higher elevations in the west. In the drier eastern region (e.g., Syria, Lebanon, and Israel), Q. calliprinos, an evergreen oak, and deciduous oaks are common. Corsican pine (P. nigra) and P. brutia often dominate in locations where wildfires occurred. Q. ilex is also found on the Atlas Mountains of North Africa at the elevation of 2000 m. Shrublands are divided into maquis, which comprises evergreen shrubs and small trees about 2.0 m tall, garrigue on calcareous soils, and jarall on siliceous soils. All communities have representative species and the size depends on local conditions. South African sclerophyll plant communities include mountain and coastal types (Moll et al. 1984). The mountain fynbos mainly consists of broad-leaved proteoid shrubs, which are found at elevations up to about 1000 m and grow to heights between 1.5-2.5 m. At higher elevations, 0.21.5 m tall ericoid shrubs are dominant. In addition, 0.2-0.4 m tall shrubs and tussocky hemicryptophytes are present in the high elevation communities. Tussocky restioid shrubs, which reach 0.3 m, dominate communities at higher elevations. In arid, high-elevation regions, the vegetation is mainly karoo with abundant succulent forms. An open ericoid cover with shrubs growing to 1.0 m tall, dominates the west coast. Small shrubs, grasses, and annuals form an open heath with 1-2 m tall proteoids along the south coast. Western Australia is dominated by forests of karri (Eucalyptus ( diversicolor) and jarrah (E. marginata). Karri is restricted to regions with acidic soils (Rossiter and Ozanne 1970) and it grows in association with other tall eucalyptus. Casuarina decussata and species of Banksia are common in the understory of these forests. Jarrah forests occur on lateritic soils in areas with lower precipitation. These forests change to wandoo (E. rudunca) woodland as the annual precipitation decreases. The western region is separated from South Australia by the acacia shrubland. Mallee is the dominant cover in the southeastern Mediterranean zone. The prevalent species are E. diversifolia and E. incrassata. In more favorable sites, species such E. behriana grow with ground cover of herbs and grasses with few sclerophyllous shrubs (Specht 1981). These communities integrate with sclerophyll forests of stryngbark (E. baxteri) and messmate (E. ( obliqua). The Chilean matorrall communities occur in the coastal lowlands and on the west facing slopes of the Andes. Most matorrall species are 1-3 m tall, evergreen shrubs with small sclerophyllous leaves. Many spinescent species and drought-deciduous shrubs are also important in these regions (Rundel 1981). Salix chilensis, Cryptocarya alba, and other trees are found in wetter regions with shrubs forming a cover. Matorrall evergreen shrubs (e.g., Lithaea caustica and Quillaja saponaria) dominate coastal regions. In more arid locations, succulent species and Fluorensia thurifera are common. The

144

Phenology: An Integrative Environmental Science

central valley of Chile is dominated by Acacia caven (Ovalle et al. 1990, 1996). California chaparrall typically consists of a dense cover of 1-4 m tall, evergreen shrubs. In California, and particularly in the south, chamise ((Adenostoma fasciculatum) is common and California lilac (Ceanothus cuneatus) is sometimes associated. In the Sierra Nevada foothills, chaparral occurs above 500 m elevation. Pure stands of California lilac are considered a fire-successional form in Southern California, but it is a dominant species of chaparrall in Northern California (Hanes 1981). Manzanita ((Arctostaphylos spp.) occurs throughout California, especially where there is snow and temperatures drop below freezing in winter. Various Quercus species may be present on lower hillsides. Coastal sage scrub (e.g., Artemisia californica) is the main vegetation along the coast.

2.3

Environmental Effects

Common characteristics of Mediterranean zones are summer drought, fire, tectonic instability, and variable floods and erosion during winter. Perhaps the most important of these is summer drought; however, drought tends to be more severe in California, Chile, and the subarid region of the Mediterranean Basin (Rundel 1995, 1998). In fact, the Mediterranean climate exhibits extreme year-to-year variability. In the last century, the rainfall trends have been relatively consistent showing a general decrease and similar or more intense tendency is expected in the future (Cubasch 2001; IPCC 2001). Dense cover and high woody biomass of shrublands make them prone to wildfire, which is an important disturbance regime in Mediterranean climates. Frequency of natural wildfire differs greatly between and within Mediterranean zones (Mooney and Conrad 1977; Rundel 1981, 1983; Trabaud and Prodon 1993; Oechel and Moreno 1994) depending on many factors. Anthropogenic disturbance is one of the biggest factors affecting Deforestation, grazing, agricultural Mediterranean ecosystems. development, and fire starting and suppression have changed vegetation community structure, especially in recent decades. One of the main factors is deforestation, in order to permit more intensive agriculture. Increased urbanization and land abandonment has led to uneven management and greater frequency and extent of wildfire as a disturbance (Rundel 1998). Fire is a natural disturbance in Mediterranean ecosystems, and the vegetation has adapted. However, the frequency and intensity of wildfires increased dramatically in the last few decades (Rundel R 1998). This has led to reduced forest vigor and degradation of forest structure and soil stability (Kuzucuoglu 1989; Naveh 1990).

Chapter 3.2: Mediterranean Climates

145

Grazing of livestock has greatly influenced Mediterranean ecosystems. A good example is in California, where livestock grazing converted much of the grassland from native perennials to exotic annuals from the Mediterranean Basin even prior to immigration by large numbers of people of European ancestry (Rundel 1998). In the late 1800s, agricultural expansion into the central valley and southern California caused extensive changes in natural communities. Later, agricultural and urban expansion led to large changes in vegetation along the coast. Human activities influenced grassland and oak woodlands of the State mainly by replacing native perennial grasses with introduced annual grasses from Europe. Native Americans purposely set fires to control vegetation, but European immigrants introduced fire suppression as a management strategy in the late 1800s. This change in management has led to fewer but more intense wildfires (Minnich 1983; Rundel and Vankat 1989). When Spanish settlers arrived in Chile in the mid-1500s, they introduced grazing and agriculture that greatly changed the natural ecosystems. The impact is most obvious in the semi-arid transition region where over-grazing has caused devegetation and desertification (Ovalle et al. 1990, 1996). Also, much of the central valley now is covered with exotic annual grasses rather than the native grasses (Gulmon 1977). Recently, Chile has become more urban and there has been an abandonment of farms and ranches as the population leaves rural areas. This has led to a big increase in mainly anthropogenic wildfires that have grown in size and intensity. Even more recently, the planting of winegrape vineyards has expanded dramatically in Chile and in California at the expense of native woodlands (Rundel 1998). Agricultural development in Southwest Australia has resulted in widespread fragmentation of mallee ecosystems mixed in with agricultural lands (Rundel 1998). The fragmented habitats tend to be too small to maintain viable plant populations, which are also impacting on animal diversity. Deforestation is a big problem in native eucalypt forests, and the resulting rise in water tables has led to problems with saline paleosoil profiles (Rundel 1998), which threatens agriculture as well as the replanting of forests. The introduction of exotic species has resulted in problems with biological diversity in the Mediterranean climate zones. Anthropogenic impacts on the Mediterranean ecosystems in South Africa are less obvious than in the other regions to a large extent because the soils of the region are not conducive to support cereal and vegetable production (Rundel 1998). However, large animal hunting and deforestation have impacted on the vegetation. There has been a large introduction of nonnative trees, especially Australian acacias along rivers and streams.

146

Phenology: An Integrative Environmental Science

3.

PHENOLOGY IN THE MEDITERRANEAN CLIMATE

3.1

An Overview of Past Studies

Mediterranean regions show seasonal changes in resource availability, which affect growth and reproductive activities of vegetation. Resource fluctuations have a strong influence not only on the structure and composition of the vegetation but also on the seasonal behavior pattern of the species. For example, the sclerophyllous forest can remain active throughout the year, but there is a distinct annual growth rhythm because photosynthesis is limited by a variety of environmental and physiological constraints (i.e., drought and nutrients). However, several other species shed leaves during summer drought period. Over the last three decades, the economic, ecological, and cultural value of Mediterranean vegetation has been increasingly recognized (Quezel 1977) and many studies were devoted to improving management and protection of Mediterranean areas. In particular, there has been comparative research on the structure of Mediterranean region ecosystems, which included a detailed assessment of phenological species behavior in the different areas. The first systematic study on Mediterranean vegetation was presented by Mooney et al. (1977) within the International Biological Program (IBP), which started in 1970. The authors summarized the results of the comparison of the structural, functional, and evolutionary features of California and Chile ecosystems. At the plant community level, there is a longer protraction of each phenological event in Chile than in California due to both the greater diversity of growth form and more moderate climate in Chile (Mooney et al. 1977). In addition, di Castri (1981) pointed out that there were more species with non-overlapping phenological activities in Chile. As more information on the phenology of ecosystems in the Mediterranean Basin, South Africa and Australia became available, it was noted that there is a pronounced seasonal rhythm in the vegetative growth throughout the year in Mediterranean regions. However, less similarity in phenological pattern was found when comparing Chile, California, and Mediterranean Basin with South Africa and Australia. In South Africa and Australia, shrubs grow in the summer as well as in the spring (Cody and Mooney 1978) because of differences in origin of the biota (Specht 1973) and nutrient availability in the soils (Specht 1979, 1981). Comparative analysis of Mediterranean species development was intensified during the 1980s with more emphasis on the interactions between temperature and water as limiting factors. Tenhunen et al. (1987) summarized the results of years of cooperative work between several

Chapter 3.2: Mediterranean Climates

147

scientists on functional analysis in Mediterranean ecosystems. The work included recent studies on plant growth and development. Montenegro (1987) discussed the difficulty in comparing these ecosystems because of different methodologies used to quantifying phenology and growth. In Portugal, phenological observations conducted on different species (Quercus coccifera and Q. suber, Arbutus unedo, and Cistus salvifolius) showed that flowering stage occurred during all times of the year except the driest months in late summer and the coldest months in winter. Shoot growth was intense in the absence of water stress, and leaf drop was possibly more intense during drought (Pereira et al. 1987). Similar results were obtained on Q. coccifera and Arbutus unedo in Greece (Arianoutsou and Mardilis 1987), although the responses to the physical environment were not synchronous for the two species. Moll (1987) observed that the differences between vegetation in South Africa and in other Mediterranean regions reported by Mooney and Kummerow (1981) were mostly due to the fact that they compared non-heath shrubland in Chile, California, and the Mediterranean Basin with heath shrubland in South Africa. In the last decade, more attention was directed to the relationship between phenological events and seasonal fluctuations in nutrient and water uptake. A phenological survey conducted in central Italy (de Lillis and Fontanella 1992) showed the effect of increasing water stress and nutrient content on several species (Cistus monspeliensis, Pistacia lentiscus, Calicotoma villosa, Quercus ilex, Erica arborea, Arbutus unedo, Phillyrea media, Smilax aspera, and Ruscus aculeatus). Phenological rhythm of the community was closely correlated with changes in environmental conditions, and large variation occurred among species. In all species, peak growth was reached between March and early July, flowering occurred before July except A. unedo and S. aspera, which flowered in autumn and winter, and fructification was unrelated to summer aridity. An analysis of water availability and growth modulation allowed for division into droughttolerant species ((Pistacia lentiscus, Phillyrea media, Arbutus unedo, and Ruscus aculeatus), drought deciduous species (Calicotoma villosa), and semi-deciduous species (Cistus monspeliensis). Carbon leaf concentration peaked and nitrogen decreased when growth stopped. Correia et al. (1992) compared the phenological characteristics of four summer semi-deciduous (species of Cistus) and evergreen ((Pistacia lentiscus) shrubs in Portugal, corresponding to earlier and later successional stages of vegetation. The Cistus species were similar in growth, flowering, and fruiting phenology, showing a long period of leaf emergence relative to P. lentiscus, which had a flush-type leaf emergence and an almost simultaneous leaf fall. In general, Pistacia showed lower leaf nitrogen contents than the Cistus species, with minimum value in winter, when the Cistus species had the highest

148

Phenology: An Integrative Environmental Science

concentrations of nitrogen. However, increased drought frequency and intensity, is likely to greatly affect phenology of these species in the future.

3.2

Phenology and Drought

Many researchers have reported studies on the phenology of various Mediterranean species; however, little information was presented on the relationship between phenological stage occurrence and duration and intensity of drought period. Because the Mediterranean climate is highly variable and there are emerging problems of water scarcity, Spano et al. (1999) reported phenological observations of native and exotic species, on the island of Sardinia, with an emphasis on the impact of drought on phenology. They recorded weekly phenology observations for a period of 11 years on the common species Pistacia lentiscus, Olea europea, Myrtus communis, Quercus ilex, Spartium junceum, and Cercis siliquastrum, and on the exotic species Robinia pseudoacacia, Salix chrysocoma, and Tilia cordata. The climate during the study and the historical averages for 19511981 (Figure 2) show that the temperatures were similar, but there was markedly less winter rainfall during the observation period from 1986-1996. The canopy drought stress index CDSI (Baldocchi 1997) was calculated to look at differences in evaporative demand and precipitation between years. The CDSI is the ratio of cumulative precipitation and cumulative reference evapotranspiration. The range of phenological event dates for the nine species varied widely (Figure 3), especially for flowering of the exotic species. The authors showed that difference in accumulated degree-days could not explain the variations in observed phenological development. Regardless of the CDSI, during the winter and spring, there seemed to be little difference in the flowering dates of common species. However, the non-native species Salix chrysocoma and Tilia cordata showed more interannual variability and both exhibited later flowering when there was more rainfall during March (i.e., prior to flowering). There was no relationship with rainfall recorded two or more months prior to flowering. With the purpose to investigate the effects of temperature, rainfall, and evapotranspiration variability on phenology, Duce et al. (2000) conducted phenological observations on three maquis species and oak trees over the period 1997 to 1999 at Giara di Gesturi, a nature reserve located in Southern Sardinia, Italy. About 46% oak trees (Quercus suber) and about 32% successional Mediterranean maquis with four dominant species ((Arbutus unedo, Pistacia lentiscus, Phillyrea angustifolia, and Myrtus communis) cover the reserve. In Figure 4, daily rainfall amount and occurrence dates of flowering and full ripe fruit of Quercus suberr and Pistacia lentiscus are shown for 1997 and 1998. Flowering and full ripe fruit stages occurred

Chapter 3.2: Mediterranean Climates

149

about 1 month later in 1997 for both species. Sensitivity of the two species seemed to be related to rainfall distribution and the onset and duration of water deficit. In 1997, both species were affected by the lack of spring rainfall, which led to a longer and more intense drought period. Moreover, Duce et al. (2002) studied the effect of temperature and water availability on the flowering date of several natural species growing in Sardinia, Italy to explain year-to-year variation of the flowering data. The results showed a large variation by species in terms of observed flowering

Figure 3.2-2. Climate diagram (Walter and Lieth 1967) for Oristano, Italy for 1951-1981 (upper graph) and for 1986-1996 (lower graph).

150

Phenology: An Integrative Environmental Science

Figure 3.2-3. Range of dates of flowering ( ), full ripe fruit ( ), and leaf drop ( nine species growing in Oristano, Italy during the period 1986-1996.

) for

Figure 3.2-4. Daily precipitation and dates for flowering and full ripe fruit in 1997 (upper graph) and 1998 (lower graph) for Quercus suberr L. (Q) and Pistacia lentiscus L. (P) at Giara di Gesturi, Sardinia, Italy.

Chapter 3.2: Mediterranean Climates

151

dates and cumulative degree-day values, indicating that other factors in addition to heat units affected plant development. A significant factor in the prediction of flowering dates was obtained by adjusting the degree-day model for accumulated precipitation from the period when available soil water content was at maximum to the earliest flowering date typical of each species. In general the flowering date was postponed when the soil water was not limiting and flowering occurred earlier during drought years.

3.3

Phenology and Climate Change

During the 21st century, there will be more concern about climate change and especially drought in Mediterranean regions. In the past century, the overall global warming was about 0.5°C (IPCC 1996; Nicholls et al. 1996; IPCC 2001). Temperature records for Mediterranean areas show similar trends (Rambal and Hoff 1998). However, it is more difficult to see definite trends in rainfall patterns. Le Houèrou (1996) reported that there were no changes in rainfall patterns in Europe. However, Gregory and Mitchell (1995) showed that the regionally averaged total annual rainfall increased at latitudes greater than 45°N, and decreased at middle latitudes (35–40°N) with a decrease in the number of rain-days. Climate variation and change will likely occur at a number of scales in time and space, influencing plant physiology, competition between species, and global distribution of major ecosystems (Lindner et al. 1997). Several investigations were conducted on the impacts of climate change on different forest ecosystems, mainly focusing on the eco-physiological level (Mooney et al. 1991; Lindner et al. 1997). However, there is a lack of information on the potential effects of climate variability and change on Mediterranean forests. In Mediterranean climates, where the structural and functional f characteristics of ecosystems are determined by annual variability of temperature and precipitation, plant response to climate change is a crucial aspect of monitoring programs (Hope 1995). Several papers have presented the possible effects of changing climate factors (i.e., temperature and water availability) on the growth of forest ecosystems. Kramer et al. (2000) presented models simulating physiological features of the annual cycle for boreal coniferous, temperate-zone deciduous, and Mediterranean forest ecosystems. The phenology was mainly water driven and the ecosystem was a maritime pine forest (Pinus ( pinaster) located in southern France. The phenological models were coupled with the process-based forest growth model FORGRO (Mohren 1987, 1994; Kramer 1995) to evaluate the effect of different climate change scenarios. The study confirmed that phenology is mainly affected by seasonality in water availability. A dry year influences the growth of conifers for several years

152

Phenology: An Integrative Environmental Science

because both initiation and elongation of needles are affected by the availability of water, and the phenology of each of the forest types shows growth responses to a given climate change scenario. Recently, Peñuelas et al. (2002) compared phenological data from 1952 and 2000 providing a complete record of abundant plant and migratory birds species and a common butterfly. The data were collected in Catalonia, Spain. A conservative linear treatment of data showed that, in 2000, leaves unfolded on average 16 days earlier, leaf fall occurred about 13 days later, and plants flowered an average of six days earlier than in 1952. In addition, fruiting occurred about nine days earlier in 2002 than in 1974. Butterflies appeared 11 days earlier and spring migratory birds arrived 15 days later than 1952. The biggest change in both temperature and phenophase timing occurred in the last 25 years. The observed phenological changes, among the different species, may alter their competitive ability, ecology, and conservation, as well as the structure and functioning of the ecosystem.

4.

CONCLUSIONS

There are five Mediterranean zones around the world that are located near the west coasts of continents between 30o and 40o latitude. The climate represents a unique transition between arid zones towards the equator and temperate zones poleward. It is characterized by cold to cool, wet winters and warm to hot summers with varying periods of drought. The vegetation is similar in each region with woody, shrubby, and evergreen shrubland plants, sparse grass, scattered evergreen trees, and many species of oak trees. In all zones, anthropogenic disturbances including deforestation, grazing, agricultural development, and fire starting and suppression have changed the vegetation community structure. In general, phenology in the five Mediterranean zones presents a pronounced seasonal rhythm related to vegetation and environmental characteristics, with large variation among species. Whereas heat unit accumulation is the main factor affecting phenology of well-watered plants, phenology of natural Mediterranean vegetation is influenced by drought and plant nutrition in addition to heat units. Climatic fluctuations and drought in particular, directly influence resources availability and indirectly phenology. Like other climate regions, more research is needed to better understand the interaction between weather factors and phenology.

Chapter 3.2: Mediterranean Climates

153

REFERENCES CITED Archibold, O. W., Ecology of World Vegetation, Chapman and Hall, London, 510 pp., 1995. Arianoutsou, M., and Th. A. Mardilis, Observations on the phenology of two dominant plants of the Greek maquis, in Plant Response to Stress: Functional Analysis in Mediterranean Ecosystems, vol. 15, edited by J. D. Tenhunen, F. M. Catarino, O. L. Lange, and W. C. Oechel, NATO Adv. Sci. Inst. Ser. G Ecol. Sci., pp. 515-519, Springer-Verlag, Berlin, Heidelberg, 1987. Baldocchi, D., Measuring and modelling carbon dioxide and water vapour exchange over a temperate broad–leaved forest during the 1995 summer drought, Plant, Cell and Environ., 20, 1108–1122, 1997. Cody, M. L., and H. A. Mooney, Convergence versus non-convergence in Mediterraneanclimate ecosystems, Annu. Rev. Ecol. Syst., 9, 265-321, 1978. Correia, O. A., A. C. Martins., and F.M. Catarino, Comparative phenology and seasonal foliar nitrogen variation in mediterranean species of Portugal, Ecol. Mediterranea, 18, 7-18, 1992. Cowling, R. M., P. W. Rundel, B. B. Lamont, M. K. Arroyo, and M. Arianoutsou, Plant diversity in Mediterranean-climate regions, Trends Ecol. Evol., 11, 352-360, 1996. Cubasch U., G. A. Meehl, G. J. Boer, R. J. Stouffer, M. Dix, A. Noda, C. A. Senior, S. Raper, and K. S. Yap, Projections on future climate change, in Climate Change 2001: The Scientific Basis, Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, edited by J. T Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. van der Linden, X. Dai, K. Maskell, C. I. Johnson, pp. 526-582, Cambridge University Press, Cambridge, 2001. de Lillis, M., and A. Fontanella, Comparative phenology and growth in different species of the Mediterranean maquis of central Italy, Vegetatio, 99-100, 83-96, 1992. di Castri, F., Climatographical comparison between Chile and the Western Coast of North America, in Mediterranean Type Ecosystems, Origin and Structure, edited by F. di Castri and H.A. Mooney, pp. 21-36, Springer-Verlag, Berlin, Heidelberg, 1973. di Castri, F., Mediterranean-type shrublands of the world, in Ecosystems of the world. Mediterranean-type shrublands, vol. 11, edited by F. di Castri, D. W. Goodall, and R. L. Specht, pp 1-45, Elsevier Scientific Publishing Company, Amsterdam, 1981. di Castri, F., D. W. Goodall, and R. L. Specht, Ecosystems of the world: Mediterranean-type shrublands, vol. 11, Elsevier Scientific Publishing Company, Amsterdam, 643 pp., 1981. Duce, P., C. Cesaraccio, D. Spano, and R. L. Snyder, Weather variability effect on phenological events in a Mediterranean-type climate, paper presented at 15th Conference on Biometeorology / Aerobiology and 16th International Congress of Biometeorology, Kansas City, USA, 2002. Duce, P., D. Spano, C. Asunis, C. Cesaraccio, C. Sirca, and A. Motroni, Effect of climate variability on phenology and physiology of Mediterranean vegetation, paper presented at 3rd European Conference on Applied Climatology, Pisa, Italy, 2000. Emberger, L., Comment comprendre le territoire phytogéographique méditerranéen francaiú et la position “systématique” de celui-ci. Nationalia Monspeliensia. Série Botanica, 14, 47-54, 1962. Gregory, J. M., and J. F. B. Mitchell, Simulation of daily variability of surface temperature and precipitation over Europe in the current and 2xCO2 climates using the UKMO climate model, Q. J. R. Meteorol. Soc, 121, 1451-1476, 1995. Gulmon, S. L., A comparative study of grasslands of California and Chile, Flora, 166, 261278, 1977.

154

Phenology: An Integrative Environmental Science

Hanes, T. L., California Chaparral, in Ecosystems of the world: Mediterranean-type shrublands, vol. 11, edited by F. di Castri, D. W. Goodall, and R. L. Specht, pp. 139-174, Elsevier Scientific Publishing Company, Amsterdam, 1981. Hope, A. S., Monitoring ecosystem response to global change: high spatial resolution digital imagery, in Ecological Studies: Global change and Mediterranean-type ecosystems, vol. 117, edited by J. M. Moreno, and W. C. Oechel, pp. 43-85, Springer-Verlag, New York, 1995. Intergovernmental Panel on Climate Change, Climate change 1995: The Science of Climate Change, edited by J. E. Houghton, L. G. Meira Filho, B. A. Callander, N. Harris, A. Kattenberg, and K. Maskell, Cambridge University Press, Cambridge, 572 pp., 1996. Intergovernmental Panel on Climate Change, Climate Change 2001: Impacts, Adaptation and Vulnerability, edited by J. J. McCarthy, O. F. Canziani, N. A. Leary, D. J. Dokken, and K. S. White, Cambridge University Press, Cambridge, 1032 pp., 2001. International Union for the Conservation of Nature, Biological diversity of dry land, Mediterranean, arid, semiarid, savanna, and grassland ecosystems, World Conservation Union: Fourth meeting of subsidiary body on scientific, technical, and technological advice, Montreal, Canada, 1999. Kramer, K., Modelling comparison to evaluate the importance of phenology for the effects of climate change on growth of temperate-zone deciduous trees, Clim. Res., 5, 119-130, 1995. Kramer, K., I. Leinonen, and D. Loustau, The importance of phenology for the evaluation of impact of climate change on growth of boreal, temperate and Mediterranean forest ecosystems: an overview, Int. J. Biometeorol., 44, 67-75, 2000. Kuzucuoglu, C., Fires in Mediterranean region, Blue Planet Ecol., 72, 371-412, 1989. Le Houèrou, H. N., Climate change, drought and desertification, J. Arid Envirin., 34, 133185, 1996. Lindner, M., H. Bugmann, P. Lasch, M. Flechsig, and W. Cramer, Regional impacts of climatic change on forests in the state of Brandenburg, Germany, Agric. For. Meteorol., 84, 123-135, 1997. Minnich, R. A., Fire mosaics in southern California and northern Baja California, Science, 219, 1287-1294, 1983. Mohren, G. M. J., Simulation of forest growth, applied to Douglas fir stands in The Netherlands, Thesis, Agricultural University Wageningen, The Netherlands, 1987. Mohren, G. M. J. Modelling Norway spruce growth in relation to site conditions and atmospheric CO2, in Vegetation, modelling and climate change effects, edited by F. Veroustraete, and R. Ceulemans, pp. 7-22, SPB Academic Publishing, The Hague, 1994. Moll, E. J., Phenology of Mediterranean plants in relation to fire season: with special reference to the Cape Province South Africa, in Plant Response to Stress: Functional Analysis in Mediterranean Ecosystems, vol. 15, edited by J. D. Tenhunen, F. M. Catarino, O. L. Lange, and W. C. Oechel, NATO Adv. Sci. Inst. Ser. G Ecol. Sci., pp. 489-502, Springer-Verlag, Berlin, Heidelberg, 1987. Moll, E. J., B. M. Campbell, R.M. Cowling, L. Bossi, M. L. Karman, and C. Boucher, A description of major vegetation categories in and adjacent to the Fynbos biome, Report No. 83, S. Afr. Nat. Sci. Program, 210 pp., 1984. Montenegro, G., Quantification of Mediterranean plant phenology and growth, in Plant Response to Stress: Functional Analysis in Mediterranean Ecosystems, vol. 15, edited by J. D. Tenhunen, F. M. Catarino, O. L. Lange, and W. C. Oechel, NATO Adv. Sci. Inst. Ser. G Ecol. Sci., pp. 469-488, Springer-Verlag, Berlin, Heidelberg, 1987.

Chapter 3.2: Mediterranean Climates

155

Mooney, H. A., and C. E. Conrad, Symposium on the environmental consequences of fire and fuel management in Mediterranean ecosystem, USDA Forest Service General, Technical Report WO-3, U.S. Government Printing Office, 498 pp., 1977. Mooney, H. A., and J. Kummerow, Phenological development of plants in mediterranean climate regions, in Ecosystems of the world: Mediterranean-type shrublands, vol. 11, edited by F. di Castri, D. W. Goodall, and R. L. Specht, pp. 303-307, Elsevier Scientific Publishing Company, Amsterdam, 1981. Mooney, H. A., B. G. Drake, R. J. Luxmoore, W. C. Oechel, and L. F. Pitelka, Predicting ecosystem response to elevated CO2 concentrantions, BioScience, 41, 96-104, 1991. Mooney, H. A., A. Johnson, D. Parson, S. Keeley, A. Hoffman, R. Hays, J. Giliberto, and C. Chu, The producers-their resources and adaptive response, in Convergent evolution in Chile and California Mediterranean Climate Ecosystems, edited by H. A. Mooney, p. 224, Dowden, Hutchinson and Ross, Stroudsburg, Pennsylvania, 1977. Naveh, Z., Fire in the Mediterranean: a landscape perspective, in Fire in ecosystem dynamics, edited by J. G. Goldhammer, and M. J. Jenkins, pp. 401-434, SPB Academic Publ., The Hague, 1990. Nicholls, N., G. V. Gruza, J. Jouzel, T. R. Karl, L. A. Ogallo, and D. E. Parker, Observed climate variability and change, in Climate change 1995: The science of climate change, edited by J. E. Houghton, L. G. Meira Filho, B. A. Callander, N. Harris, A. Kattenberg, and K. Maskell, pp. 133-192, Cambridge University Press, Cambridge, 1996. Oechel, W. C., and M. J. Moreno, The role of fire in Mediterranean ecosystems, SpringerVerlag, Berlin, Heidelberg, 527 pp., 1994. Ovalle, C., J. Aronson, A. Del Pozo, and J. Avendano, The espinal: agroforestry system of the mediterranean-type climate region of Chile, Agrof. Syst., 10, 213-239, 1990. Ovalle, C., J. Avendano, A. Del Pozo, and J. Aronson, Land occupation patterns and vegetation structure of the anthropogenic savannas (espinales) of central Chile, For. Ecol. Manage., 86, 129-139, 1996. Penuelas, J., I. Filella, and P. Comas, Changed plant and animal life cycles from 1952 to 2000 in the mediterranean region, Global Change Biol., 8, 532-544, 2002. Pereira, J. S., G. Beyschlag, O. L. Lange, W. Beyschlag, and J. D. Tenhunen, Comparative phenology of four Mediterranean shrub species growing in Portugal, in Plant Response to Stress: Functional Analysis in Mediterranean Ecosystems, vol. 15, edited by J.D. Tenhunen, F. M. Catarino, O. L. Lange, and W. C. Oechel, NATO Adv. Sci. Inst. Ser. G Ecol. Sci., pp. 503-513, Springer-Verlag, Berlin, Heidelberg, 1987. Quezel, P., Forests of the Mediterranean basin, in Mediterranean forests and maquis: Ecology conservation and management, pp. 9-33, Unesco, Paris, 1977. Rambal, S., and C. Hoff, Mediterranean ecosystems and fire: the threats of global change, in Large forest fires, edited by J. M. Moreno, pp. 187-213, Backhuys Publishers, Leiden, The Netherlands, 1998. Rossiter, R. C., and P. G. Ozanne, South-western temperate forests, woodlands and heaths, in Australian Grassland, edited by R. M. Moore, pp. 199-218, Australian National University Press, Canberra, 1970. Rundel, P. W., The matorral zone of central Chile, in Ecosystems of the world: Mediterranean-type shrublands, vol. 11, edited by F. di Castri, D. W. Goodall, and R. L. Specht, pp. 175-201, Elsevier Scientific Publishing Company, Amsterdam, 1981. Rundel, P.W., Impact of fire on nutrient cycles in Mediterranean-type ecosystems, with reference to chaparral, in Mediterranean-type ecosystems: the role of nutrients, edited by F. J. Kruger, D. T. Mitchell, and J. U. M. Jarvis, pp.192-207, Springer-Verlag, Berlin, Heidelberg, 1983.

156

Phenology: An Integrative Environmental Science

Rundel, P. W., Adaptive significance of some morphological and physiological characteristics in mediterranean plants: facts and fallacies, in Time scales of biological responses to water constraints: the case of mediterranean biota, edited by J. Roy, J. Aronson, and F. di Castri, pp. 119-139, SPB Academic Publishers, Amsterdam, 1995. Rundel, P. W., Landscape disturbance in Mediterranean-type Ecosystems: an overview, in Ecological Studies: Landscape Degradation and Biodiversity in Mediterranean-Type Ecosystems, vol. 136, edited by P. W. Rundel, G. Montenegro, and F. M. Jaksic, pp. 3-22, Springer-Verlag, Berlin, Heidelberg, 1998. Rundel, P. W., and J. L. Vankat, Chaparral communities and ecosystems, in The California chaparral: paradigms reexamined, edited by S. Keeley, pp. 127-139, Los Angeles County Museum of Natural History, Los Angeles, California, 1989. Spano, D., C. Cesaraccio, P. Duce, and R. L. Snyder, Phenological Stages of Natural Species and their use as Climate Indicators, Int. J. Biometeorol., 42, 124-133, 1999. Specht, R. L., Structure and functional response of ecosystems in the Mediterranean climate of Australia, in Mediterranean-type ecosystems, Origin and Structure, edited by F. Castri, and H. A. Mooney, pp. 113-120, Springer-Verlag Berlin, Heidelberg, 1973. Specht, R. L., Ecosystems of the world: Heathlands and related shrublands, vol. 9A, Elsevier, Amsterdam, 497 pp., 1979. Specht, R. L., Mallee ecosystem in Southern Australia, in Mediterranean-type shrublands, edited by F. Castri, D. W. Goodall, and R. L. Specht, pp. 203-231. Elsevier, Amsterdam, 1981. Tenhunen, J. D., F. M. Catarino, O. L. Lange, and W. C. Oechel, Plant Response to Stress: Functional Analysis in Mediterranean Ecosystems, vol. 15, NATO Adv. Sci. Inst. Ser. G Ecol. Sci., Springer-Verlag, Berlin, Heidelberg, 1987. Thrower, N. J. W., and D. E. Bradbury, The physiography of the Mediterranean lands with special emphasis on California and Chile, in Mediterranean Type Ecosystems, Origin and Structure, edited by F. di Castri and H. A. Mooney, pp. 37-52, Springer-Verlag, Berlin, Heidelberg, 1973. Trabaud, L., and R. Prodon, Fire in Mediterranean ecosystems, Commission of European Communities, Brussels, 441 pp., 1993. Walter, H., and H. Lieth, Klimadiagramm Weltatlas, G Fischer Verlag, Jena, irreg. pp., 1967. Zinke, P. J., Analogies between the soil and vegetation types in Italy, Greece and California, in Mediterranean Type Ecosystem, Origin and Structure, edited by F. di Castri, and A. H. Mooney, pp. 61-80, Springer-Verlag, Berlin, Heidelberg, 1973.

Chapter 3.3 GRASSLANDS OF THE NORTH AMERICAN GREAT PLAINS Geoffrey M. Henebry Center for Advanced Land Management Information Technologies (CALMIT), School of Natural Resources, University of Nebraska, Lincoln, NE, USA

Key words:

1.

Tallgrass prairie, Anthesis, Transient Maxima Hypothesis, Great Plains

INTRODUCTION Question: What is Spring?— Growth in everything— Flesh and fleece, fur and feather, Grass and greenworld all together Gerard Manley Hopkins, The May Magnificat

The study of appearances of growth, development, and senescence in grassland communities is not an enterprise commonly pursued. Phenology of mid-latitude grasslands is, nevertheless, too large and diverse a collection of phenomena to cover within a single chapter, as it ought to embrace at once the vast Kazakh steppe, the chalk grasslands of southern England, and myriad other grassy landscapes. Thus, the view here shall be on the grasslands of the North American Great Plains, with particular reference to the tallgrass prairie as a type model. This chapter approaches phenological observations of grasslands from a perspective on ecological dynamics that is informed by hierarchy theory. A survey of the literature on what constitutes the expected phenological patterns of the grasses—rather than the forbs—within grasslands is provided. Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 157-174 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

158

Phenology: An Integrative Environmental Science

Various influences on grassland phenology are reviewed and the chapter concludes with an eye to future research directions.

2.

CHARACTERIZING GRASSLAND DYNAMICS

The grassland biome of North America emerges from interaction of vegetation with a particular climatic regime and a panoply of quasi-periodic influences, including drought, fire, and grazing by large ungulates (Bragg 1995). Four major types of grasslands can be distinguished within the Great Plains: (1) the tallgrass prairie that occurs principally east of 97°W; (2) the shortgrass prairie that occurs principally west of 101°W; (3) the mixed grass prairie that intergrades between these extremes; and (4) the sandhills prairie, which occurs on the inland sand dune system of the central Great Plains, principally in Nebraska, Colorado, and South Dakota (Joern and Keeler 1995). These prairies occur along two environmental gradients: the east-west gradient of diminishing total annual precipitation and the south-north gradient of diminishing average annual temperature. Stature of the community decreases as precipitation declines westward. Composition of the grass community shifts from dominance by C4 species in the southern and central Great Plains to increasing prevalence by C3 species, with a crossover near 44°N latitude (Sims 1988). The edaphic constraints that distinguish the sandhills prairie lead to a distinctive community composition and responsiveness to disturbances (Bragg 1995). From what perspective ought grassland phenology be approached? The Transient Maxima Hypothesis (TMH) posed by Seastedt and Knapp (1993) portrays subhumid grasslands in general and tallgrass prairies in particular as subject to the dynamic availability of multiple limiting resources. Thus, the interannual variation observed in primary production (Briggs and Knapp 1995; Knapp and Smith 2001) emerges from the interactivity of processes at different tempos: recent weather and atmospheric teleconnections; soil texture, nutrients, and moisture; topographic relief and fire frequency; and grazing by vertebrates and invertebrates above and below ground. When various windows of opportunity for resource capture develop from this interactivity, the vegetation exploits these transitory releases from constraints (Blair 1997). In effect, the “transient maxima” emerge from the constructive and destructive inference of constraints and forcings. However, which factors limit net primary production in a grassland at a given time and place depends strongly on recent and past events and the landscape context. Such a complex dynamic suggests that long-term study is required to articulate the boundaries of ecosystem behavior and develop an “ecological

Chapter 3.3: Grasslands of the N. American Great Plains

159

expectation” of a location, analogous to the operational definition of climate as the “meteorological expectation” based on thirty years of weather data. Indeed, Fuhlendorf et al. (2001) found that canopy structural attributes from a 44-year grazing experiment lent support to both equilibrium and nonequilibrium models of vegetation dynamics, depending on the temporal scale considered. A temporal corollary of the TMH predicts non-transitivity in constraint sequencing, e.g., drought followed by fire is not equivalent in ecological effect to fire followed by drought (Collins et al. 1998a). A spatial corollary of the TMH predicts that differential phasing of resource availability across the landscape can generate self-reinforcing patches of vegetation assemblages and maintain or increase heterogeneity in space and species (Fuhlendorf and Smeins 1997; Collins et al. 1998b). Coffin et al. (1996) reached similar conclusions examining community composition 53 years after disturbance in shortgrass prairie. Burke et al. (1998) further argued that since dry grasslands are water-limited, they are not characterized by the “indeterminate dominance” of the tallgrass prairie but by “below ground dominance” that leads to the development of resource islands and discontinuous ground cover across the landscape. However, resource islands are a particular dynamical basin of attraction within the predictive purview of the TMH. This understanding of grassland ecosystem dynamics has significant implications for the study of phenological patterns in grasslands and for linking these patterns to appropriate regulatory influences and constraints. Spatial observations of grassland dynamics cover the landscape setting, composition, and configuration, including edaphic and terrain factors. Temporal observations describe current and recent conditions in terms of the energy and moisture regimes of the climate—long-term averages, variances, and extremes—and the characteristics of quasi-periodic disturbances that structure the grassland, including disturbance intensity, extent, duration, time since event, and the mean and variance of the return interval. In the study of appearances that is phenology, hierarchy theory (Allen and Starr 1982) provides a conceptual tool to organize various factors that affect phenology in grasslands. By considering the characteristic frequencies of these factors in space and/or time, constraints that change slowly in space and/or time relative to growing season phenology can be distinguished from forcings that change rapidly in space and/or time (Allen and Hoekstra 1986). Meteorological forcings on grassland phenology include the onset, severity, and duration of drought (Weaver 1968), excess precipitation, hail, snow, and frost (Inouye 2000), in addition to recent and current weather, including the pace and tempo of insolation (Frank and Hoffmann 1989; Goodin and Henebry 1997). Climatic constraints include regional climatic complexes

160

Phenology: An Integrative Environmental Science

(Hayden 1998), atmospheric teleconnections (Goodin et al. 2003), and the seasonality of plant-atmosphere interactions (Schwartz and Marotz 1986). Perturbations affecting primary production—fire, grazing, disease or pest outbreak, and air pollution/atmospheric deposition—are forcings that can affect phenology. However, the characteristic return intervals of these forcings may be considered to constitute constraints on the system. Frequently encountered events may be called perturbations to the system with the implication that their familiarity enables some degree of system resistance or resilience. Infrequently encountered events in terms of extent, duration, or severity may be called disturbances with the implication that their unusual aspect leads to system stress. Both terrain features—elevation, slope, aspect, landscape position—and edaphic properties, such as soil texture and geochemical composition, are constraints. However, the dynamics of soil moisture, soil nutrients, and surface energy balance are all forcings relative to the phenology of a prairie canopy. One final example of this distinction is germane to the spatiotemporal duality of forcings and constraints: the seasonality of daylength is a temporal forcing but the maximum daylength at a location is a constraint that is a function of both latitude and time of year. For a given limiting resource, phenology will be strongly modulated by slowly changing constraints on resource availability. Forcings affecting resource availability that change rapidly in time and/or space may exert less influence on grassland phenology. Yet, in the presence of multiple limiting resources, phenology can be strongly affected within a given growing season by abrupt switching between what constitutes the current constraint on plant growth and development (Figure 1, Seastedt and Knapp 1993; Blair 1997). This switching between primary controls, as predicted by the TMH, leads to a diversity of potential spatio-temporal modulations on canopy development, which translates into significant breadth for the coexistence of different phenological patterns, even among the dominant species of the grassland matrix. Although the TMH was formulated to understand specifically the above ground biomass dynamics in the tallgrass prairie, mid-latitude grasslands across the planet can be understood, to lesser or greater degrees, within the dynamical complexity available through the TMH framework.

3.

OBSERVING PHENOLOGY IN GRASSLANDS

Phenology is a field of inquiry where the role of the observer is made quite explicit. While the primary challenge to phenological study in grasslands is the range of forcings and constraints that can influence plant growth and development, an additional related challenge is the relative lack

Chapter 3.3: Grasslands of the N. American Great Plains

161

of what Leopold and Jones (1947) identified as desirable qualities in items used for phenological survey: sharpness, visibility, and recurrence.

Figure 3.3-1. Grasslands phenology emerges from the interactivity of multiple influences as filtered through the specifics of spatial relationships and genetic heritage and the process of observation.

Sharpness is the relative distinctiveness in the item that reduces variation between observers. Leopold and Jones (1947) point to grasses that do not extrude their pollen as an example of lack of item sharpness for the detection of first bloom. Visibility may be best illustrated by its deficit: Rice (1950) resorted to dissection and microscopic examination to determine whether inflorescences had initiated in principal grasses of a mixed grass prairie. Recurrence relates to low interannual variation in the phenological item. Phenological studies of the grasses that compose the ecosystemic matrix of the prairies must face low recurrence, poor visibility, and blunted sharpness. In contrast, focus on the forbs and woody plants that dwell within the prairie yields phenological items that display sharpness, visibility, and recurrence. Not surprisingly, the scant literature on grasslands phenology tends to focus on showy forbs embedded within grass matrices, rather than the grasses per se. Tallgrass prairies have received more phenological study than other prairies, possibly due to their diversity and productivity. Leopold and Jones (1947) gathered cross-taxa phenological data over 11 years on the average blooming dates of 52 forbs and 7 grasses typical of the sandy soils of

162

Phenology: An Integrative Environmental Science

Wisconsin prairies. The 7 grasses, which are typical species of tallgrass prairie, and their flowering sequence were: (1) Scribner’s panicum, Dichanthelium oligosanthes var. scribnerianum (Nash) Gould—early June to mid-July; (2) junegrass, Koeleria macrantha (Ledeb.) J.A. Schultes— mid-June to early July; (3) switchgrass, Panicum virgatum L.—late July to mid-September; (4) big bluestem, Andropogon gerardii Vitman—late July to late August; (5) sideoats grama, Bouteloua curtipendula (Michx.) Torr.— early August to mid-September; (6) little bluestem, Schizachyrium scoparium (Michx.) Nash—mid-August to mid-September; and (7) indiangrass, Sorghastrum nutans (L.) Nash—mid-August to early September. Scribner’s panicum and junegrass are C3 species and the rest are C4 species. Noting that some prairie grasses and forbs commence growth relatively late in the season, they opined: “Could this be an evolutionary device for avoiding damage from spring fires?” Ahshapenek (1962) described grass phenology during 1957-1958 in an unburned, ungrazed tallgrass prairie in central Oklahoma dominated by C4 species. The sequence of anthesis of these grasses was switchgrass (late July), little bluestem (early August), big bluestem (late August), indiangrass (early September), and tall dropseed, Sporobolus compositus (Poir.) Merr. (mid-September). Proportion of flowering tillers ranged from 1.5% (tall dropseed) and 3.6% (big bluestem) to 21% (switchgrass) and 26% (little bluestem). Duration of phenological events in the prairie exhibits a range of variation. Leopold and Jones (1947) noted that the prairie has a “peculiar” interspersion of long and short blooming plants, primarily forbs. To examine the hypothesis that prairie plants “stagger” flowering times to reduce competition for pollinators, Anderson and Schelfhout (1980) analyzed flowering patterns of 77 tallgrass prairie plants (primarily forbs) from the Curtis Prairie at the University of Wisconsin Arboretum during 1950-1951. They found consistency in blooming sequences and duration over two years and speculated that many prairie plants use precise environmental cues such as photoperiod to initiate blooming. In a related study, Anderson and Adams (1981) compared flowering patterns of the Curtis Prairie with one year (1974) of observation at a tallgrass prairie in central Oklahoma dominated by little bluestem. The Oklahoma site had been formerly grazed and hayed, but not otherwise disturbed for five years prior to the study. From midMarch to mid-November, a clear bimodality was evident in the number of species flowering in Oklahoma. One peak occurred from May to early June and a second in early September. In contrast, the Wisconsin data suggested bimodality only weakly, with a broad peak occurring in mid-July. Extending this investigation, Kebart and Anderson (1987) examined flowering patterns in a tallgrass prairie community in Illinois in 1983 and

Chapter 3.3: Grasslands of the N. American Great Plains

163

compared those data with the Wisconsin and Oklahoma data. The Illinois pattern was distinct from the other two and suggested trimodality, with the principal peak in mid-July and minor peaks in mid-March and mid-August through mid-September. Parrish and Bazzaz (1979) also found trimodality in forb flowering sequences in a remnant tallgrass prairie in Illinois. Correlation analysis for the Wisconsin, Oklahoma, and Illinois sites revealed moderate to strong significant positive correlations between the number of species flowering per month with the mean monthly precipitation and temperature. However, there was not a consistent pattern in the correlation coefficients across sites and environmental variables. In addition, Rabinowitz et al. (1981) found no significant differences in phenological curves for 82 tallgrass species pollinated by either wind or insect, concluding that community flowering sequences were indistinguishable from random assemblages. In other words, distinct differences among species’ phenologies did not translate into ensemble flowering patterns that exhibit either temporal convergence or divergence that might support hypotheses regarding competition for pollinators. Speculating that combinations of soil moisture availability and daylength differentially trigger flowering, Anderson and Adams (1981) concluded that prairie species’ phenological patterns respond both to the environmental conditions of a particular growing season and across a latitudinal gradient of photoperiod and temperature. Phenological studies have also been conducted in shortgrass prairie with similarly inconclusive results. Dickinson and Dodd (1976) investigated phenological pattern in a shortgrass prairie in northeastern Colorado. The phenological timing of 34 species (12 grasses and sedges; 17 forbs; four half shrubs; and one succulent) was observed in the first study year (1972). These observations were extended to a second year for a subset of six species. One study objective was the identification of a single species that could indicate the phenological behavior of distinct groups of vegetation. Six phenological groups were identified from the 1972 survey. Their indicator species and time of flowering were (1) spike-rush sedge, Carex duriuscula C.A. Mey.—late April to mid-May; (2) prairie pepperweed, Lepidium densiflorum Schrad.—mid-May to early June; (3) needle and thread, Hesperostipa comata (Trin. & Rupr.) Barkworth—June; (4) red threeawn, Aristida purpurea var. longiseta (Steud.) Vasey—June to July; (5) blue grama, Bouteloua gracilis (Willd. ex Kunth) Lag. ex Griffith—July to August; and (6) fringed sagebrush, Artemisia frigida Willd.—late August to early September. Results showed the conventional division between warm-season and cool-season grasses was not a reliable predictor of flowering period. In particular, the nominally warm-season C4 species buffalograss, Buchloe dactyloides (Nutt.) Engelm., flowered at similar times

164

Phenology: An Integrative Environmental Science

to the cool-season C3 species western wheatgrass, Pascopyrum smithii (Rydb.) A. Löve. Another example of a C4 species that flowers early like a cool-season grass is eastern gamagrass, Tripsacum dactyloides (L.) L. (Dewald and Louthan 1979). From their survey, Dickinson and Dodd (1976) distinguished four groups of shortgrass prairie plants based on flowering pattern: very early single bloom; double bloom requiring a summer dormancy; drought-retarded midsummer single bloom; and late season single bloom. Given the importance of water in limiting productivity in the shortgrass prairie, particular attention has been given to its effect on phenology. To assess this response, Dickinson and Dodd (1976) applied water amendments to study plots by. Blue grama was found to be particularly responsive to a manipulated moisture regime: a delay of flowering followed by synchronization of flowering and acceleration of seed dispersal. For example, multiple blooms were observed on blue grama following downpours from convective thunderstorms that occurred approximately every 10 d. However, the secondary blooms observed following natural precipitation were not observed in the well-watered plots, suggesting that serial blooming requires wet-dry environmental cueing. A common thread running through many of these phenological studies in grasslands is their short duration relative to the high interannual variation in weather experienced across the Great Plains grasslands (Hayden 1998). While the broad patterns of phenology have been described, there are many variations on those themes that can serve to confound brief observational studies of communities of these long-lived perennials.

4.

FACTORS AFFECTING PHENOLOGY IN GRASSLANDS

4.1

Ecotypic Variation

During the middle of the last century, a significant body of research focused on the effect of photoperiod on phenology of rangeland grasses and the variation in that effect among geographic clones (Benedict 1941; Olmsted 1943, 1944, 1945; Rice 1950). Building on this work through a series of experiments on geographic clones of dominant and subordinate grasses transplanted to uniform gardens and in reciprocal transplants, McMillan found a strong gradient of more rapid phenological development from south to north and to a lesser extent from east to west (McMillan 1956a,b, 1957, 1959a,b). The longer growing seasons and more mesic

Chapter 3.3: Grasslands of the N. American Great Plains

165

environments toward the south and east of the Great Plains allowed grasses to initiate growth earlier and flower later than more northerly or westerly clones. McMillan (1960) argued that these patterns of ecotypic variation within species emerge from natural selection through the continuing interaction between the “habitat variable”, i.e., variation in local environmental forcings and constraints, and the “genetic variable”, i.e., the differential responses to the same habitat arising from the differential genetic potential among individual members of a population. Heide (1994) provides a contemporary review on the role of daylength and ecotypic variation on induction of flowering in temperate C3 temperate grasses.

4.2

Fire

Fire has substantial effects on grassland physiognomy, productivity, and phenology. In the tallgrass prairie, a disturbance-maintained ecosystem, periodic fire is required to prevent the encroachment by woody plants (Collins et al. 1998a). Moreover, dormant season fire following a few years of no burning can produce significant increases in the productivity of the dominant grasses, especially if the burn occurs in late spring. Frequent dormant season fire in the tallgrass prairie, however, can decrease species diversity (Knapp et al. 1998). Research indicates that the Nebraska sandhills prairie is also fire-adapted, with shifts in fire-positive and fire-negative species occurring subsequent to occasional wildfires (Bragg 1995). In the northern mixed-grass prairies dominated by C3 species, fire at any time of year generally decreases net primary production, even in years of normal precipitation, which is a response opposite to that observed in tallgrass prairie (Bragg 1995). Similarly, in the southern mixed-grass prairie dominated by C4 species, fire decreases production. In the water-limited shortgrass prairie of the western and southern Great Plains, fire during a dry year can have lasting impacts, although recovery from fire in normal-to-wet years can be relatively rapid (Bragg 1995; Burke et al. 1998). Fire frequency in these prairies is limited due to by accumulation and spatially patchy distribution of flammable litter (Bragg 1995). The timing and frequency of fire can affect various aspects of plant phenology but the effect of fire on subsequent flowering of dominant tallgrass species has been a focus of recurrent interest since the middle of the last century. Most of this work has centered on the responses of big bluestem under a regime of late spring burning, although some attention has been given to other prevalent species. Frequent burning can decrease the prevalence of flowering but long fire return intervals can also attenuate the flowering (Ehrenreich and Aikman 1963; Hulbert and Wilson 1983; Towne 1995). Significant interannual variations in flowering responses have been observed inn tallgrass prairies

166

Phenology: An Integrative Environmental Science

under annual burning. Knapp and Hulbert (1986) suggested that this variability points to shifts in controlling factors. Other investigators have contended that significant variation in intraspecific n responses to fire point to the importance of current site conditions (Zedler and Loucks 1969; Pemble et al. 1981). All these interpretations fall within the expected range of behaviors emerging from switching among primary controls as predicted by the TMH (Seastedt and Knapp 1993). The fire return interval also yields species-specific plant responses. Hulbert and Wilson (1983) found highly significant differences between annual and biennial burns for flowering density, flower stem biomass and height of big bluestem and indiangrass: the biennial burn yielded greater stem density and biomass. In contrast, there was no significant difference in inflorescence density or biomass in little bluestem. Furthermore, the significant differences between two-year and six-year spring burns were not uniform across species. The longer time between burns translated into taller, heavier, and more flowering stems for big bluestem. Fewer, smaller flowering stems in both indiangrass and little bluestem resulted from the longer fire interval. Towne (1995) observed that, during two uncharacteristically wet years, annual burning stimulated 15% of big bluestem and indiangrass tillers to flower. In contrast, only 3% of tillers developed inflorescences in unburned prairie. Flowering of big bluestem peaked (44%) in the year of burning following three years without burning (a quadrennial fire return interval), but indiangrass flowering was greatly diminished with this burning regime. Season of fire also affects flowering response. Henderson et al. (1983) examined the interaction of site condition and seasonality of burning for three C3 and four C4 grasses. They found that the C3 species flowered significantly less vigorously following late spring burns. A range of flowering responses in the C4 species were observed with a general—though not universal—trend toward greater flowering with late spring burns. Towne (1995) found the percentage of reproductive indiangrass tillers to be significantly less when annual burning occurred in November or March. Big bluestem, in contrast, showed no significant effect of the season of annual burning. Benning and Bragg (1993) found that the specific timing of burning within late spring could yield significantly different flowering responses, suggesting that the proximate environmental context (i.e., constraint sequencing) of the burning event is important in floral induction. Flowering stalks of dominant tallgrasses increase in number, density, and stature following spring fire in areas where prior years’ growth has formed a thick litter layer (Curtis and Partch 1950; Kucera and Ehrenreich 1962; Ehrenreich and Aikman 1963; Old 1969; Knapp 1984a; Knapp and Hulbert 1986). Petersen (1983) suggested that the presence of fire is a more

Chapter 3.3: Grasslands of the N. American Great Plains

167

important environmental cue for flowering in big bluestem than the removal of the litter layer or the addition of nutrient-rich surface-darkening ash as argued by others (Curtis and Partch 1950; Ehrenreich and Aikman 1963). Current thinking points to changes in microclimate, particularly due to the removal of the litter layer, as the primary stimuli to flowering following spring burns. The grassland detritus layer composed of previous years’ growth acts as a biotic buffer, which limits irradiance, increases the threshold for effective precipitation, and retards bare soil evaporation, thus reducing productivity (Knapp and Seastedt 1986; Knapp et al. 1998). Burning removes the accumulation of previous years’ growth, thereby increasing the quality of the light environment, the potential rate of evapotranspiration, and the variability of the surface energy balance. Furthermore, for all the biomass that has accumulated above ground, there is even more biomass below ground as fine root material. The pulse in primary production is fueled by the interaction of favorable light environment with a pool of organic nitrogen that can be mineralized and made available for canopy development, if the soil moisture is favorable (Blair 1997; Blair et al. 1998). The resulting pulse of available nitrogen can support relatively high flowering rates (Towne 1995). However, burning over the long term reduces nitrogen availability. Big bluestem has high nitrogen use efficiency and becomes a superior competitor under suboptimal conditions for its growth and development (Seastedt et al. 1991; Seastedt 1995).

4.3

Water Stress

Deficits of water leading to plant stress can affect phenological development. Investigating the effect of water stress in the tallgrass prairie on growth of big bluestem, little bluestem, and switchgrass in wet and droughty years, Knapp (1984b) found that reproductive effort (density, height, and biomass of flowering stems) could be significantly diminished during drought. Supplemental irrigation during a wetter year increased reproductive effort of switchgrass but not big bluestem or little bluestem. During a droughty year, however, irrigation increased primary production and reproduction in all three species. Big bluestem had a greater capacity for osmotic adjustment than little bluestem or switchgrass; however, big bluestem exhibited no reproductive effort in the droughty year of 1983. Late season water stress can reduce flowering in semi-arid C4 grasses (Alcocer-Ruthling et al. 1989). Dickinson and Dodd (1976) found that several shortgrass species could flower multiple times during a single season when short dry spells were followed by sufficient precipitation. Similarly, Beatley (1974) described how heavy rains trigger phenological events in vegetation of the Mojave Desert vegetation, including its perennial grasses.

168

4.4

Phenology: An Integrative Environmental Science

Spatial Influences

Topographic position can affect microclimate, including insolation, drainage, deposition, and exposure to wind. Although topoposition has been shown to have a significant effect on the phenological development of spring wildflowers (Jackson 1966), the influence for grassland species is variable. Rice (1950) saw no topographic effect on flowering in little bluestem in Oklahoma. McMillan (1956b) found that little bluestem clones from different topopositions in steep loess bluffs exhibited variations in flowering date after transplanting to a uniform garden. Topoposition has been shown to interact with fire to results in different plant responses. Zedler and Loucks (1969) found topoposition influences the susceptibility of grass production and flowering response to fire in a disturbed prairie. Pemble et al. (1981) attributed the significant intraspecific variations they found in the flowering responses for grasses at a Minnesota prairie to site-specific modulation of moisture conditions following a spring burn. Similarly, Knapp (1985) found that early season production in big bluestem in the more mesic lowlands of Konza Prairie were significantly greater than in the more exposed, warmer, and windier uplands.

4.5

Herbivory and Other Influences

Other influences on phenological timing and pace may include any disturbance that removes or kills above ground biomass, including herbivory, mowing, and hail. Hover and Bragg (1981), for example, found that summer mowing significantly decreased big bluestem flowering stem density compared to spring mowing or burning. Herbivory, which is a more selective process to remove plant biomass both above ground and below ground, affects plant growth, development, and reproduction in myriad ways, including through indirect influences and delayed effects. The literature on grazing in grasslands is too extensive to engage here. The primary effect of grazing is the reduction in flowering stem density as a result of the removal of apical meristems that may produce inflorescences. However, an indicative study (Vinton and Hartnett 1992) points to the complexities of interactions that may found. For example, bison grazing and simulated herbivory by clipping of big bluestem and switchgrass increased relative growth rates during the growing season that compensated for the lost tissue. In contrast, big bluestem tillers that had been repeatedly grazed in the previous growing season exhibited reductions in relative growth rates, survival and accumulated biomass in the subsequent growing season, although one was distinguished by a severe drought. In fire-related interactions, increases in relative growth rates of big bluestem following

Chapter 3.3: Grasslands of the N. American Great Plains

169

defoliation were greater in unburned than burned prairie; a topographic effect was also noted. Switchgrass, in contrast, was less responsive to environmental conditions, showing similar responses to defoliation in burned and unburned prairie (Vinton and Hartnett 1992).

5.

FUTURE DIRECTIONS

Phenological observations in mid-latitude grasslands have intrinsically low information density because of the high interannual variation in weather patterns. Only from several years of data can robust patterns emerge. Yet many studies of grassland phenology in the North American Great Plains have relied on only one or two years of observation. Dedicated, intensive phenological observations are expensive and thus rare—even the seminal cross-taxa phenology of Leopold and Jones (1947) was constructed from data collected largely in passing. Moreover, the sensitivity of some dominant species to topoposition and disturbance history complicates simple surveying. What then may be future directions for observing, monitoring, and modeling the phenologies found in grassland ecosystems? One bottom-up direction is in situ wireless sensor networks (Withey et al. 2001), which promise distributed monitoring capabilities of perisurficial environments. The use of spaceborne sensors to measure and monitor land surface phenology, including grasslands, is reviewed in Chapter 5.1. One further top-down direction stems from recent advances in the use of fine spectral resolution sensing to retrieve pigment concentrations, including anthocyanins (Gitelson et al. 2001), chlorophylls (Gitelson et al. 2002a), and carotenoids (Gitelson et al. 2002b). These techniques hold promise for the detection of anthesis and senescence as well as the onset of spring. However, it is important to note that, given the critical role of the scale of observation in phenological studies, the translations from remote observations to canopy dynamics as well as from leaf phenomena to vegetated landscape are intrinsically difficult and likely require site-specific rather than generic solutions. The question of whatt Spring is may be addressed more readily than the dual when questions about Springs past and Springs future. Indeed, it may be fair to characterize this simple question—When will Spring arrive?—as a canonical query for ecological forecasting (Clark et al. 2001; Henebry and Goodin 2002) and global change research (Myneni et al. 1997; Schwartz 1999). To integrate the various phenologies of “grass and greenworld” within the purview of ecosystem ecology and biogeosciences, the myriad ecophysiological responses of grassland canopies must be placed within the

170

Phenology: An Integrative Environmental Science

contexts of their landscapes, land uses, and disturbance regimes, by means of longer term empirical investigations—using field-based and space-based monitoring—modeling studies, and the development of ecological forecasting techniques.

ACKNOWLEDGEMENTS The accuracy and clarity of this chapter were significantly enhanced by comments and suggestions from an anonymous reviewer, Mark Schwartz, Andrés Viña, and especially from the careful readings by Tom Bragg and Gene Towne. Thanks to Ana Braga-Henebry for assistance with the figure. This survey was partially supported through the NSF Biodiversity and Ecosystem Informatics (BDEI) program (EIA #0131937).

REFERENCES CITED Ahshapanek, D., Phenology of a native tall-grass prairie in Central Oklahoma, Ecology, 43, 135-138, 1962. Alcocer-Ruthling, M., R. Robberecht, and D. C. Thill, The response of Bouteloua scorpioides to water stress at two phenological stages, Botanical Gazette, 150, 454-461, 1989. Allen, T. F. H., and T.W. Hoekstra, Description of complexity in prairies through hierarchy theory, in Proceedings of the Ninth North American Prairie Conference edited by G.K. Clambey and R.H. Pemble, pp. 71-73, Tri-college University Center for Environmental Studies, Fargo, ND, 1986. Allen, T. F. H., and T. B. Starr, Hierarchy: Perspectives for Ecological Complexity, University of Chicago Press, Chicago, 310 pp., 1982. Anderson, R. C., and D. E. Adams, Flowering patterns in a central Oklahoma grassland, Ohio Biological Survey Biological Notes, 15, 232-235, 1981. Anderson, R. C., and S. Schelfhout, Phenological patterns among tallgrass prairie plants and their implications for pollinator competition, Amer. Midland Naturalist, 104, 253-263, 1980. Beatley, J. C., Phenological events and their environmental triggers in Mojave Desert ecosystems, Ecology, 55, 856-863, 1974. Benedict, H. M., Growth of some range grasses in reduced light intensities at Cheyenne, Wyoming, Botanical Gazette, 102, 582-589, 1941. Benning, T. L., and T. B. Bragg, Response of big bluestem ((Andropogon gerardii Vitman) to timing of spring burning, Amer. Midland Naturalist, 130, 127-132, 1993. Blair, J. M., Fire, N availability, and plant response in grasslands: A test of the transient maxima hypothesis, Ecology, 78, 2359-2368, 1997. Blair, J. M., T. R. Seastedt, C.W. Rice and R.A. Ramundo, Terrestrial nutrient cycling in tallgrass prairie, in Grassland Dynamics: Long-Term Ecological Research in Tallgrass Prairie edited by A. K. Knapp, D. C. Hartnett, J. M. Briggs, and S. Collins, pp. 222-243, Oxford University Press, New York, 1998.

Chapter 3.3: Grasslands of the N. American Great Plains

171

Bragg, T. B., The physical environment of great plains grasslands, in The Changing Prairie, edited by A. Joern and K. H. Keeler, pp. 49-81, Oxford University Press, New York, 1995. Briggs, J. M., and A. K. Knapp, Interannual variability in primary production in tallgrass prairie: climate, soil moisture, topographic position and fire as determinants of aboveground biomass, Amer. J. Botany, 82, 1024-1030, 1995. Burke, I. C., W. K. Lauenroth, M. A. Vinton, P. B. Hook, R. H. Kelly, H. E. Epstein, M. R. Aguiar, M. D. Robles, M. O. Aguilera, K. L. Murphy, and R. A. Gill, Plant-soil interactions in temperate grasslands, Biogeochemistry, 42, 121-143, 1998 Clark J. S., R. Carpenter, M. Barber, S. Collins, A. Dobson, J. Foley, D. Lodge, M. Pascual, R. Pielke Jr., W. Pizer, C. Pringle, W. V. Reid, K. A. Rose, O. Sala, W. H. Schlesinger, D. Wall, and D. Wear, Ecological forecasts: an emerging imperative, Science, 293, 657–660, 2001. Coffin, D. C., W. K. Lauenroth, and I. C. Burke, Recovery of vegetation in a semiarid grassland 53 years after disturbance, Ecol. Applications, 6, 538-555, 1996. Collins, S., A. K. Knapp, D. C. Hartnett, and J. M. Briggs, The dynamic tallgrass prairie. Synthesis and research opportunities, in Grassland Dynamics: Long-Term Ecological Research in Tallgrass Prairie, edited by A. K. Knapp, D. C. Hartnett, J. M. Briggs, S. Collins, pp. 301-315, Oxford University Press, New York, 1998a. Collins, S. L., A. K. Knapp, J. M. Briggs, J. M. Blair, and E. M. Steinauer, Modulation of diversity by grazing and mowing in native tallgrass prairie, Science, 280, 745-747, 1998b. Curtis, J. T. and M. L. Partch, Some factors affecting flower production in Andropogon gerardi, Ecology, 31, 488-489, 1950. Dewald, C.L, and V. H. Louthan, Sequential development of shoot system components in eastern gamagrass, J. Range Management, 32, 147-151, 1979. Dickinson, C. E., and J. L. Dodd, Phenological pattern in the shortgrass prairie, Amer. Midland Naturalist, 96, 367-378, 1976. Ehrenreich, J. H., and J. M. Aikman, An ecological study of the effect of certain management practices on native prairie in Iowa, Ecol. Monographs, 33, 113-130, 1963. Frank, A. B., and L. Hofmann, Grazing management, growing degree-days, and morphological development for native grasses on the northern Great Plains, J. Range Management, 42, 199-202, 1989. Fuhlendorf, S. D., and F. E. Smeins, Long-term vegetation dynamics mediated by herbivores, weather and fire in a Juniperus-Quercus savanna, J. Veg. Science, 8, 819-828, 1997. Fuhlendorf, S. D., D. D. Briske, and F.E. Smeins, Herbaceous vegetation change in variable rangeland environments: The relative contribution of grazing and climatic variability, Appl. J. Veg. Science, 4, 177-188, 2001. Gitelson, A. A., M. N. Merzlyak, O. B. Chivkunova, Optical properties and nondestructive estimation of anthocyanin content in plant leaves, Photochemistry and Photobiology, 74, 38-45, 2001. Gitelson, A. A., Y. J. Kaufman, R. Stark, and D. Rundquist, Novel algorithms for remote estimation of vegetation fraction, Remote Sens. Environ., 80, 76-87, 2002a. Gitelson, A. A., Y. Zur, O. B. Chivkunova, and M.N. Merzlyak, Assessing carotenoid content in plant leaves with reflectance spectroscopy, Photochemistry and Photobiology, 75, 272281, 2002b. Goodin, D. G., and G. M. Henebry, Monitoring ecological disturbance in tallgrass prairie using seasonal NDVI trajectories and a discriminant function mixture model, Remote Sens. Environ., 61, 270-278, 1997.

172

Phenology: An Integrative Environmental Science

Goodin, D. G., P. A. Fay, and M. J. McHugh, Climate variability in tallgrass prairie at multiple time scales: Konza Prairie Biological Station, USA, in Climate Variability and Ecosystem Response, edited by D. E. Greenland, D. G. Goodin, and R. Smith, pp. 411423, Oxford University Press, New York, 2003. Hayden, B. P., Regional climate and the distribution of tallgrass prairie, in Grassland Dynamics: Long-Term Ecological Research in Tallgrass Prairie, edited by A. K. Knapp, D. C. Hartnett, J. M. Briggs, and S. Collins, pp. 19-34, Oxford University Press, New York, 1998. Heide, O. M., Control of flowering and reproduction in temperate grasses, New Phytologist, 128, 347-362, 1994. Henderson, R. A., D. L. Lovell, and E. A. Howell, The flowering responses of 7 grasses to seasonal timing of prescribed burns in remnant Wisconsin prairie, in Proceedings of the Eighth North American Prairie Conference, edited by R. Brewer, pp. 7-10, Western Michigan University, Kalamazoo, MI, 1983. Henebry, G. M., and D. G. Goodin, Landscape trajectory analysis: toward spatio-temporal models of biogeophysical fields for ecological forecasting, in Workshop on Spatiotemporal Data Models of Biogeophysical Fields for Ecological Forecasting: April 8-10, 2002, San Diego Supercomputer Center, La Jolla, California, edited by G. M. Henebry, pp. 9-13, CALMIT, University of Nebraska, Lincoln, NE, 2002. Hover, E. I. and T. B. Bragg, Effect of season of burning and mowing on an eastern Nebraska Stipa-Andropogon prairie, Amer. Midland Naturalist, 105, 13-18, 1981. Hulbert, L. C., and J. K. Wilson, Fire interval effects on flowering of grasses in Kansas Bluestem Prairie, in Proceedings of the Seventh North American Prairie Conference; 1980 August 4-6; Springfield, MO, edited by C. L. Kucera, pp. 255-257, University of Missouri, Columbia, MO, 1983. Inouye, D. W., The ecological and evolutionary significance of frost in the context of climate change, Ecol. Letters, 3, 457-463, 2000. Jackson, M., Effects of microclimate on spring flowering phenology, Ecology, 47, 407-415, 1966. Joern, A., and K. H. Keeler, Getting the lay of the land: Introducing North American native grasslands, in The Changing Prairie, edited by A. Joern and K. H. Keeler, pp. 11-24, Oxford University Press, New York, 1995. Kebart, K. K., and R. C. Anderson, Phenological and climatic patterns in three tallgrass prairies, Southwestern Naturalist, 32, 29-37, 1987. Knapp, A. K., Post-burn differences in solar radiation, leaf temperature and water stress influencing production in a lowland tallgrass prairie, American Journal of Botany, 71, 220-227, 1984a. Knapp, A. K., Water relations and growth of three grasses during wet and drought years in tallgrass prairie, Oecologia, 65, 35-43, 1984b. Knapp, A. K., Early season production and microclimate associated with topography in a C4 dominated grassland, Acta Oecologica/ Oecologica Plantarum, 6, 337-346, 1985. Knapp, A. K., J. M. Briggs, J. M. Blair, and C. L. Turner, Patterns and controls of aboveground net primary production in tallgrass prairie, in Grassland Dynamics: LongTerm Ecological Research in Tallgrass Prairie edited by A. K. Knapp, D. C. Hartnett, J. M. Briggs, and S. Collins, pp. 193-221, Oxford University Press, New York, 1998. Knapp, A. K., and L. C. Hulbert, Production, density and height of flower stalks of three grasses in annually burned and unburned eastern Kansas tallgrass prairie: a four year record, Southwestern Naturalist, 31, 235-241, 1986.

Chapter 3.3: Grasslands of the N. American Great Plains

173

Knapp, A. K., and T. R. Seastedt. Detritus accumulation limits productivity of tallgrass prairie, BioScience, 36, 662-668, 1986. Knapp, A. K., and M. D. Smith, Variation among biomes in temporal dynamics of aboveground primary production, Science, 291, 481-484, 2001. Kucera, C. L., and J. H. Ehrenreich, Some effects of annual burning on central Missouri prairie, Ecology, 43, 334-336, 1962. Leopold, A., and E. Jones, A phenological record for Sauk and Dane Counties, Wisconsin, 1935-1945, Ecol. Monographs, 17, 83-122, 1947. McMillan, C., Nature of the plant community. I. Uniform garden and light period studies of five grass taxa in Nebraska, Ecology, 37, 330-340, 1956a. McMillan, C., Nature of the plant community. II. Variation in flowering behavior within populations of Andropogon scoparius, Ecology, 43, 429-436, 1956b. McMillan, C., Nature of the plant community. III. Flowering behavior within two grassland communities under reciprocal transplanting, Amer. J. Botany, 44, 144-153, 1957. McMillan, C., Nature of the plant community. V. Variation within the true prairie community-type, Amer. J. Botany, 46, 418-424, 1959a. McMillan, C., The role of ecotypic variation in the distribution of the central grassland of North America, Ecol. Monographs, 29, 285-308, 1959b. McMillan, C., Ecotypes and community function, Amer. Naturalist, 94, 245-255, 1960. Myneni, R. B., C. D. Keeling, C. J. Tucker, G. Asrar, R. R. Nemani, Increased plant growth in the northern high latitudes from 1981 to 1991. Nature, 386, 698-701, 1997. Old, S. M., Microclimate, fire, and plant production in an Illinois prairie, Ecol. Monographs, 39, 355-384, 1969. Olmsted, C. E., Growth and development in range grasses. III. Photoperiodic responses in the genus Bouteloua, Botanical Gazette, 105, 165-181, 1943. Olmsted, C. E., Growth and development in range grasses. IV. Photoperiodic responses in twelve geographic strains of side-oats gramma, Botanical Gazette, 106, 46-74, 1944. Olmsted, C. E., Growth and development in range grasses. V. Photoperiodic responses of clonal divisions of three latitudinal strains of side-oats gramma, Botanical Gazette, 106, 382-401, 1945. Parrish, J. A. D., and F. A. Bazzaz, Difference in pollination niche relationships in early and late successional plant communities, Ecology, 60, 597-610, 1979. Pemble, R. H., G. L. Van Amburg, and L. Mattson, Intraspecific variation in flowering activity following a spring burn on a Northwestern Minnesota prairie, Ohio Biological Survey Biological Notes, 15, 235-239, 1981. Petersen, N. J., The effects of fire, litter, and ash on flowering in Andropogon gerardii, in Proceedings of the Eighth North American Prairie Conference, 1982, Aug 1-4, Kalamazoo, Michigan, edited by R. Brewer, pp. 21-24, Western Michigan University, Department of Biology, Kalamazoo, MI, 1983. Rabinowitz, D., J. K. Rapp, V. L. Sork, B. J. Rathcke, G. A. Reese, and J. C. Weaver, Phenological properties of wind- and insect-pollinated prairie plants, Ecology, 62, 49-56, 1981. Rice, E. L., Growth and floral development of five species of range grass in central Oklahoma, Botanical Gazette, 111, 361-377, 1950. Schwartz, M. D., Advancing to full bloom: planning phenological research for the 21st century, Int. J. Biometeorol., 42, 113-118, 1999. Schwartz, M. D. and G. A. Marotz, An approach to examining regional atmosphere-plant interaction with phenological data, J. Biogeography, 13, 551-560, 1986.

174

Phenology: An Integrative Environmental Science

Seastedt, T. R., Soil systems and nutrient cycles of the North American prairie, in The Changing Prairie, edited by A. Joern and K. H. Keeler, pp. 49-81, Oxford University Press, New York, 1995. Seastedt, T. R., J. M. Briggs, and D. J. Gibson, Controls of nitrogen limitation in tallgrass prairie, Oecologia, 87, 72-79, 1991. Seastedt, T. R., and A. K. Knapp, Consequences of non-equilibrium resource availability across multiple time scales: the transient maxima hypothesis, Amer. Naturalist, 141, 621633, 1993. Sims, P. L., Grasslands, in North American Terrestrial Vegetation, edited by M. G. Barbour, and W. D. Billings, pp. 265-286, Cambridge University Press, New York, 1988. Towne, E. G., Influence of fire frequency and burning date on the proportion of reproductive tillers in big bluestem and indiangrass, in Proceedings of the 14th Annual North American Prairie Conference, edited by D. C. Hartnett, pp. 75-78, Kansas State University, Manhattan, Kansas, 1995. Vinton, M. A., and D. C. Hartnett, Effects of bison grazing on Andropogon gerardii and Panicum virgatum in burned and unburned tallgrass prairie, Oecologia 90:374-382, 1992. Weaver, J. E., Prairie Plants and Their Environment: a Fifty-Year Study in the Midwest, University of Nebraska Press, Lincoln and London, 276 pp., 1968. Withey, A., W. Michener, and P. Tooby, Scalable Information Networks for the Environment (SINE), Report of an NSF-sponsored workshop, San Diego Supercomputer Center, October 29-31, 2001. Zedler, J. B., and O. L. Loucks, Differential burning responses of Poa pratensis fields and Andropogon scoparius prairies in central Wisconsin, Amer. Midland Naturalist, 81, 341352, 1969.

Chapter 3.4 HIGH LATITUDE CLIMATES Frans E. Wielgolaski1 and David W. Inouye2 1

Department of Biology, University of Oslo, Oslo, Norway; University of Maryland, College Park, MD, USA

Key words:

1.

2

Department of Biology,

Biocalendar, Climate change, Experimental phenology, interception, Prediction, Species-specific responses

Phenological

INTRODUCTION

High latitudes are characterized by strong variation in day-length during different seasons of the year. North of the Arctic Circle there is sun 24 hours of the day near the Summer Solstice, but no sun at all six months earlier or later. The sun angle is always low compared to further south, which means that the aspect of slopes strongly influences light conditions. Temperatures generally decrease towards the poles and the growing seasons are shorter; e.g., at the northernmost coast of Norway there are fewer than 100 days with a daily mean temperature above 5°C (Aune 1993). In many parts of northern lowland Fennoscandia, the snow-free period is less than 120 days (Björbekk 1993). Therefore, organisms living at high latitudes have to be adapted to these conditions. This means that plants have to flower relatively soon after snowmelt (Bliss 1971) in order to ripen seeds successfully. Growth of many plant species may start even before all snow has disappeared as observed in maritime Norway (Wielgolaski, pers. obs.). Heide (1985) stated that the more severe the environment, the more important survival adaptations seemed to be, while biological competition tended to be less important. It is difficult to decide where to place the southern limit of “High Latitudes.” The Arctic Circle might be one possibility, but in many ways Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 175-194 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

176

Phenology: An Integrative Environmental Science

this definition is too narrow. Summer days are long even further south (Figure 1), and plant photosynthetic activity during the important early summer goes on for many hours every day. In Europe, warm ocean currents keep temperatures higher, particularly near the west coast, compared to North America, and this of course is very important for phenology, especially in spring. Therefore in this paper “High Latitudes” are arbitrarily set to 60°N in western Eurasia, but close to 50°N in North America.

Figure 3.4-1. Variation in sun hours during a year at various high latitudes in Norway.

“Modern” phenology was started in European high latitudes by the Swedish botanist Linné (1751). He presented definitions for the study of bud break, flowering, fruit ripening and leaf coloring in autumn. He also established the first regional phenological network. His main aim was to prepare phenological plant calendars that could be cheap supplements to meteorological measurements for delineating biological zones. Ever since, this has been an important purpose of making phenological observations, particularly in the sparsely populated high latitude regions of the world. However, mainly after the First World War, phenological information was also used in agrometeorology for selection of districts for cultivation of certain crops and fruits, especially in Central Europe (Schnelle 1955), but also at higher latitudes. Many phenological networks were established after the Second World War, e.g. the International Phenological Gardens (IPG) in Europe from the

Chapter 3.4: High Latitude Climates

177

early 1960s (see Chapter 2.3 in this volume). Some stations in Fennoscandia were active in this network. Particularly during the 1970s and the 1980s, the interest in phenology decreased in large parts of the world, including Fennoscandia. Therefore, IPG observations were often not continued. A change came in the 1990s because of interest in the effect of Global Change (e.g., Lieth 1997). In this context the old phenological observations, even in remote areas of sparsely populated districts of high latitudes, could be very important (e.g., Klaveness and Wielgolaski 1996), by providing baselines for new phenological information, for example by satellite techniques (Myneni et al. 1997; Högda et al. 2001, 2002). Within the International Tundra Experiment (ITEX) several studies have been carried out in various Arctic areas to see whether experimentally changed microclimate would influence phenology and growth of plants over a few years (e.g., Arft et al. 1999).

2.

HISTORY

The first organized phenological observation program, or network, in the world was established at 18 stations in Sweden and Finland in 1751 by the Swedish botanist Carl von Linné (Johansson 1946, 1953). Most of the older data collections at high latitudes did not last for long periods, and often years are missing in between the periods of observations. However, in Finland phenological data of both plants and animals from the last part of the 1700s and the first part of the 1800s were recovered by Moberg (1857, 1894). Here, it is possible to find, e.g., that, Betula bud break occurred on 9 May 1751 in Turku and 3 May the next year; while in 1797 the same event happened on 25 June in Utsjoki, in northernmost Finland. Phenological observations at high latitudes in Europe (including Russia) have been more common since the mid-1850s. In Finland (then a country within Russia), a monitoring program was started in 1846 by the Finnish Society of Sciences and Letters and continued by the Finnish Meteorological Institute (when it was established in 1881). The Finnish phenological plant observations for the period 1896-1965 were entered into a database by Lappalainen and Heikinheimo (1992). Apparently, phenological data in Finland have been collected more widely than in other countries in Fennoscandia, although in Sweden such data have been presented monthly through many years in reports from the Swedish Meteorological and Hydrological Institute (SMHI). Mean phenological values for various parts of Sweden from 1873 to the 1920s are presented by H.W. Arnell (1923), K. Arnell (1927), and K. Arnell and S. Arnell (1930). In Norway, the first observations of plant phenology for a more scientific purpose took place starting about 1850 (Printz 1865). During the period

178

Phenology: An Integrative Environmental Science

1851-1859 phenological data of both plants and migratory birds were collected in north easternmost Norway (about 70°N). Both these data, and data from the capitol of Norway (now Oslo) for the period 1860-1884, were published by Schübeler (1885). For the end of the 1800s and beginning of the 1900s, long-term phenological observations are mainly known from the southern part of the country (Moe 1928; Lie 1931), but some observations were carried out in Troms County in northern Norway in the same period and particularly from 1910-1911 (Holmboe 1913). Later, however, in 1928, a phenological network was established that lasted to 1952 with some continuation to 1977 (Lauscher 1980; Lauscher et al. 1955, 1959, 1978; Lauscher and Lauscher 1990), and at a few Norwegian sites in the European Phenological Garden using vegetatively propagated plants (Lauscher 1985). There are no references to older phenological studies in Iceland, but recently Thórhallsdóttir (1998) has studied flowering phenology for 11 years. The best-known phenological observations on Greenland were carried out by Sörensen (1941) in the Northeast, mainly through three years in the 1930s. However, Böcher (1938) has also reported some phenological information in his plant studies from Greenland. In North America phenological observations started later than in Europe, although the field has been very active in more recent times (see Chapter 2.4 in this volume). In North American higher latitudes, the only older long-term phenological study on plants and birds was a survey in western Canada by the Royal Society of Canada from the 1890s to 1922, published annually in Proceedings and Transactions of the Royal Society (Beaubien and Freeland 2000). Since then, some phenological studies in higher North American latitudes were carried out in eastern Canada with the lilac/honeysuckle surveys of eastern USA, and some studies in western Canada (Beaubien and Johnson 1994; Beaubien 1996). However, more local phenological observations have been performed (e.g. brief overview in Erskine 1985).

3.

RECENT PHENOLOGY STUDIES

Several phenological studies have been initiated recently at high latitudes, both regionally and in international networks. Some of them are mainly descriptive and form traditional biocalendars. These can be used for agricultural planning and for educational purposes as described in more detail in other chapters of the present volume. Others are mainly experimental studies in controlled chambers and open top chambers, or transplant studies of various ecotypes, or modeling studies. Climate change is a key reason behind studies of phenology today, and the old data are of the

Chapter 3.4: High Latitude Climates

179

greatest importance for comparisons with newer ones collected in a traditional way or by satellites. Although phenological studies on plants are mainly reported here, phenological observations have also been made on animals at high latitudes, mainly on migration of birds, but on other animal groups as well (see Chapter 3.5 in this book). Forchhammer et al. (2002) reported that in northern Europe, generally migrants arrived earlier in a district after high NAO winters (mild) than after colder ones. According to Barrett (2002) and Jonzén et al. (2002), however, there has not been a significant long-term trend in the arrival dates in the north for all birds through the last 30 years of the last century.

3.1

Phenological Biocalendars and their Applications

In the tundra projects of the International Biological Program (IBP) in the early 1970s, phenological information was collected in Arctic Canada (Wielgolaski 1974; Bliss 1977). More recently, there have also been phenological studies in the Canadian Arctic (e.g., Woodley and Svoboda 1994) and in forested land in northern Canada (e.g., Colombo 1998). In Russia, phenological spectra or phenograms are given for some plant species on the Kola Peninsula for the period 1994-2000 (Makarova et al. 2001). The time of snowmelt is often considered to be the primary initiator of phenological events in tundra plants (Böcher 1938; Sörensen 1941; Wielgolaski and Kärenlampi 1975; Eriksen et al. 1993; Woodley and Svoboda 1994; Chapter 3.5 in this volume). This seems particularly to be true in the less oceanic alpine and arctic regions, as was shown experimentally by Woodley and Svoboda (1994) by snow removal at sites on Ellesmere Island, Arctic Canada. In more oceanic, snow-rich regions, however, such as the outer Troms County in northern Norway, mountain birch may have green leaves before melting of the winter snow in spring. Growth in herbaceous plants may also start before snowmelt when there is enough light through the snow, if snowmelt is slow as often occurs in maritime high-latitude climates. Thórhallsdóttir (1998) stated that time of snowmelt was likely to influence flowering only after very cold springs with exceptionally late ablation. She says that flowering in oceanic cold climates is normally not linked to snow-free conditions at all. In a survey of the flora of sub-arctic Sweden, Molau (1993) found that populations with normal pollination and seed setting flowered early, while apomictic and viviparous species were found among the later flowering ones. In short and cool summers, he found there was a low proportion of seeds ripening and a reduced seed quality and power of germination. Therefore, vegetative reproduction is very common at high latitudes (e.g.,

180

Phenology: An Integrative Environmental Science

Bliss 1971) and annuals are few, as in high-altitude regions (see Chapter 3.5 in this volume). Most of the current international phenological networks, such as the European Phenological Gardens (Chmielewski and Rötzer 2001; Chmielewski and Rötzer 2002), have few high latitude sites. One reason for this is that many of the plant species chosen for international use cannot stand the harsh climate at high latitudes. Norwegian schools have been involved since 2001 in the GLOBE network for education of school children and data have been submitted to the database from 10 schools. Other regional networks have also been established in northern latitudes in Europe, like the Norwegian Environmental Education Network launched in 2001, with about 70 participating schools by the end of 2002. Most often, the newer time series in phenology at higher latitudes have been short (e.g., Woodley and Svoboda 1994; Diekmann 1996; Karlsen et al. 1998; Wielgolaski 1999; Arft et al. 1999). It is then important that the observations be combined with other approaches, e.g. experiments, to provide insights into phenology. Another valuable approach is to collect information from stations with great differences in climate, which has to be studied at each site, and in other environmental factors, to facilitate correlations between phenology and environmental variables. For example, this approach was used in a three-year study at nearly 60 sites in western Norway along a 300km long fjord penetrating the country from west to east with strong variation from maritime to continental conditions and with steep mountains both to the north and the south of the fjord (e.g., Wielgolaski 1999, 2001). The sites included all aspects from sea level nearly to the tree line, with a climate station at each site. Several plants were studied, both native and cultivated, and woody as well as herbaceous (Wielgolaski 1999). The plants reacted differently to day and night temperatures in various developmental phenophases (Wielgolaski 1974). As expected, the plants also clearly showed different temperature requirements for development, both by plant types and phenophases (Wielgolaski 1999), but temperature requirements were generally lowest for spring phases of early plants. In that study, the author found the lowest air temperature for development (basic or threshold temperature) to leaf bud burst in Prunus padus and flowering in Salix caprea (Figure 2). Leaf bud break of the late sprouting Fraxinus excelsiorr on the other hand was found to have considerably higher basic air temperature, as did both bud break and flowering of pear (‘Moltke’). While Prunus padus showed a low basic temperature for leaf bud break, it needed a relatively high basic air temperature for flowering. Plants growing at low temperatures in high latitudes and altitudes have to be adapted to relatively low basic temperatures to be able to finish their life

Chapter 3.4: High Latitude Climates

181

cycle in a short period and/or to be restricted to the warmest places of a district. When the same crop plant variety is cultivated at different latitudes, mean annual temperature sums to reach a specific phenophase normally decreased with increasing latitude (e.g., Strand 1965). In addition to the

Figure 3.4-2. Basic air temperatures found for several plant species in western Norway from the starting date in spring, to leaf bud break and flowering (based on Wielgolaski 1999).

temperature, the photoperiod might have an influence as observed in specific varieties of grass species (e.g., Skjellvåg 1998). Although temperature clearly was the most important environmental factor for phenology in all plants studied at high latitudes, edaphic factors (Wielgolaski 2001) also played some role. Strand (1965) pointed out that in Norway heat sums for plant development in agriculture were higher in clay than in sandy soil and that fertilization also influenced the necessary heat sums. Similarly, Woodley and Svoboda (1994) in Arctic Canada found that fertilization caused an earlier flowering of Salix arctica. Water conditions are also of some importance to phenology even in high latitudes. Wielgolaski (in press) has found in his studies in western Norway that precipitation was important for the development of many plants, even in relatively oceanic areas. For instance a higher number of days with precipitation caused an acceleration of bud break in Betula pubescens. This is probably due to softening of the bud scale in moist air as was also observed by Junttila et al. (1983). Flowering of the extremely early Corylus avellana and the early Salix caprea also seemed to be favored by increased precipitation, probably for the same reasons. In later flowering species, increased precipitation most often delayed the blooming, particularly in

182

Phenology: An Integrative Environmental Science

plants found to have high basic air temperatures (Figure 2). A 12-year study of herb phenology in a Swedish temperate deciduous forest showed that precipitation may be even more important than temperature for flowering the year after (Tyler 2001). In most cases high precipitation during the previous autumn was favorable for flowering, but in Anemone nemorosa low precipitation resulted in more flowers. In Arctic Canada, Woodley and Svoboda (1994) found that irrigation of a dry riverside during the growing season caused a phenological shift in Papaver lapponicum, and an increased length of the budding and flowering periods. It is clear that various plant species react differently to diverse environmental factors (Köppen 1927), sometimes called “phenological interception”, and even in different phenophases within the same species. Therefore, timing of one specific phase in a species can vary between districts of diverse day lengths (i.e., latitude), but also different continentality (e.g., climates with high humidity and moderate temperatures can have different timing for the same phase than climates with high temperatures and lower humidity). This was obvious in the study along the fjord of western Norway (Wielgolaski, in press). Studies like this are important in phenology of all parts of the world, but more so in high latitude and altitude districts with short growing seasons and low temperatures, than in districts with less environmental variation between nearby areas. Phenological interception is found both in native plants, e. g. in leaf bud break of Fraxinus excelsior in relation to Quercus roburr (Batta 1969), and in agricultural ones. Knowledge about such variation in response to environmental factors can be a valuable tool to indicate districts with the best climate for certain plants (e.g., Wielgolaski, in press). In temperate regions, leaf and flower buds of woody plants are normally initiated in the last part of the previous summer (Kramer 1922; Guimond et al. 1998). It might, therefore, be possible to predict the timing of leaf bud burst and flowering in spring based on late phenophases the year before. Wielgolaski (2000) has carried out predictions of leaf bud break and flowering of several native and cultivated woody plants in his three-year study at several sites in western Norway. He correlated these spring phenophases with time of visible bud (mean August 21), bud with color (mean September 20), and flowering (mean September 30) of the perennial plant Aster novi-belgii in the previous autumn. In all cases, the last phase (flowering) showed the highest correlation (often highly significant), and visible bud the lowest and normally insignificant correlations with both leaf bud break and flowering of plants the year after. In Arctic Greenland, Böcher (1938) and Sörensen (1941) found that flower buds were often initiated two and more years before flowering. This indicates that energy is

Chapter 3.4: High Latitude Climates

183

built up in the short Arctic growing seasons through one or more years before it is high enough to start the flower initiation.

3.2

Experimental Phenology

In forestry it has been known for many years that transplantation of coniferous tree provenances between southern and northern latitudes has a considerable impact on both phenology and productivity of the trees (Hagem 1931, Kalela 1938, Heikinheimo 1949, Magnesen 1992, Beuker 1994). Planting in the north of southern provenances may cause continuous growth too long in the summer and, therefore, a weak hardening of the shoots by the end of the season. This often has led to frost damage during winter and spring. The well-hardened buds in coniferous trees of the northernmost ecotypes flushed earlier in spring than plants of more southern origin (Beuker 1994). It has also been observed that high latitude populations of Picea abies ended growth earliest in autumn (Morgenstern 1996). Generally, the photoperiod at the site of origin was a dominant factor in determining the timing of cessation in northern plants (Partanen and Beuker 1999). Recently, similar transplant studies have been carried out on Nordic mountain birch ((Betula pubescens ssp. czerepanovii) between oceanic and continental districts in northern Fennoscandia, and by transplantation of southern birch provenances to the north. Phenological observations over ten years have shown that the northernmost mountain birch ecotypes (from 7071°N) also ended growth earliest in the autumn when grown at the same site (Ovaska et al., in press). Oceanic northern provenances probably were somewhat earlier in ending growth in autumn than the more continental ones from similar latitudes. Both oceanic and relatively continental ecotypes of mountain birch from southern latitudes (60-64°N) showed a longer growing season when planted in more northern districts (e.g. about 68°N) than the northern provenances, being nearly green on September 10 when the northern ones were yellow and red (Ovaska et al., in press). In transplantation to oceanic districts, survival was better for oceanic provenances, while for transplantation to a continental region the survival rate was lowest in the southernmost and westernmost ecotypes where the height growth was also lower. Also, in an oceanic district, the northernmost plants were tallest after ten years. As found in coniferous plants the leaf bud break of mountain birch in spring was later in plants of southern and in particular oceanic origins, than in northern provenances transplanted and grown both in oceanic and continental districts at about 68°N (Ovaska et al., in press). The bud-burst of Betula pubescens provenances of various latitudes and continentality has also been studied in controlled climate (Myking and Heide

184

Phenology: An Integrative Environmental Science

1995; Myking 1997). It was found that plants native to mild and unstable winters (as in the south and along the coast of Norway) were released later from dormancy than those from regions with cold and stable winters (Figure 3), probably due to adaptation to avoid frost damage. The early dormancy in northern ecotypes (about 69°N), Myking (1999) stated, could be a decisive adaptation to a short growing season and hardening of the shoots at short days in time to avoid autumn frost. He also found that a period of chilling (below 10°C) was necessary for bud-burst of B. pubescens, but the natural chilling period is normally long enough even in southern and oceanic sites at latitudes above 60°N. Long photoperiods significantly reduced time to leaf bud break in partly dormant buds, but not when dormancy was fully released. However, there may be increasing experimental evidence that light conditions play some role in the timing of spring phenology (Linkosalo et al. 2000). According to Heide (1993) dormancy in B. pubescens, as well as in B. pendula and Prunus padus was released as early as December, while in Alnus sp. it was not until February.

Figure 3.4-3. Days to bud-burst in an eight-hour (short-day) photoperiod, after different periods of chilling at 5°C in Betula pubescens provenances along two gradients. a. Latitudinal gradient from 56°N (Denmark), 64°N (Mid – Norway) to 69°N (North – Norway). b. Coastalinland gradient in Norway below 150 m a.s.l. at about 60°N (reprinted from Myking 1999, Figure 2, p. 142, in Phyton, 39, used with permission).

Chapter 3.4: High Latitude Climates

185

Similarly, Hannerz (1999) suggested that a chilling requirement for Picea abies was fulfilled in December even somewhat south of 60°N in Fennoscandia. Leinonen (1996), however, generally observed that chilling requirements increased in Pinus sylvestris and Betula pendula for coastal populations compared to the more continental ones. Hänninen (1995) concluded that chilling temperature is the major environmental factor regulating rest break, but premature leaf bud break seemed not to be any serious problem for frost damage, according to models of bud-burst phenology of trees from cool and temperate regions. In North America, the arctic deciduous shrub species Salix pulchra and Betula nana were also found to have a chilling requirement before bud burst (Pop et al. 2000). According to studies in growth chambers, however, this probably played a lesser role in bud break in nature because the plants easily met their chilling requirement during winter. The ITEX (International Tundra Experiment) project was established at several sites in the Northern Hemisphere to study the influence of temperature and wind shields on vegetative and reproductive growth and development in arctic and alpine plants. Open top chambers with transparent walls were placed in the field for one to four years (Arft et al. 1999). In addition to increasing the temperature, the chambers also altered the light, moisture and gas exchange somewhat, but these side effects were minimized. It was observed that in the warmer, Low Arctic sites, the strongest response was in vegetative growth, particularly by herbaceous plants, while colder High Arctic sites produced a greater reproductive response. The better opportunities to use energy for investment in flowering and development of seeds afforded by increased temperatures in the High Arctic may provide an opportunity for species to colonize patches of bare ground (Robinson et al. 1998). At the semi-desert Svalbard site, there was a strong effect of temperature increment on the flowering of Dryas octopetala and also on seed setting (Wookey et al. 1993). While the key phenological events such as leaf bud burst and flowering occurred earlier in the warmer chambers throughout the study period, there was little impact on growth cessation at the end of the season. This may show that the photoperiod plays a more important role than temperature in late-season phenology of high latitudes (e. g., Barnes et al. 1998). Some studies, however, may also indicate a possible delayed senescence in the ITEX chambers (e.g., Molau 1997; Stenström et al. 1997).

3.3

Climate Change

Global warming is expected to cause greater increases in temperatures and precipitation during winter than summer (Dickinson 1986; Maxwell

186

Phenology: An Integrative Environmental Science

1997). The most pronounced predicted changes are in models of northern latitudes. The North Atlantic Oscillation (NAO) seems to be responsible for a large component of the increased temperature in Europe (Post and Stenseth 1999). According to Chmielewski and Rötzer (2001) the positive phase of NAO has increased clearly in Europe in the period February-April through the past several years, leading to prevailing westerly winds and thus to higher temperatures, particularly since the end of the 1980s. In most of the higher latitudes an increased winter precipitation has caused stronger snow accumulation. Despite that, increased temperature has led to earlier snowmelt and longer annual snow-free season in most regions (Maxwell 1992), an earlier and longer growing season (Bliss and Matveyeva 1992; Oechel and Billings 1992) and increased rates of plant population growth (Carlsson and Callaghan 1994). However, in areas of Fennoscandia with low winter temperatures, as in high mountain areas of Norway and inner parts of northern Fennoscandia, the higher winter precipitation (predicted to increase about 1.5% per decade

Figure 3.4-4. Change in onset of spring in Fennoscandia from 1992 to 1998 on the basis of the GIMMS NDVI dataset (reprinted from Högda et al. 2001, used with permission of the first author, NORUT IT, Tromsö).

Chapter 3.4: High Latitude Climates

187

at Finnmarksvidda according to Hanssen-Bauer et al. 2001) may have caused longer lasting snow cover in spring. This was observed by satellite inventory (GIMMS) NDVI maximum values (Myneni et al. 1997) despite higher temperatures (predicted as 0.5°C per decade at Finnmarksvidda, according to Hanssen-Bauer et al. 2000). Thus, a later onset of spring (Figure 4) was observed in some places during 17 years (1982-1998) of study (Högda et al. 2001). The strongest delay (approximately one week) occurred in the most continental areas of northern Fennoscandia and correlated well with colder April and May and with an increased snow cover. In most of Fennoscandia, the autumn was delayed and, therefore, the growing season generally increased in the region, again except for the northern continental section, where a decrease of approximately one week was observed (Högda et al. 2001). Using a similar technique it was demonstrated that starting dates of birch pollen seasons were delayed in the same regions as the delay of spring, but earlier in all other parts of Fennoscandia (Högda et al 2002). Plant responses to climate change in northern latitudes can be predicted by modeling phenological data from experiments studying the effect of temperature changes in growth chambers (e. g., Hänninen 1995; Hannerz 1999; Pop et al. 2000), studies of changes in weather variables as described above, and by models or indices of biosphere response, e. g. based on some average of plant phenology of various species (Schwartz 1997; Schwartz 1998; Chmielewski and Rötzer 2001). In high latitudes a warmer climate has caused a higher altitude tree line (Kullman 2000; Skre 2001), and an increased biodiversity is expected in many districts because of global warming, e.g. in the High Arctic due to better seed production (Philipp et al. 1990; Arft et al. 1999). Late-flowering arctic species that now only rarely ripen their seed may do so more regularly with increasing temperatures (Thórhallsdóttir 1998). Kramer et al. (2000) concluded that there are significant differences between the ways various tree species respond to climate change. Earlier in this chapter, it was stated that various plants reacted differently to environmental factors, also during various phenophases (Wielgolaski, in press). All this means that there might be rather large changes in biodiversity because of climate change, even in the more southern and temperate parts of high latitudes. By comparison of plant phenology from old data sets with the same phenophase of more recent observations at a site, it is possible to see changes that may be a result of climate change. In Norway this has been done for the first flowering date of some species in the second half of the 1800s compared with the same species in the second quarter of the 1900s (1928-1952) and the third quarter of the century (1952-1972) in three areas: south-eastern coastal district (Oslo) at about 60°N, an elevated inland district

188

Phenology: An Integrative Environmental Science

Figure 3.4-5. Scatter-plot of mean first flowering days (FFDs) in various plant species at stations in Finnmark (y-axis), 1928-52 (dots) and 1952-72 (triangles), plotted against mFFDs from Finnmark (Nyborg) 1851-59 (x-axis) (reprinted from Klaveness and Wielgolaski 1996, used with permission).

Figure 3.4-6. Scatter-plot of mean first flowering days (FFDs) in various plant species at 10 stations in Oslo plotted against mFFDs from Christiania (Oslo) 1860-84, with 95% C.I. calculated for n=25 years (reprinted from Klaveness and Wielgolaski 1996, used with permission).

Chapter 3.4: High Latitude Climates

189

of southern Norway at about 61°N, and in the far north-east at about 71°N (Klaveness and Wielgolaski 1996). The species responded differently between the periods. However, most later-flowering species were earlier in Oslo and in north-easternmost Norway (Figures 5 and 6) in the 1900s and particularly in the last period of the century. In early spring, on the other hand, there were no differences between the periods. In the inland district of southern Norway, no differences were seen between the various periods. Menzel (2000) and Chmielewski and Rötzer (2001) have reported from Central Europe that generally the leaf bud break in spring of trees from 1960 to the end of the century was more than 0.2 days/year earlier, while the autumn phases were delayed by about 0.15 days/year, causing longer growing seasons. Similarly, Beaubien and Freeland (2000) reported the first bloom of the early flowering Populus tremuloides to be 0.27 days/year earlier in a long term-study (1900-1977) at Edmonton, Canada. In Finland, however, phenological data did not permit reliable estimates of the effect of climate change on spring in boreal trees (Linkosalo 2000).

4.

CONCLUSION

At high latitudes, there are large differences between species-specific responses to environmental factors. The responses also vary between geographical districts or continentality and within a species at different times of the year. In many cases phenology may be used in climate change studies, but then there must be a clear description of the sites used in the study: geographically, climatically and edaphically, as well as clear definitions of the phenophases studied and the state of the organism.

REFERENCES CITED Arft, A. M., M. D. Walker, J. Gurevitch, J. M. Alatalo, M. S. Bret-Harte, M. Dale, M. Diemer, F. Gugerli, G. H. R. Henry, M. H. Jones, R. D. Hollister, I. S. Jonsdottir, K. Laine, E. Levesque, G. M. Marion, U. Molau, P. Mölgaard, U. Nordenhäll, V. Raszhivin, C. H. Robinson, G. Starr, A. Stenström, M. Stenström, Ö. Totland, P. L. Turner, L. J. Walker, P. J. Webber, J. M. Welker, and P. A. Wookey, Responses of tundra plants to experimental warming: Meta-analyses of the International Tundra Experiment, Ecol. Monographs 69, 491-511, 1999. Arnell, H. W., Vegetationens aarliga utvecklingsgaang i Svealand. Medd. Statens Meteorologisk - Hydrografiska Anstalt, 2, 1-80, (German abstr. 74-80), 1923. Arnell, K., Vegetationens utvecklingsgaang i Norrland. Medd. Statens Meteorologisk – Hydrografiska Anstalt, 4, 1-28, (German abstr. 28), 1927. Arnell, K., and S. Arnell, Vegetationens utveckling i Götaland. Medd. Statens Meteorologisk – Hydrografiska Anstalt, 6, 1-70, (German abstr. 69-70), 1930.

190

Phenology: An Integrative Environmental Science

Aune, B., Aarstider og vekstsesong. Kartblad (Map) 3.1.7., scale 1:7 mill., in Nasjonalatlas for Norge, Det norske meteorologiske institutt – Statens kartverk, Oslo – Hönefoss, 1993. Barnes, B. V., D. R. Zak, S. R. Denton, and S. H. Spurr, Forest tree variation, in Forest Ecology, pp. 63-93, John Wiley, New York, 1998. Barrett, R.T., The phenology of spring bird migration to north Norway, Bird Study, 49, 270277, 2002. Batta, J., Variasjoner i tid for bladsprett hos ask og eik, (Engl. summary), Aarsskr. Planteskoledrift Dendrologi, 14-15, 78-85, 1969. Beaubien, E. G., Plantwatch, a model to initiate phenology in school classes, Phenol. Season., 1, 33-35, 1996. Beaubien, E. G., and H. J. Freeland, Spring phenology trends in Alberta, Canada: links to ocean temperature, Int. J. Biometeorol., 44, 53-59, 2000. Beaubien, E.G., and D. L. Johnson, Flowering plant phenology and weather in Alberta, Canada, Int. J. Biometeorol., 38, 23-27, 1994. Beuker, E., Adaptation to climatic changes of the timing of bud burst in populations of Pinus sylvestris L. and Picea abies (L.) Karst., Tree physiology, 14, 961-970, 1994. Björbekk, G., Snö. Kartblad (Map) 3.1.4., scale 1.7 mill., in Nasjonalatlas for Norge, Det norske meteorologiske – Statens kartverk, Oslo – Hönefoss, 1993. Bliss, L.C., Arctic and alpine life cycles, Ann. Rev. Ecol. Syst., 2, 405-438, 1971. Bliss, L.C. (Editor), Truelove Lowland, Devon Island, Canada: A High Arctic Ecosystem, University of Alberta Press, Edmonton, 714 pp., 1977. Bliss, L.C., and N. V. Matveyeva, Circumpolar arctic vegetation, in Arctic Ecosystems in a Changing Climate, edited by F. S. Chapin III, R. L. Jefferies, J. F. Reynolds, G. R. Shaver and J. Svoboda, pp. 59-89, Academic, New York, 1992. Böcher, T. W., Studies on the vegetation of the east coast of Greenland, Medd. Grönl., 104, 132, 1938. Carlsson, B.A., and T. V. Callaghan, Impact of climate change factors on the clonal sedge Carex bigelowii: implication for population growth and vegetative spread, Ecography, 17, 321-330, 1994. Chmielewski, F.-M., and T. Rötzer, Response of tree phenology to climate change across Europe, Agricul. Forest Meteorol., 108, 101-112, 2001. Chmielewski, F.-M., and T. Rötzer, Annual and spatial variability of the beginning of growing season in Europe in relation to air temperature changes, Clim. Res., 19, 257-264, 2002. Colombo, S. J., Climatic warming and its effect on bud burst and risk of frost damage to white spruce in Canada, For. Chron., 74, 567-577, 1998. Dickinson, R., The climate system and modelling of future climate, in The Greenhouse Effect, Climate Change, and Ecosystem, edited by B. Bolin, B. Doos, J. Jager and R.A. Warrick, pp. 207-270, Wiley, Chichester, UK, 1986. Diekmann, M., Relationship between flowering phenology of perennial herbs and meteorological date in deciduous forests of Sweden, Can. J. Bot. 74, 528-537, 1996. Eriksen, B., U. Molau, and M. Svensson, Reproductive strategies in two arctic Pedicularis species (Scrophulariaceae), Ecography, 16, 154-166, 1993. Forchhammer, M. C., E. Post, and N. C. Stenseth, North Atlantic Oscillation timing of longand short-distance migration, J. Animal Ecol., 71,1002-1014, 2002. Erskine, A. J., Some phenological observations across Canada’s boreal regions, Can. FieldNat., 99, 188-195, 1985. Guimond, C. M., P. K. Andrews, G. A. Lang, Scanning electron microscopy of floral initiation in sweet cherry, J. Am. Soc. Hort. Sci., 123, 509-512, 1998.

Chapter 3.4: High Latitude Climates

191

Hagem, O., Forsök med vestamerikanske traeslag, (German summary), Medd. Vestl. Forstl. Forst., 12, 1-217, 1931. Hannerz, M., Evaluation of temperature models for predicting bud burst in Norway spruce, Can. J. For. Res., 29, 9-19, 1999. Hänninen, H., Effects of climate change on trees from cool and temperate regions: an ecophysiological approach to modelling of bud burst phenology, Can. J. Bot. 73, 183-199, 1995. Hanssen-Bauer, I., O. E. Tveito, and E. J. Förland, Temperature scenarios for Norway: Empirical downscaling from the ECHAM4/OPYC3 GSDIO integration, Det norske meteorologiske institutt, Report 24, 1-53, 2000. Hanssen-Bauer, I., O. E. Tveito, and E. J. Förland, Precipitation scenarios for Norway: Empirical downscaling from the ECHAM4/OPYC3 GSDIO integration, Det norske meteorologiske institutt, Report 10, 1-39, 2001. Heide, O. M., Physiological aspects of climatic adaptation in plants with special references to high-latitude environments, in Plant Production in the North, edited by Aa. Kaurin, O. Junttila and J. Nilsen, pp. 1-22, Norwegian University Press, Tromsö, 1985. Heide, O. M., Daylength and thermal time response of budburst during dormancy release in some northern deciduous trees, Physiol. Plant., 88, 531-540, 1993. Heikinheimo, O., Results of the experiments on the geographical races of spruce and pine, (in Finnish with English summary), Comm. Inst. For. Fenn., 37, 1-44, 1949. Högda, K. A., S. R. Karlsen, and I. Solheim, Climatic change impact on growing season in Fennoscandia studied by a time series of NOAA AVHRR NDVI data, Proc. IGARSS 2001, Sydney, Australia, ISBN 0-7803-7033-3, pp. 1-3, 2001. Högda, K. A., S. R. Karlsen, I. Solheim, H. Tömmervik, and H. Ramfjord, The start dates of birch pollen seasons in Fennoscandia studied by NOAA AVHRR NDVI data, Proc. IGARSS 2002, Toronto, Canada, ISBN 0-7803-7536-X, pp. 1-3, 2002. Holmboe, J., Vaarens utvikling i Tromsö amt, (in Norwegian), Bergens Mus. Aarb., 1912, 1248, 1913. Johansson, O. V., Det fenologiska observationsmaterialet i Finland och provstudier av detsamma, (in Swedish), Finlands Natur och Folk, 88(8), 1-118, 1946. Johansson, O. V., Die Phänologie in Finland, Soc. Sci. Fenn., Comm. Biol., 11(1), 1-55, 1953. Jonzén, N., A. Hedenström, C. Hjort, Å. Lindström, P. Lundberg, and A. Andersson, Climate patterns and the stochastic dynamics of migratory birds, Oikos, 97, 329-336, 2002. Junttila, O., C. Stushnoff, and L. V. Gusta, Dehardening in flower buds of saskatoon-berry, Amelanchier alnifolia, in relation to temperature, moisture content, and spring bud development, Can. J. Bot., 61, 164-170, 1983. Kalela, A., Zur Synthese der experimentellen Untersuchungen über Klimarassen der Holzarten, Comm. Inst. For. Fenn., 26, 1-445, 1938. Karlsen, S. R., H. Tömmervik, I. Solheim, B. Johansen, K. A. Högda, and O. Engelsen, Satellite based phenology study in eastern part of Finnmark, northern Norway, NORUT IT 450/15-98, ISBN 82-7747-094-0, pp. 1-35, NORUT, Tromsö, 1998. Klaveness, D., and F. E. Wielgolaski, Plant phenology in Norway – a summary of past and present first flowering dates (FFDs) with emphasis on conditions within three different areas, Phenol. Season., 1, 47-61, 1996. Köppen, W., Wechsel der phänologischen Zeitenfolge, Meteorol. Z., 44, 175-177, 1927. Kramer, K., I. Leinonen, and D. Loustau, The importance of phenology for the evaluation of impact of climate change on growth of boreal, temperate and Mediterranean forests ecosystems: an overview, Int. J. Biometeorol., 44, 67-75, 2000.

192

Phenology: An Integrative Environmental Science

Kramer, O., Über die Blütenknospen und der Zeitpunkt der Entstehung von Blütenanlagen bei einigen Obstsorten, Dtsch. Obstbauztg., 68, 306-308. 1922. Kullman, L., Tree-limit rise and recent warming: a geoecological case study from the Swedish Scandes. Norsk geogr. Tidsskr., 54, 49-59, 2000. Lappalainen, H., and M. Heikinheimo, Relations between climate and plant phenology. Part 1. Survey of plant phenological observations in Finland from 1896 to 1965, Meteorol. Publ. Finland, 20, 1-74, 1992. Lauscher, F., Klima, Klimaschwankungen und phänologischer Jahresablauf am europäischen Nordkap, Mitt. Österr. Geogr. Ges., 122, 193-220, 1980. Lauscher, F., Zur Phänologie vegetativ vermehrter Pflanzen einheitlicher Herkunft – Beobachtungen in phänologischen Pflanzgärten in Norwegen 1963-1982, Phyton (Horn, Austria), 25, 253-272, 1985. Lauscher, A., and F. Lauscher, Phänologie Norwegens, Teil IV, Private edition, pp. 1-13 + Anhang 1-3 (16+103+120 pp.), 1990. Lauscher, A., F. Lauscher, and H. Printz, Die Phänologie Norwegens, Teil I, Allgemeine Übersicht, Skr. Det Norske Videnskaps-Akademi Oslo, 1, Mat.-Naturv. Kl., No. 1 1955, 199, 1955. Lauscher, A., F. Lauscher, and H. Printz, Die Phänologie Norwegens, Teil II, Phänologische Mittelwerte für 260 Orte, Skr. Det Norske Videnskaps-Akademi Oslo, 1, Mat.-Naturv. Kl. No.1 1959, 1-176, 1959. Lauscher, A., F. Lauscher, and H. Printz, Die Phänologie Norwegens, Teil III, TabellenKarten der Mittelwerte, Skr. Det Norske Videnskaps-Akademi Oslo, 1, Mat.-Naturv. Kl. Ny Serie No.37, 1-253, 1978. Leinonen, I., Dependence of dormancy release on temperature in different origins of Pinus sylvestris and Betula pendula seedlings, Scand. J. For. Res., 11, 122-128, 1996. Lie, H., Faenologiske noteringar fraa Telemark, (in Norwegian), Tidsskr. Norske Landbruk, 38, 204-206, 1931. Lieth, H., Aims and methods in phenological monitoring, in Phenology in Seasonal Climates II, edited by H. Lieth and M. D. Schwartz, pp. 1-21, Backhuys Publ., Leiden, 1997. Linkosalo, T., Analyses of the spring phenology of boreal trees and its response to climate change, Univ. Helsinki Dept. For. Ecol., Publ. 22, ISBN 951-45-9362-6, ISSN 1235-4449, 1-55. 2000. Linkosalo, T., T. R. Carter, R. Häkkinen, and P. Hari, Predicting spring phenology and frost damage risk of Betula spp. Under climatic warming: a comparison of two models, Tree Physiol., 20, 1175-1182, 2000. Linne, C., Philosophia Botanica, (in Latin), Kiesewetter, Stockholm, Climate and phenology pp. 263-277, 1751. Magnesen, S., Injuries on forest trees related to choice of the species and provenances: A literature survey of a one hundred year epoch in Norwegian forestry, Rep. Skogforsk, 7, 146, 1992. Makarova, O.A., A. A. Pohilko, and J. A. Kushel, Seasonal life of the nature in Kola Peninsula, (in Russian, translated in English), Murmansk, ISBN 5-7744-0102-2, 1-68, 2001. Maxwell, B., Arctic climate: potential for change under global warming, in Arctic Ecosystems in a Changing Climate, edited by F. S. Chapin III, R. L. Jefferies, J. F. Reynolds, G. R. Shaver and J. Svoboda, pp. 11-34, Academic Press, New York, 1992. Maxwell, B., Recent climate patterns in the Arctic, in Global Change and Arctic Terrestrial Ecosystems, edited by W. C. Oechel, T. Callaghan, T. Gilmanov, J. I. Holten, B. Maxwell, U. Molau and B. Sveinbjörnsson, pp. 21-46, Springer-Verlag, Heidelberg, 1997.

Chapter 3.4: High Latitude Climates

193

Menzel, A., Trends in phenological phases in Europe between 1951 and 1996, Int. J. Biometeorol. 44, 76-81, 2000. Moberg, A., Naturalhistoriska daganteckningar gjorda i Finland aaren 1750-1845, (in Swedish), Förh. Sällsk. Fauna Flora Fenn., 3, 95-250, 1857. Moberg, A., Fenologiska iakttagelser i Finland aaren 1750-1845 (in Swedish), Finlands Natur Folk, 55, 1-165, 1894. Moe, A., Dates of flowering for native and garden plants at Stavanger 1897-1926, Skr. Det Norske Videnskaps-Akademi Oslo, 1, Mat.-Naturv. Kl. No.3, 1-50, 1928. Molau, U., Relationship between flowering phenology and life history strategies in tundra plants, Arctic Alpine Res., 25, 391-402, 1993. Molau, U., Responses to natural climatic variation and experimental warming in two tundra plant species with contrasting life forms: Cassiope tetragona and Ranunculus nivalis, Global Change Biology, 3, (Suppl. 1), 97-107, 1997. Morgenstern, E. K., Environmental influences and geographic variation, in Geographic variation in forest trees, pp.45-89, UBC Press, Vancouver, 1996. Myking, T., Dormancy, budburst and impacts of climatic warming in coastal-inland and altitudinal Betula pendula and B. pubescens ecotypes, in Phenology in Seasonal Climates I, edited by H. Lieth and M. D. Schwartz, pp. 51-66, Backhuys Publ., Leiden, 1997. Myking, T., Winter dormancy release and budburst in Betula pendula ROTH and B. pubescens EHRH. ecotypes, Phyton (Horn, Austria), 39(4), 139-145, 1999. Myking, T., and O. M. Heide, Dormancy release and chilling requirement of buds of latitudinal ecotypes of Betula pendula and B. pubescens, Tree Physiol. 15, 697-704, 1995. Myneni, R. B., C. D. Keeling, C. J. Tucker, G. Asrar, and R. R. Nemani, Increased plant growth in the northern latitudes from 1981 to 1991, Nature, 386, 698-702, 1997. Oechel, W. C., and W. D. Billings, Effects of global change on the carbon balance of arctic plants and ecosystems, in Arctic Ecosystems in a Changing Climate, edited by F.S. Chapin III, R. L. Jefferies, J. F. Reynolds, G. R. Shaver and J. Svoboda, pp. 139-168, Academic Press, New York, 1992. Ovaska, J. A., J. Nilsen, F. E. Wielgolaski, H. Kauhanen, R. Partanen, S. Neuvonen, L. Kapari, O. Skre, and K. Laine, Phenology and performance of mountain birch provenances in transplant gardens: latitudinal, altitudinal and oceanity-continentality gradients, in Plant Ecology, Herbivory and Human Impact in Nordic Mountain Birch Forests, edited by F. E. Wielgolaski (in press), Springer-Verlag, Heidelberg, (anticipated in 2004). Partanen, J., and E. Beuker, Effects of photoperiod and thermal time on the growth rhythm of Pinus sylvestris seedlings, Scand. J. For. Res., 14, 487-497, 1999. Philipp, M., J. Böcher, O. Mattson, and S. R. J. Woodell, A quantitative approach to the sexual reproductive biology and population structure in some arctic flowering plants: Dryas integrifolia, Silene acaulis and Ranunculus nivalis, Medd. Grönland, Biosci., 34, 160, 1990. Pop, E. W., S. F. Oberbauer, G. Starr, Predicting vegetative bud break in two arctic deciduous shrub species, Salix pulchra and Betula nana, Oecologia, 124, 176-184, 2000. Post, E., and N. C. Stenseth, Climatic variability, plant phenology, and northern ungulates, Ecology, 80, 1322-1339, 1999. Printz, H. C., Beretning om en i Sommeren 1864 foretagen botanisk Reise i Valders, (in Norwegian), Nyt Mag. Naturv., 14, 51-96, 1865. Robinson, C. H., P. A. Wookey, J. A. Lee, T. V. Callaghan, and M. C. Press, Plant community responses to simulated environmental change at a High Arctic polar semidesert, Ecology, 79, 856-866, 1998.

194

Phenology: An Integrative Environmental Science

Rötzer, T., and F. -M. Chmielewski, Phenological maps of Europe, Clim. Res., 18, 249-257, 2001. Schnelle, F., Pflanzen-Phänologie, Geest & Portig, Leipzig, 299 pp., 1955. Schübeler, F. C., Viridarium Norvegicum, Norges Vaextrige, Et Bidrag til Nord-Europas Natur-og Kulturhistorie, 1ste bind, (in Norwegian), Phenology etc., pp. 1-184, Climatic, pp. 185-195, W. C. Fabritius, Christiania, 610 pp., 1885. Schwartz, M. D., Spring index models: an approach to connecting satellite and surface phenology, in Phenology in Seasonal Climates II, edited by H. Lieth and M. D. Schwartz, pp. 23-38, Backhuys Publ., Leiden, Netherlands, 1997. Schwartz, M. D., Green-wave phenology, Nature, 394, 839-840, 1998. Skjellvåg, A. O., Climatic conditions for crop production in Nordic countries, Agr. Food Sci. Finland, 7, 149-160, 1998. Skre, O., Climate change impact on mountain birch ecosystems, inn Nordic Mountain birch ecosystems, edited by F. E. Wielgolaski, pp. 343-357, UNESCO, Paris and Parthenon Publ. Group, New York and London, 2001. Sörensen, T., Temperature relations and phenology of the Northeast Greenland flowering plants, Medd. Grönl., 125, 1-307, 1941. Stenström, M., F. Gugerli, and G. H. R. Henry, Response of Saxifraga oppositifolia L. to simulated climate change at three contrasting latitudes, Global Change Biology, 3, (Suppl. 1), 44-54, 1997. Strand, E., Forelesning i plantekultur, (in Norwegian), Norges landbrukshögskole, Aas, 73 pp., 1965. Thórhallsdóttir, T. E., Flowering phenology in the central highland of Iceland and implications for climatic warming in the Arctic, Oecologia, 114, 43-49, 1998. Tyler, G., Relationships between climate and flowering of eight herbs in a Swedish deciduous forest, Ann. Bot. 87, 623-630, 2001. Wielgolaski, F. E., Phenology in agriculture in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 369-381, Springer-Verlag, New York, 1974. Wielgolaski, F. E., Starting dates and basic temperatures in phenological observations of plants, Int. J. Biometeorol., 42, 158-168, 1999. Wielgolaski, F. E., Predictions in plant phenology, paper presented at Int. Congress: Progress in Phenology, Freising, Germany, October 2000. Wielgolaski, F. E., Phenological modifications in plants by various edaphic factors, Int. J. Biometeorol., 45, 196-202, 2001a. Wielgolaski, F. E., Climatic factors governing plant phenological phases along a Norwegian fjord, Int. J. Biometeorol., 47(4), in press, 2003. Wielgolaski, F. E., and L. Kärenlampi, Plant phenology of Fennoscandian tundra areas, in Fennoscandian Tundra Ecosystems. Part1: Plants and Microorganisms, edited by F. E. Wielgolaski, pp. 94-102, Springer-Verlag, Heidelberg, 1975. Woodley, E.J., Svoboda, J. Effects of habitat on variations of phenology and nutrient concentration among four common plant species of the Alexandra Fiord Lowland, in Ecology of a Polar Oasis, Alexandra Fiord, Ellesmere Island, Canada, edited by J. Svoboda and B. Freedman, pp. 157-175, Captus Univ. Press, Toronto, 1994. Wookey, P. A., A. N. Parsons, J. M. Welker, J. A. Potter, T. V. Callaghan, and M. C. Press, Comparative responses of phenology and reproductive development to simulated environmental change in sub-arctic and high arctic plants, Oikos, 67, 490-502, 1993.

Chapter 3.5 HIGH ALTITUDE CLIMATES David W. Inouye1 and Frans E. Wielgolaski2 1

Department of Biology, University of Maryland, College Park, MD, USA; 2Department of Biology, University of Oslo, Oslo, Norway

Key words:

1.

Alpine, Montane, Snowpack, Subalpine, Rocky Mountains

INTRODUCTION

Phenology at high altitudes differs from that in most other habitats in four significant ways. First, for much (sometimes the majority) of the calendar year these habitats may be under snow or ice, and there is little photosynthetic activity. Consequently (and second), there is a very short growing season delimited by a combination of temperature and snowpack. Third, this may be one of a few habitats where almost all phenology is tied to a single highly variable event, the timing of snowmelt; few high-altitude plants appear to exhibit photoperiodic responses for phenological events. And finally, high altitudes may differ from other habitats in the way that global climate change is affecting phenology. What is a high altitude? The answer is not as obvious as it might seem. It probably makes more sense to use an ecosystem definition rather than an absolute altitude, as what constitutes a high altitude at high latitudes differs from a high altitude at mid- or low latitudes. For the purposes of this chapter we will consider “high altitude” to refer to alpine or montane ecosystems. Alpine is defined as the area above the natural limit of trees, and it extends over a wide latitudinal and altitudinal range. Another way of defining it, in climatic terms, is that its lower elevational limit corresponds well to the 10°C isotherm for the warmest summer month (Wardle 1974). The alpine Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 195-214 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

196

Phenology: An Integrative Environmental Science

shares many characteristics with high-latitude or arctic ecosystems (Bliss 1971, and see Chapter 3.4 in this book), but from a phenological perspective one big difference is the much longer day lengths during the arctic growing season. Montane ecosystems are less clearly defined, but would include the ecosystem between grasslands on the lower end and the alpine on the upper end; the term subalpine applies to the upper end of the montane. There are not as many long-term studies of phenology at high altitudes as there are of low altitudes. In fact, there do not appear to be any studies longer than the one described in this chapter that was initiated in 1973. Thus some of the discussion in this chapter will be colored by the fact that Inouye has worked for most of his research career at a single high-altitude field station, the Rocky Mountain Biological Laboratory (RMBL). Perhaps the information in this chapter will help to stimulate the initiation of other studies.

2.

THE HIGH-ALTITUDE CLIMATE

Mountains have been described as “generating their own climate”, due to the effect of their mass on circulation patterns, precipitation, and radiation. This creates abundant variation of the climate within mountain regions, but also some general patterns that help to differentiate high altitude climate, and hence phenology, from that at lower altitudes. Kittel et al. (2002) go into detail about elevation dependence of climate, which can be attributed to factors at high altitudes such as closer contact to the free troposphere, decoupling from convective mixing of the lower troposphere, and snowalbedo feedback. Within the Rocky Mountains, there is a classic orographic precipitation pattern that creates increased precipitation on the west (windward) side of the mountains and a rain shadow on the east, although some winter air masses create up-slope conditions on the east side that generate precipitation there. In all areas of the Rocky Mountains, total precipitation increases significantly with elevation (Kittel et al. 2002), with the highest precipitation occurring near the peak elevations (this is true for the Alps also, Theurillat and Guisan 2001, and in the Scandinavian mountains). Maximum and minimum temperatures decrease strongly with increasing altitude throughout the Rocky Mountains (Kittel et al. 2002), although there are some interesting variations. For example, variation of maximum temperatures decreases with elevation in the winter while that of minimum temperatures increases during the summer, and at lower elevations in the central Rockies, temperatures can be colder than those at similar elevations in the north (due to the influence of continental vs. maritime air masses).

Chapter 3.5: High Altitude Climates

197

An analysis of historical data shows that there are significant centuryscale positive trends for annual and seasonal precipitation and mean minimum temperature in the U.S. Rocky Mountains (Kittel et al. 2002). Some of these are quite striking; in the northern and central Rockies summer precipitation has increased by 30% and 33% over the past century. There has also been a trend for increasing annual mean minimum temperature (0.7 – 0.9ºC for the northern and central mountains). It is probable that these changes, and those forecast for the future, will have consequences for phenology at high altitudes. Giorgi et al. (1997) predicted similar changes for temperature over an altitudinal range in the Swiss Alps, and suggested “... that high elevation temperature change could be used as an early detection tool for global warming.” Both large- and small-scale events can affect high-altitude phenology. For example, the El Niño Southern Oscillation (ENSO) and the North Pacific Oscillation (or Pacific Decadal Oscillation) can affect winter precipitation in the Rocky Mountains. At the other extreme, microclimate can have very large effects. Areas where snow is deposited by wind (on the lee side of ridges, trees, etc.) can melt out much later than nearby sites, and deposition by snow slides and avalanches may create such deep snow depths that certain areas may not melt out at all in a given summer. Wagner and Reichegger (1997) found that a north-facing study site subject to deep snow deposition took about a month longer to melt out than sunlit sites. The cold water from melting snowbanks can have an effect on the phenology of plants it reaches. Holway and Ward (1963) found in an experimental study that meltwater resulted in delays of flowering in 12 of 14 species growing in irrigated plots (typically of about a week, but up to a month). Another aspect of high-altitude phenology that is unique to areas with great topographic relief is the potential for phenological inversions. Thermal inversions through cold air drainage can have phenological consequences. Lynov (1984) reported statistically significant effects of such phenological inversions for eight species of trees and shrubs, with delays of 2 – 5 days in times of bud opening and flowering. Areas where this cold air collects are also described as frost hollows or frost pockets. In the southern Colorado Rocky Mountains the ecosystem at 2900 m is sub-alpine, or montane, as trees are still common at this altitude. The date of first permanent winter snowpack at the RMBL averages about 4 November (range 15 October-24 November, data from 1974-2001), and the length of snowcover is about 200 days (range 165-233, data from 1975-2001). The mean date of first bare ground at a permanent snow measurement station is 21 May, with a range of 25 May to 19 June (data from 1975-2002). Summer precipitation also appears to play a role in some aspects of phenology at the RMBL. Precipitation for June-August at the NOAA weather station in

198

Phenology: An Integrative Environmental Science

Crested Butte (2704 m, 9.5 km from RMBL) has averaged 13.5cm (range 5.3-22.3 cm, data from 1973-2001; the data for 2002 will probably set a new record low). The influence of climate on the growing season, and hence on phenology, can also be seen by transplant experiments. Plants transplanted to lower altitudes will typically develop much sooner than those left in their native high-altitude sites (e.g., Wagner and Reichegger 1997).

3.

LITERATURE REVIEW

Perhaps more than any other bioclimatic zone except high latitudes, phenological events at high altitudes are constrained by a short growing season, delimited by cold temperatures and snowpack. Time of snowmelt appears to have an almost universal effect on high-altitude phenology, and variation in phenology can usually be linked to variation in accumulation and then melting of snow, whether this is across time or space. This interaction has been reported by many studies, including Bliss (1956), Holway and Ward (1963, 1965), and Mark (1970). Canaday and Fonda (1974) found that the timing and duration of phenophases of a variety of subalpine plants in the Olympic Mountains (WA) were a function of snowmelt. In general terms Ratcliffe and Turkington (1989) found the same results, although they argue that the identity of dominant species and, to a lesser degree, aspect, are responsible for variations they observed in phenology. Some species tended to flower earlier on south-facing slopes, indicating the potential importance of aspect and microenvironment. Most studies of phenology at high altitudes have been short (e.g., (Wielgolaski and Kärenlampi 1975); such studies can probably define relatively well the spatial pattern of snowmelt and hence phenology in a particular site, but longer studies are required to gain insights into the effects of climate variables on phenology. In one relatively long study, Walker et al. (1995) followed the phenology of two forbs for six years in five different plant communities and found significant differences among years and plant communities. Phenophase condensation has been observed in several studies of alpine plants, with full development being accomplished more rapidly where snow persists longer. Examples of this have been reported by Knight et al. (1977, and references therein), and Billings and Bliss (1959). Snowmelt gradients are a common phenomenon at high altitude, as some areas will receive more or less snow and receive more or less insolation and result in earlier or later snowmelt. The consequences of these gradients have been investigated in several studies. Kudo (1992) investigated five herbaceous species along a

Chapter 3.5: High Altitude Climates

199

snowmelt gradient on Mt. Kaun in the Taisetsu Mountains of Hokkaido, where the snowfree period ranged from 55 to 95 days. The later snowmelt occurred, the later flowering and fruiting began, and in the plot that melted out last no species was able to mature all fruits because of the short growing season. In a study of 56 species over three years, Kudo also found that a shorter snow-free period reduced flowering and seeding (Kudo 1991). Plants found in alpine tundra are often remarkable for the speed with which they can flower and fruit, but this adaptation is required for success given the short growing season (Bliss 1971; Wielgolaski and Kärenlampi, 1975). This early flowering is facilitated by the fact that floral initiation often occurs one or more years in advance of flowering; preformation of flower buds is characteristic of many high-altitude plants (e.g., Resvoll 1917; Forbis and Diggle 2001; Meloche and Diggle 2001, and see references in Bliss 1971). The climate constraints of high altitudes could result in significant selection against late flowering, to allow sufficient time for the development of seeds before the first killing frost in the fall, and this may be why annual and even biennial plants are relatively uncommon at high altitudes (Bliss 1971; Jackson and Bliss 1982). Annuals may be restricted to sites that melt out early or that don’t dry out early in the summer (Reynolds 1984). Some studies do show that most species in alpine communities initiate flowering rapidly after snowmelt, with relatively few species flowering late (e.g., Holway and Ward 1965; Billings and Mooney 1968; Totland 1993). In a few cases, plants may be able to initiate growth under the snow and get a head start on the growing season. Billings and Bliss (1959) found Geum turbinatum, Carex elynoides, and Deschampsia caespitosa growing under 15 cm of snow near the edge of a melting snowbank, and observed that the latter two of these had started growth under 50 cm of snow that did not become snow free for another four days. Young red leaves of Polygonum bistortoides were found under 110 cm of old snow (Mooney et al. 1981). Williams and Cronin (1968) found that Delphinium species could emerge and develop green cotyledons when snow melted to a depth of 30 cm or less, and Spomer (cited in Richardson and Salisbury 1977) found green plants of Ranunculus adoneus under 1m of snow. Arroyo et al. (1981) recorded the earliest-flowering alpine species in the Chilean Andes as actually blooming precociously under 5-6 cm of snow, and Bliss (1971) reported flowering by species of Caltha and Ranunculus under 10 cm or more of snow. Theurillat and Schlüssel (2000) studied seven subalpine-alpine species in the Alps and characterized them by the number of degree-days to bud burst and the end of flowering. Each species differed in its requirements, but only Vaccinium myrtillus was closely tied to snowmelt, while the others fit heat sum models that depended on degree-days after a chilling requirement,

200

Phenology: An Integrative Environmental Science

intensity of chilling from a threshold, or constant degree-days following thaw (Heide 1993; Myking and Heide 1995). It would be interesting to have some studies that have compared models of heat sums and time since snowmelt to see which works best for predicting phenology of a variety of species.

3.1

Temporal, Spatial, and Altitudinal Gradients

Plants growing in microsites with shorter growing seasons may be able to compensate somewhat by shortening the time before flowering starts, allowing them to complete seed production before the weather becomes unfavorable. Jackson and Bliss (1984) studied Polygonum minimum, a very small (<2 cm tall) subalpine annual plant at 3,000m in the Sierra Nevada. Its indeterminate growth pattern, and the fact that it can dehisce seeds in as little as two months after snowmelt, help it to survive the phenological constraints of its habitat. The indeterminate growth pattern permits plants to take advantage of longer growing seasons for increased seed production. Another example is provided by Carex species in the Austrian Alps (Wagner and Reichegger 1997). However, if the microsite differences in snowmelt lead to significant differences in flowering time, the end result may be fitness differences among individuals flowering at different times, and eventually, divergence. This kind of differentiation has been seen in mountain environments in Japan (Kudo 1991), Colorado (Galen and Stanton 1991) and North Swedish Lapland (Stenström and Molau 1992). Galen and Stanton (1991) found fitness differences related to emergence phenology; plants in the latestmelting parts of snowbeds produced smaller seeds than earlier-flowering plants, and seed size was correlated with seedling survival rates. Despite the potential advantages of early flowering, some species do seem to be able to withstand the constraints of late flowering. For example, Gentianella caucasea is an annual species found growing up to the subniveal zone in the Central Caucasus that can complete its life cycle with four to five weeks (Akhalkatsi and Wagner 1996). Different microclimates appear to have resulted in different phenotypes with significant differences in developmental times. Populations at the lower end of the altitudinal range take at least 24-25 days for floral bud formation, flowering, and fruit ripening, while those at high elevations flower about three weeks later than the earliest lower population and take only 18 days to complete reproduction. Gentianella germanica, a biennial species found at high altitudes in Europe and the Balkans, is unusual for its late flowering (August to the beginning of November when winter begins (Wagner and Mitterhofer 1998). Phenological differences in flowering between two morphs of this

Chapter 3.5: High Altitude Climates

201

species may be leading to their isolation. The Norwegian alpine species Leontodon autumnalis var. taraxaci also flowers late (Totland 1997). Elevational gradients provide an interesting opportunity for phenological studies at high altitudes. Schuster et al. (1989) studied gene flow in limber pine (Pinus flexilis) over its altitudinal range (1650 m to 3350 m) in Colorado, and found that pollination phenology is strongly affected by altitude. Sites that differed in elevation by more than 400m did not usually have overlapping pollination periods, so gene flow by pollination was quite restricted. Mitton et al. (1980) found similar results for ponderosa pine trees ((Pinus ponderosa). Hoffmann and Walker (1980) looked at phenology of two drought-deciduous shrubs along an altitudinal gradient on the Coastal and Andean Ranges of Chile. Vegetative growth at the upper limit (2000 m) started nine weeks later than at the lowest altitude (700 m), and lasted longer at the lower altitude, but flowering and fruiting occurred at about the same time at both ends of the gradient (perhaps indicating a dependence on daylength as a cue). Environmental gradients may produce some of the same effects as elevational gradients; for example, Stanton et al. (1997) suggested that differences in flowering time of Ranunculus adoneus along a snowmelt gradient could reduce opportunities for pollen transfer, and presented evidence for some genetic differentiation among early, middle, and latemelting cohorts. The longer growing season in early-melting sites enhanced vegetative growth at all life-history states and increased fecundity of seedlings (Stanton and Galen 1997).

3.2

Community-Level Patterns and Temporal Patterns

Relatively few community-level surveys of plant phenology have been made of alpine plants. Douglas and Bliss (1977) conducted weekly samples in 2x2m plots in the North Cascade Range of Washington and recorded times of vegetative growth, flowering, fruiting, seed dispersal, and dormancy of 32 species in weekly surveys. Most species flowered within 14-24 days following initiation of growth, remained in bloom for 8-20 days, and then had a 10-24 day fruiting stage prior to seed dispersal. This is also reported in various communities of the low alpine region in southern Norway (Wielgolaski and Kärenlampi 1975). Ratcliffe and Turkington (1989) used similar methods looked at 45 vascular plant species in southern British Columbia. Most species flowered between 15 and 40 days after snowmelt. Bauer (1983) reported data on the seasonal flowering phenology of 24 species of tundra plants visited by bumblebees, from weekly surveys. Molau (1993) conducted a study of the relationships between flowering phenology and life history strategies of plants in high-altitude Sweden. Some of the immense variation among tundra species in reproductive traits

202

Phenology: An Integrative Environmental Science

(e.g., seed:ovule and fruit:flower ratios) in this habitat is correlated with flowering phenology (which in turn is linked to snowmelt patterns). Earlyflowering species show high outbreeding rates and low seed:ovule ratios, while late-flowering species showed the reverse. Apomixis and vivipary were restricted to late-flowering species, and ploidy levels increased from early- to late-flowering times. Relatively little is known about phenology at high altitudes in the southern hemisphere, or in tropical areas. Inouye and Pyke (1988) did report phenological data for alpine Australia from a one-year study. Ralph (1978) found that plants of Azorella compacta (Apiaceae) flowered year-round in populations at 3960 and 4500 m. Phenological patterns of the high Andean Cordillera of central Chile were studied by Arroyo et al. (1981), who looked at 97 alpine species. They characterized these species as falling into nine different patterns of phenological behavior. The most common category was perennials that overwintered in a dormant state. At all sites they examined, there was a single prominent maximum in flowering activity, which was earlier on north-facing than south-facing slopes. An interesting observation about flowering longevity was that duration of flowering tended to be longer at higher altitudes (individual flowers lasted longer, and the total flowering period was also longer). (Gómez 1993 also saw longer flowering periods at higher altitudes.) Over their two-year study, they found that phenological patterns were “remarkably constant from one year to the next”; more than 90% of species had identical phenology in both years. Melampy (1987) studied flowering phenology of Befaria resinosa, an ericaceous shrub of the eastern Andes of Colombia. This species seems to have two peaks of flowering each year, although one is more significant, and there was some suggestion from the data that the two peaks of rainfall might have been responsible for initiation of flowering. The weather at this 2200 m site is quite dry for much of the year, and 22% and 31% of the annual precipitation occurs during the two peaks. There are often sympatric congeneric wildflower species in high altitude habitats, and it’s not surprising that in such cases they might flower at different times as a way to avoid competition for pollinators. Adams (1983) studied five sympatric species of Pedicularis (Scrophulariaceae) on Mount Rainier in the Washington Cascade Mountains, all of which are pollinated by bumblebees. Two bloom early in the growing season, one in mid-season, and two in lateseason. At RMBL, pairs of Mertensia and Delphinium species include species that flower early and at mid-season (Inouye, unpublished). Gómez (1993) investigated the effects of flowering synchrony on reproductive success of 80 Hormathophylla spinosa (Cruciferae) plants in the Sierra Nevada mountains of Spain. He measured flowering synchrony as the number of days that the flowering of an individual overlaps with the

Chapter 3.5: High Altitude Climates

203

flowering of every other plant in the sample, and found that synchrony ranged from 0.25 to 1.00, with more than 70% of plants having synchrony levels greater than 0.75. Plants with a lower degree of flowering synchrony were visited by more pollinators and also eaten by fewer herbivores, but the phenological traits did not affect female fertility so Gómez concluded that flowering synchrony in this species is not regulated by selective pressures from pollinators or herbivores. Although it may be more common than the literature indicates, mast flowering does occur in at least some alpine or montane ecosystems. Frasera speciosa (Gentianaceae) has a flowering pattern that has been described as ‘sporadic seasonal synchrony’, in which significant flowering events of this long-lived monocarp occurs only once every several years (Beattie et al. 1973; Taylor and Inouye 1985). A similar flowering pattern, but with even longer intervals between significant flowering years, is seen in the lily Veratrum tenuipetalum (Inouye, unpublished observations). Mast flowering has also been reported from a few alpine species in New Zealand (Kelly et al. 2000; Rees et al. 2002)

4.

FLOWERING PHENOLOGY IN THE COLORADO ROCKY MOUNTAINS

Inouye has conducted since 1973 a study of the timing and abundance of flowering by Rocky Mountain wildflowers, in permanent 2x2m plots near the RMBL. This site is located in the West Elk Mountains of Colorado, at an altitude of about 2,900m (38º57.5’N, 106º59.3’W). Every other day for most or all of the growing season all flowers in the plots are counted. Phenology of flowering and many other events at this site is dependent on the amount of snow that fell during the previous winter and when that snow melts, as the growing season does not begin until the snow is gone. Since 1975 the annual snowfall (measured as daily snowfall or even more frequently during storms; unpublished data courtesy of Billy Barr) has ranged from 474-1641cm (mean = 1105). The date on which the permanent snowpack has begun has ranged from 15 October to 24 November (mean = 4 November), and the snow has disappeared from the measurement site as early as 25 April and as late as 19 June (mean = 24 May). Of course there is much variation in these dates across the landscape, with differences in microclimate, slope aspect, and snow deposition through wind activity resulting in some sites that hold or lose snow earlier or later. One question that arises is whether the range of dates described here is typical of a longer-term period. Although we do not have snowfall data from RMBL earlier than 1975, a proxy for snowfall, the peak runoff in the

204

Phenology: An Integrative Environmental Science

East River (which runs through the Lab) is measured at Almont, Colorado, and has been recorded continuously since 1935. The correlation between measurements of winter snowfall at RMBL and the peak runoff measured at Almont is strong (rr2 = .757, p = .0001). If peak runoff is then used to estimate historical snowfall at RMBL, we find that the period 1975 – 2002 contains both the maximum and minimum snowfall records for the 68 years of data. This could be an indication of the increased variation in precipitation that has been predicted by some models of global climate change. For all species that we have examined so far, from the earliest flowering (Claytonia lanceolata) to the latest (Artemisia ( tridentata), the timing of flowering is strongly linked to the amount of snow that falls during the previous winter, and hence the timing of snowmelt. For some species the relationship is linear (e.g., Delphinium nuttallianum, Figure 1a), while for others it is curvilinear (e.g., Mertensia ciliata, Figure 1b). The curvilinear relationship seen for some species may indicate an interaction between temperature and number of days since snowmelt; in years with early snowmelt, it takes more days (i.e., more growing degree days) for some heat sum to be achieved before flowering begins. Such curvilinear relationships may characterize later-flowering species (e.g., Inouye et al. 2002), as we have not observed them yet in early-flowering species. This strong reliance on snowmelt as an environmental factor determining flowering time, seems to characterize all of the herbaceous species at this site, and possibly the relatively few woody ones as well. There do not seem to be species at this altitude that rely on day length for timing of flowering. Another factor that is important in the abundance of flowers and is linked to phenology is frost (Inouye 2000). In some years frost can kill almost all of the buds of some species of wildflowers at RMBL, such as Helianthella quinquenervis, the aspen sunflower. In eight of the past 28 years there has been significant frost damage (Inouye, unpublished). The critical factor for this interaction between snow, frost, and flower abundance appears to be phenology. If there is sufficient snow during the winter, snowmelt, and hence the beginning of the growing season, will be delayed compared to those years with light snowpack. Hard frosts (e.g., temperatures down to -7° C) can occur as late as the third week in June. If plants begin growth too early in the season, they can have buds at a sensitive stage of development by the third week of June, and are then susceptible to frost damage. If there is sufficient snow to delay snowmelt, and hence the growing season, then development of frost-sensitive buds will be delayed beyond the time when frost is likely to occur. There is some evidence that frost risk at lower altitudes may be declining as a consequence of global climate change (Moonen et al. 2002).

Chapter 3.5: High Altitude Climates

205

Given that there is so much evidence supporting the dependence of highaltitude phenology on snowpack, it may be possible to use historical data on

Figure 3.5-1. The relationship between the first date of bare ground and the date of first flower for a) Delphinium nuttallianum and b) Mertensia ciliata.

206

Phenology: An Integrative Environmental Science

snowpacks in other areas to draw conclusions about historical patterns of phenology. For example, spring pulse dates of streamflow exhibit trends toward earlier spring timing, matching patterns of flowering by lilacs and honeysuckles in phenology studies (Cayan et al. 2001).

5.

PHENOLOGY AND ANIMAL ECOLOGY

Although most of this review has focused on plant phenology, there are indications from the few available studies that animal phenology at high altitudes is also influenced by the same kinds of environmental cues that affect plants. Morton (1994) assessed timing of reproduction for both wild onion (Allium ( validum) and White-crowned Sparrows (Zonotrichia leucophyrus oriantha) at the same subalpine meadow in the California Sierra Nevada. Data from 21 years indicated that both flowering date and clutch initiation date were highly correlated with snow conditions (occurring later as snowpack increased). Langvatn et al. (1996) found that plant phenology had important consequences for diet quality, and hence reproduction, for red deer. They found a negative correlation between various fitness measures for female red deer and growing degree-days during the summer. Their interpretation was that when herbage growth is retarded (via cooler weather) the digestibility of plants declines more slowly than in warm summers, resulting in a longer period with good grazing. Although this study was conducted at high latitude rather than high altitude, the same principle may apply at higher altitudes too. Merrill and Boyce (1991) make a similar story for high-elevation summer range for elk in the Yellowstone ecosystem, describing a link between heavy winter snowfall, delayed phenology the next summer, and consequent high-quality forage through late summer.

6.

CLIMATE CHANGE AND PHENOLOGY IN THE COLORADO ROCKY MOUNTAINS

Inouye et al. (2000) reported that there was no sign of an effect of climate change on the phenology of wildflowers in the Rocky Mountains. This finding contrasts strongly with results from lower altitudes, where many studies have reported that phenological events are happening earlier than they used to (e.g., Bradley et al. 1999; Brown et al. 1999; Roy and Sparks, 2000; Peñuelas and Filella 2001; Sagarin and Micheli 2001; Fitter and Fitter 2002; Peñuelas et al. 2002; Walther et al. 2002). The paper by Inouye et al.

Chapter 3.5: High Altitude Climates

207

pointed out that this difference between low and high altitudes in phenological responses to climate change could cause problems for altitudinal migrants, and reported data suggesting that robins (Turdus migratorius) were showing effects of this change in synchrony. The reason that flowering phenology was not changing at high altitudes is probably related to the trend observed for increased winter precipitation (see Figure 1 in Inouye et al. 2000, and Figure 3.4-4 in this book). Thus, even though air temperatures appeared to be warming, the net result was that there was no trend for earlier snowmelt dates (and hence phenology). Some models of global climate change predict increased precipitation. In contrast to flowering phenology, there is evidence that some animals are responding to warming temperatures with changes in phenology. Inouye et al. (2000) reported that marmots were emerging from hibernation significantly earlier than they did a few decades earlier, and therefore emerging when there was much more snow left on the ground than previously. Another change is seen in the emergence dates of a butterfly species that overwinters as adults. Emergence date has become significantly later during the period 1975 – 2002 (Figure 2), for reasons that are not clear. Emergence dates of ground squirrels and chipmunks from winter hibernation are also changing, but in the opposite direction from that of marmots (Billy Barr, unpublished). Apparently they are responding to different cues than the marmots, or they are responding differently to the same cues as marmots.

Figure 3.5-2. The change in emergence date for adults of the butterfly Nymphalis milberti at the Rocky Mountain Biological Laboratory. Unpublished data from Billy Barr.

208

Phenology: An Integrative Environmental Science

If the emergence phenology of pollinators and herbivores is changing relative to their host plants and food plants, it seems likely that coevolved relationships may change significantly in the future. This kind of change in synchrony has also been reported from lower altitudes (Visser and Holleman 2001). High-altitude aquatic habitats may also be showing signs of changing phenology. Caine (2002) reported that spring ice thickness on an alpine lake is declining at a rate of about 2 cm/yr, and the duration of ice cover also appears to be declining. Presumably a variety of aspects of plant and animal phenology in such alpine lakes is being influenced by the change in ice cover (which may be a consequence of increasing precipitation). The changes that are predicted for distribution of alpine plants in the future as a consequence of climate change (e.g., Guisan and Theurillat 2000; Theurillat and Guisan 2001) will be in part a function of changes in phenology. The ultimate result is likely to be an overall trend for reduced availability of suitable habitat, and shifts in the distribution of species richness (either shifting upward or spreading out of patches) (Guisan and Theurillat 2000). There is some evidence that plants in the Swiss Alps may already be showing signs of such a shift upward in distribution (Hofer 1992; Grabherr et al. 1995; and Chapter 3.4 in this book).

7.

CONCLUSIONS

The disappearance of snow cover appears to be the primary factor influencing phenology at high altitudes in the temperate zone. Not enough is known yet about other high-altitude areas without significant snow cover to confirm what is controlling their phenologies. One consequence of the importance of snow in controlling phenology is that flowering, and other phenological events involving both plants and animals, can be highly variable because of variation across years in snowpack depth and across space because of aspect and microsite differences in snow accumulation and melting. A consequence of this variation may be that no single set of phenological and physiological characteristics is optimally adapted to all of this variability, which would then encourage the evolution and maintenance of a diversity of adaptive strategies in high altitude communities. Although the timing of events exhibits significant variation among years, their relative timing is much less variable. The sequence of flowering species is typically consistent across years, with the same species flowering each year in early, middle or late season. Variation across space is also consistent, as sites that tend to accumulate snow in one year are likely to do so in other years.

Chapter 3.5: High Altitude Climates

209

Lest flowering phenology at high altitudes be considered an esoteric topic, consider the economic importance of strawberry production. Strawberries used to be highly seasonal, but it is now possible to produce multiple crops across most temperate climates. This change is the result of developing new multiple-cropping cultivars with cyclic flowering. The genes for this change for all modern day-neutral cultivars come from a single clone of Fragaria virginiana ssp. glauca from the Wasatch Mountains of Utah (Sakin et al. 1997).

8.

AVENUES FOR FUTURE INVESTIGATION

Experimental studies of phenology are rare, even though snowpack is relatively easily manipulated (e.g., Galen and Stanton 1993; Dunne et al. 2003), but they offer the potential to elucidate the significance of variation in phenology more quickly than observational studies. The International Tundra Experiment has investigated responses of tundra plants to simulated climate warming, but most of its sites are at high latitudes (Henry and Molau 1997). Suzuki and Kudo (2000) did carry out an experiment in Japanese alpine tundra using open top chambers and five species of shrubs. Although there was some evidence of earlier leafing and flowering, this was not a consistent response. Additional experimental studies are likely to provide interesting and novel results. It would be valuable to have additional longterm studies of phenology at high altitudes, not only of flowering and other phenological events for plants, but also for phenology of animals. Not much is known about the phenology of hibernating and migrating species, or insects at high altitudes. There are been some interesting reports of differences in phenological characteristics within species across altitudinal gradients, for example, that flowering may last longer at higher altitudes. Additional studies across such gradients would be useful to confirm this and other aspects of phenology. Studies of the consequences of variation in phenology are uncommon. What are the consequences of flowering early or late, or at the tail vs. the peak of flowering time? What are the consequences of variation in time of emergence from hibernation or in arrival or departure of migrating species? Very little is known about what controls the phenology of high-altitude tropical ecosystems, where snow is not a significant controlling factor. Given that snow may play little or no role in some of those areas, what is the influence of other environmental variables? We suggest that future studies of high-altitude phenology will be particularly interesting in the context of climate change, and hope that this review will help to stimulate additional work on this topic.

210

Phenology: An Integrative Environmental Science

REFERENCES CITED Adams, V. D., Temporal patterning of blooming phenology in Pedicularis on Mount Rainier, Can. J. Botany, 61, 786-6791, 1983. Akhalkatsi, M., and J. Wagner, Reproductive phenology and seed development of Gentianella caucasea in different habitats in the central Caucasus, Flora, 191, 161-168, 1996. Arroyo, M. T. K., J. J. Armesto, and C. Villagran, Plant phenological patterns in the high Andean Cordillera of central Chile, J. Ecology, 69, 205-223, 1981. Bauer, P. J., Bumblebee pollination relationships on the Beartooth Plateau tundra of southern Montana, Amer. J. Botany, 70, 134-144, 1983. Beattie, A., D. Breedlove, and P. Ehrlich, The ecology of the pollinators and predators of Frasera speciosa, Ecology, 54, 81-91, 1973. Billings, W. D., and L. C. Bliss, An alpine snowbank environment and its effects on vegetation, plant development, and productivity, Ecology, 40, 388-397, 1959. Billings, W. D., and H. A. Mooney, The ecology of arctic and alpine plants, Biological Reviews of the Cambridge Philosophical Society, 43, 481-529, 1968. Bliss, L. C., A comparison of plant development in microenvironments of arctic and alpine tundras, Ecol. Monographs, 26, 303-337, 1956. Bliss, L. C., Arctic and alpine plant life cycles, Annual Review of Ecology and Systematics, 2, 405-438, 1971. Bradley, N. L., A. C. Leopold, J. Ross, and W. Huffaker, Phenological changes reflect climate change in Wisconsin, Proc. Nat. Acad. Science (USA), 96, 9701-9704, 1999. Brown, J. L., S. -H. Li, and N. Bhagabati, Long-term trend toward earlier breeding in an American bird: A response to global warming?, Proc. Nat. Acad. Science(USA), 96, 55655569, 1999. Caine, N., Declining ice thickness on an alpine lake is generated by increased winter precipitation, Climatic Change, 54, 463-470, 2002. Canaday, B. B., and R. W. Fonda, The influence of subalpine snowbanks on vegetation pattern, production, and phenology, Bull. Torrey Botanical Club, 101, 340-350, 1974. Cayan, D. R., S. A. Kammerdiener, M. D. Dettinger, J. M. Caprio, and D. H. Peterson, Changes in the onset of spring in the western United States, Bull. Amer. Meteorol. Soc., 82, 399-415, 2001. Douglas, G. W., and L. C. Bliss, Alpine and high subalpine plant communities of the North Cascades Range, Washington and British Columbia, Ecol. Monographs, 47, 113-150, 1977. Dunne, J. A., J. Harte, and K. J. Taylor, Subalpine meadow flowering phenology responses to climate change: integrating experimental and gradient methods, Ecol. Monographs, 73, 69-86, 2003. Fitter, A. H., and R. S. R. Fitter, Rapid changes in flowering time in British plants, Science, 296, 1689-1691, 2002. Forbis, T. A., and P. K. Diggle, Subnivean embryo development in the alpine herb Caltha leptosepala (Ranunculaceae), Can. J. Botany (Revue Canadienne de Botanique), 79, 635642, 2001. Galen, C., and M. L. Stanton, Consequences of emergence phenology for reproductive success in Ranunculus adoneus (Ranunculaceae), Amer. J. Botany, 78, 978-988, 1991. Galen, C., and M. L. Stanton, Short-term responses of alpine buttercups to experimental manipulations of growing season length, Ecology, 74, 1052-1058, 1993.

Chapter 3.5: High Altitude Climates

211

Giorgi, F., W. Hurrell, M. Marinucci, and M. Beniston, Elevation dependency of the surface climate change signal: a model study, J. Climate, 10, 288-296, 1997. Gómez, J. M., Phenotypic selection on flowering synchrony in a high mountain plant, Hormathophylla spinosa (Cruciferae), J. Ecology, 81, 605-613, 1993. Grabherr, G., M. Gottfried, A. Gruber, and H. Pauli, Patterns and current changes in alpine plant diversity, in Arctic and alpine biodiversity: patterns, causes and ecosystem consequences, Ecological Studies, edited by F. S. Chapin, III and C. Koerner, pp. 167-181, Springer Verlag, Berlin, 1995. Guisan, A., and J. -P. Theurillat, Equilibrium modeling of alpine plant distribution: how far can we go?, Phytocoenologia, 30, 353-384, 2000. Heide, O., Daylength and thermal time responses of budburst during dormancy release in some northern deciduous trees, Physiologia Plantarum, 88, 531-540, 1993. Henry, G. H. R., and U. Molau, Tundra plants and climate change: the International Tundra Experiment (ITEX), Global Change Biology, 3, Suppl. 1, 1-9, 1997. Hofer, H. R., Veränderungen in der Vegetation von 14 Gipfeln des Berninagebietes zwischen 1905 und 1985, Berichte des Geobotanischen Instituts der ETH Zürich, Stiftung Rüubel, 58, 39-54, 1992. Hoffmann, A. J., and M. J. Walker, Growth habits and phenology of drought-deciduous species in an altitudinal gradient, Can. J. Botany, 58, 1789-1796, 1980. Holway, J. G., and R. T. Ward, Snow and meltwater effects in an area of Colorado alpine, Amer. Midland Naturalist, 69, 189-197, 1963. Holway, J. G., and R. T. Ward, Phenology of alpine plants in northern Colorado, Ecology, 46, 73-83, 1965. Inouye, D. W., The ecological and evolutionary significance of frost in the context of climate change, Ecol. Letters, 3, 457-463, 2000. Inouye, D. W., B. Barr, K. B. Armitage, and B. D. Inouye, Climate change is affecting altitudinal migrants and hibernating species, Proc. Nat. Acad. Science(USA), 97, 16301633, 2000. Inouye, D. W., M. Morales, and G. Dodge, Variation in timing and abundance of flowering by Delphinium barbeyi Huth (Ranunculaceae): the roles of snowpack, frost, and La Niña, in the context of climate change, Oecologia, 139, 543-550, 2002. Inouye, D. W., and G. H. Pyke, Pollination biology in the Snowy Mts. of Australia, with comparisons with montane Colorado, U. S. A., Australian J. Ecology, 13, 191-210, 1988. Jackson, L. E., and L. C. Bliss, Distribution of ephemeral herbaceous plants near treeline in the Sierra Nevada, California, U.S.A., Arctic and Alpine Research, 14, 33-42, 1982. Jackson, L. E., and L. C. Bliss, Phenology and water relations of three plant life-forms in a dry treeline meadow, Ecology, 65, 1302-1314, 1984. Kelly, D., A. L. Harrison, W. G. Lee, I. J. Payton, P. R. Wilson, and E. M. Schauber, Predator satiation and extreme mast seeding in 11 species of Chionochloa (Poaceae), Oikos, 90, 477-488, 2000. Kittel, T. G. F., P. E. Thornton, J. A. Royle, and T. N. Chase, Climates of the Rocky Mountains: historical and future patterns, in Rocky Mountain Futures: An Ecological Perspective, edited by J. Baron, D. Fagre and R. Hauer, pp. 59-82, Island Press, Covelo, CA, 2002. Knight, D. H., B. S. Rogers, and C. R. Kyte, Understory plant growth in relation to snow duration in Wyoming subalpine forest, Bull. Torrey Botanical Club, 104, 314-319, 1977. Kudo, G., Effects of snow-free period on the phenology of alpine plants inhabiting snow patches, Arctic and Alpine Research, 23, 436-443, 1991.

212

Phenology: An Integrative Environmental Science

Kudo, G., Performance and phenology of alpine herbs along a snow-melting gradient, Ecological Research, 7, 297-304, 1992. Langvatn, R., S. D. Albon, T. Burkey, and T. H. Clutton Brock, Climate, plant phenology and variation in age of first reproduction in a temperate herbivore, J. Animal Ecology, 65, 653670, 1996. Lynov, Y. S., Phenological inversions in alpine terrain, Western Tien Shan, Ékologiya, 4, 2933, 1984. Mark, A. F., Floral initiation and development in New Zealand alpine plants, New Zealand J. Botany, 8, 67-75, 1970. Melampy, M. N., Flowering phenology, pollen flow and fruit production in the Andean shrub Befaria resinosa, Oecologia, 73, 293-300, 1987. Meloche, C. G., and P. K. Diggle, Preformation, architectural complexity, and developmental flexibility in Acomastylis rossii (Rosaceae), Amer. J. Botany, 88, 980-991, 2001. Merrill, E. H., and M. S. Boyce, Summer range and elk population dynamics in Yellowstone National Park, in The Greater Yellowstone Ecosystem: Redefining America's Wildlife Heritage, edited by R. B. Keiter and M. S. Boyce, pp. 263-273, Yale University Press, New Haven, 1991. Mitton, J. B., K. B. Sturgeon, and M. L. Davis, Genetic differentiation in ponderosa pine along a steep elevational transect, Silvae Genetica, 29, 100-103, 1980. Molau, U., Relationships between flowering phenology and life history strategies in tundra plants, Arctic and Alpine Research, 25, 391-402, 1993. Moonen, A. C., L. Ercoli, M. Mariotti, and A. Masoni, Climate change in Italy indicated by agrometeorological indices over 122 years, Agricul. Forest Meteorol., 111, 13-27, 2002. Mooney, H. A., K. S. Williams, D. E. Lincoln, and P. R. Ehrlich, Temporal and spatial variability in the interaction between the checkerspot butterfly, Euphydryas calcedona and its principal food source, the Californian shrub, Diplacus aurantiacus, Oecologia, 50, 195198, 1981. Morton, M. L., Comparison of reproductive timing to snow conditions in wild onions and White-Crowned Sparrows at high altitude, Great Basin Naturalist, 54, 371-375, 1994. Myking, T., and O. M. Heide, Dormancy release and chilling requirement of buds of latitudinal ecotypes of Betula pendula and B. pubescens, Tree Physiology, 15, 697-704, 1995. Peñuelas, J., and I. Filella, Responses to a warming world, Science, 294, 793-795, 2001. Peñuelas, J., I. Filella, and P. Comas, Changed plant and animal life cycles from 1952 to 2000 in the Mediterranean region, Global Change Biology, 8, 531-544, 2002. Ralph, C. P., Observations on Azorella compacta (Umbelliferae), a tropical Andean cushion plant, Biotropica, 10, 62-67, 1978. Ratcliffe, M. J., and R. Turkington, Comparative phenology of some alpine vascular plant species on Lakeview Mountain, Southern British Columbia, Canadian Field Naturalist, 103, 348-352, 1989. Rees, M., D. Kelly, and O. N. Bjornstad, Snow tussocks, chaos, and the evolution of mast seeding, Amer. Naturalist, 160, 44-59, 2002. Resvoll, T. R., Om planter som passer til kort og kold sommer (in Norwegian), Arciv Mathematik Naturvidenskab, 35, 1-224, 1917. Reynolds, D. N., Alpine annual plants: phenology, germination, photosynthesis, and growth in three Rocky Mountain species, Ecology, 65, 759-766, 1984. Richardson, S. G., and F. B. Salisbury, Plant responses to the light penetrating snow, Ecology, 58, 1152-1158, 1977.

Chapter 3.5: High Altitude Climates

213

Roy, D. B., and T. H. Sparks, Phenology of British butterflies and climate change, Global Change Biology, 6, 407-416, 2000. Sagarin, R., and F. Micheli, Climate change in nontraditional data sets, Science, 294, 811, 2001. Sakin, M., J. F. Hancock, and J. J. Luby, Identifying new sources of genes that determine cyclic flowering in Rocky Mountain populations of Fragaria virginiana ssp glauca Staudt, J. Amer. Soc. Hort. Science, 122, 205-210, 1997. Schuster, W. S., D. L. Alles, and J. B. Mitton, Gene flow in limber pine: evidence from pollination phenology and genetic differentiation along an elevational transect, Amer. J. Botany, 76, 1395-1403, 1989. Stanton, M. L., and C. Galen, Life on the edge: Adaptation versus environmentally mediated gene flow in the snow buttercup, Ranunculus adoneus, Amer. Naturalist, 150, 143-178, 1997. Stanton, M. L., C. Galen, and J. Shore, Population structure along a steep environmental gradient: Consequences of flowering time and habitat variation in the snow buttercup, Ranunculus adoneus, Evolution, 51, 79-94, 1997. Stenström, M., and U. Molau, Reproductive ecology of Saxifraga oppositifolia: phenology, mating system, and reproductive success, Arctic and Alpine Research, 24, 337-343, 1992. Suzuki, S., and G. Kudo, Responses of alpine shrubs to simulated environmental change during three years in the mid-latitude mountain, northern Japan, Ecography, 25, 553-564, 2000. Taylor, O. R., Jr., and D. W. Inouye, Synchrony and periodicity of flowering in Frasera speciosa (Gentianaceae), Ecology, 66, 521-527, 1985. Theurillat, J. P., and A. Guisan, Potential impact of climate change on vegetation in the European Alps: A review, Climatic Change, 50, 77-109, 2001. Theurillat, J. P., and A. Schlüssel, Phenology and distribution strategy of key plant species within the subalpine-alpine ecocline in the Valaisan Alps (Switzerland), Phytocoenologia, 30, 439-456, 2000. Totland, Ø. Pollination in alpine Norway: flowering phenology, insect visitors, and visitation rates in two plant communities, Can. J. Botany (Revue Canadienne de Botanique), 71, 1072-1079, 1993. Totland, Ø. Effects of flowering time and temperature on growth and reproduction in Leontodon autumnalis var. taraxaci a late-flowering alpine plant, Arctic and Alpine Research, 29, 285-290, 1997. Visser, M. E., and L. J. M. Holleman, Warmer springs disrupt the synchrony of oak and winter moth phenology, Proceedings of the Royal Society B, 268, 1-6, 2001. Wagner, J., and E. Mitterhofer, Phenology, seed development, and reproductive success of an alpine population of Gentianella germanica in climatically varying years, Botanica Acta, 111, 159-166, 1998. Wagner, J., and B. Reichegger, Phenology and seed development of the alpine sedges Carex curvula and Carex firma in response to contrasting topoclimates, Arctic and Alpine Research, 29, 291-299, 1997. Walker, M. D., R. C. Ingersoll, and P. J. Webber, Effects of interannual climate variation on phenology and growth of two alpine forbs, Ecology, 76, 1067-1083, 1995. Walther, G. -R., E. Post, P. Convey, A. Menzel, C. Pamesan, T. J. C. Beebee, J. -M. Fromentin, O. Hoegh-Guldberg, and F. Bairlein, Ecological responses to recent climate change, Nature, 416, 389-395, 2002. Wardle, P., Alpine timberlines, in Arctic and alpine environments, edited by J. D. Ives and R. G. Barry, pp. 371-402, Methuen, London, 1974.

214

Phenology: An Integrative Environmental Science

Wielgolaski, F. E., and L. Kärenlampi, Plant phenology of Fennoscandinan tundra areas, in Fennoscandian Tundra Ecosystems, Part 1., vol. 16, Ecological Studies: Analysis and Synthesis, edited by F. E. Wielgolaski, pp. 94-102, Springer-Verlag, Berlin, 1975. Williams, M. C., and E. H. Cronin, Dormancy, longevity, and germination of seeds of three larkspurs and western false hellebore, Weeds, 8, 452-461, 1968.

PART 4

PHENOLOGICAL MODELS AND TECHNIQUES

Chapter 4.1 PLANT DEVELOPMENT MODELS Isabelle Chuine1, Koen Kramer2, and Heikki Hänninen3 1

CEFE-CNRS, Montpellier, France; 2Alterra, Department of Ecology and Environment, Wageningen University, Wageningen, The Netherlands; 3Department of Ecology and Systematics, University off Helsinki, Helsinki, Finland

Key words:

1.

Statistical and mechanistic models, Budburst, Flowering, Frost hardiness, Species range

AN OVERVIEW OF PHENOLOGY MODELING DURING THE LAST THREE CENTURIES

Phenology modeling has a long history starting in 1735 with a publication by Reaumur (1735). Reaumur suggested that differences between years and locations in the date of phenological events could be explained by differences in daily temperatures from an arbitrary date to the date of the phenological event considered. This is still the most important assumption in plant phenology modeling. The main advances in phenology modeling took place in the late 20th century (Table 1) for two main reasons: (i) the revolution in computer science, and (ii) concerns about global climate change. Global warming is expected to have major impacts on plant functions and fitness, as increasing temperatures will change the timing of phenological events. Most plant phenology models predict budburst (leaf unfolding), flowering, and fruit maturation (Table 1), but no model can predict leaf coloration so far. Leaf unfolding and flowering are the most widely observed phenophases, because the timing of these events can be observed accurately. This is much less the case for fruit maturation and leaf coloration. Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 217-235 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

218

Phenology: An Integrative Environmental Science

Additionally, leaf unfolding is important for primary productivity models (see section 4). Most phenology models were developed for tree species, rather than nonwoody species (Table 1). Many more phenological modeling studies focus Table 4.1-1. Classification of phenology modeling studies according to the phenophase, the species studied, and the type of observations used Phenophase Budburst (Robertson 1968; Cannell and Smith 1983; Kobayashi and Fuchigami 1883a, 1983b; Cannell and Smith 1986; Nizinski and Saugier 1988; Murray, y et al. 1989; Hänninen 1991; Hunter and Lechowicz 1992; Kramer 1994b; Hänninen 1995; Kramer 1995b; Häkkinen 1999) Flowering (Boyer 1973; Kupias and Mäkinen 1980; Ellis, et al. 1988; Roberts et al. 1988; Frenguelli et al. 1989; Andersen 1991; Frenguelli et al. 1991; Sinclair et al. 1991; Frenguelli et al. 1992; Marletto et al. 1992; Pipper et al. 1996; Chuine et al. 1998; Frenguelli and Bricchi 1998; Oliveira 1998; Chuine et al. 1999; Osborne et al. 2000) Fruit maturation (Pipper et al. 1996; Lescourret et al. 1999) Species Woody (Boyer 1973; Landsberg 1974; Cannell and Smith 1983; Kobayashi and Fuchigami 1983a, 1983b; Anderson; et al. 1986; Cannell and Smith 1986; Winter 1986; Nizinski and Saugier 1988; Murray et al. 1989; Hänninen 1990a, 1990b; Andersen 1991; Hänninen 1991; Kikuzawa 1991; Hunter and Lechowicz 1992; Hänninen et al. 1993; Kramer 1994a, 1994b; Hänninen 1995; Kramer 1995b; Leinonen et al. 1995; Kikuzawa 1996; Chuine et al. 1998, 1999; Häkkinen 1999; Lescourret et al. 1999; Osborne et al. 2000) Non woody (Robertson 1968; Ellis et al. 1988; Roberts et al. 1988; Sinclair et al. 1991; Pipper et al. 1996) Observations Phenology networks (Boyer 1973; Nizinski and Saugier 1988; Frenguelli et al. 1989; Frenguelli et al. 1991; Frenguelli et al. 1992; Hunter and Lechowicz 1992; Marletto et al. 1992; Kramer 1994a, 1994b, Häkkinen et al. 1995; Linkosalo et al. 1996; Chuine et al. 1998; Häkkinen et al. 1998; Chuine et al. 1999; Häkkinen 1999; Linkosalo 1999; Menzel and Fabian 1999; Linkosalo et al. 2000; Osborne et al. 2000) Experiments (Hänninen 1987; Ellis et al. 1988; Roberts et al. 1988; Cannell 1989; Murray et al. 1989; Hänninen 1990b; Sinclair et al. 1991; Hänninen et al. 1993; Hänninen 1995)

Chapter 4.1: Plant Development Models

219

on temperate biota than on boreal, tropical or sub-tropical biota (Kupias and Mäkinen 1980; Phillipp et al. 1990; Reich 1994; Thorhallsdottir 1998). Phenological observations used to develop phenology models have two main origins: historical observations in wild populations or phenological gardens and experimental results (Table 1). Both data types are useful for phenology modeling, but imply different methodologies as we discuss in section 3.

2.

AN OVERVIEW OF PLANT PHENOLOGY MODELS

Reaumur (1735) first introduced the concept of degree-day sum, i.e. daily average temperatures accumulated between an arbitrary date of onset and the date of an observed phenological event. Reaumur showed that the late harvesting date of crop and grapes in 1735 had a lower sum of degree days in April, May and June than in 1734, which had early harvests. He realized that a plant develops quicker at a higher temperature, thus shortening the interval between sowing and crop harvest, or flowering and vine harvest. He proposed that plant development is proportional to the sum of temperature over time rather than to temperature during the phenological event itself. In many studies since Reaumur, accumulated temperature is recognized as the main factor influencing year-to-year variation in phenology. The evidence for the role of photoperiod in tree phenology is conflicting, depending on species and location (Heide 1993; Kramer 1994b; Falusi and Calamassi 1996). However, photoperiod without interaction with temperature cannot explain the annual variability of phenology at a given location because photoperiod is the same each year. Three main types of phenology models exist: theoretical, statistical, and mechanistic models. Theoretical models are based on the cost/benefit tradeoff of producing leaves to optimize resource acquisition (Kikuzawa 1991, 1995a, 1995b, 1996; Kikuzawa and Kudo 1995) and are designed to understand the evolution of leaf lifespan strategies in trees, rather than the annual variation in plant phenology. Statistical phenology models relate the timing of phenological events to climatic factors. Model parameters are estimated from empirical data using various statistical fitting methods. While these models may not consider specific biological processes, some are more mechanistic than others. Statistical models are varied. Some are simple correlations with average temperature in different periods of the year (Boyer 1973; Spieksma et al. 1995; Emberlin et al. 1997), some are more complex (Schwartz and Marotz 1986, 1988; Schwartz 1997, 1998, 1999). In particular the Spring Index

220

Phenology: An Integrative Environmental Science

Model (Schwartz and Marotz 1986, 1988; Schwartz 1997) has been successfully used to predict the start of the growing season in the United States. This is a multiple regression model of the type:

( y ) −1 = C +

∑A X k

k

( y ) −1

(1)

n

where y is the date of the phenological event, C and Ak are constants and Xk are predictor variables: e.g., degree-day sums for two threshold temperatures (-0.6°C or 5°C), mean temperature, intensity, and number of synoptic weather events (in particular warm air advection, see Chapter 4.8). Mechanistic models formally describe known or assumed cause-effect relationships between biological processes and some driving factors in the plant’s environment. New relationships should be allowed in a mechanistic model only if information on its impacts on the process is available. It is important to note that parameters of mechanistic models have physical dimensions that can, in principle (see section 3), be measured directly instead of being estimated by fitting. However, this is rarely possible in phenology models. As most of the models described in the literature are of this type, the following paragraphs provide a detailed overview of their hypotheses and formulation. Experimental evidence shows that dormancy (rest) needs to be broken before plants enter the phase of quiescence during which the rate of ontogenetic development increases with increasing temperature (Lamb 1948; Landsberg 1974; Campbell and Sugano 1975; Cannell and Smith 1983). Temperature is also involved in breaking dormancy, although in a different manner, i.e. lower temperatures are required. This is why we refer to a “chilling requirement” (Sarvas 1974; Hänninen 1990b). The rate of development during both phases has been related to temperature in many ways, usually called chilling and forcing units (of development). Reaumur’s approach later referred to as the Thermal Time model (Robertson 1968; Cannell and Smith 1983) or Spring Warming model (Hunter and Lechowicz 1992) has been used for many species and many locations. It is the simplest plant phenology model, requiring only three parameters: y

y such that Sf =

∑R t0

f

( xt ) = F*

(2)

Chapter 4.1: Plant Development Models

if x t ≤ Tb1 ⎧0 Rf(xt) = ⎨ x − T if xt > Tb1 b1 ⎩ t

221 (3)

where y is the phenological event date; xt is the daily mean temperature; Rf(xt,) is the rate of development during quiescence; Sf is the state of development; Tb1 is the threshold temperature and t0 is the day when rate of development starts to accumulate. Rf is also commonly called degree-days, or forcing units, and Sf is the sum of degree-days, or state of forcing. The Thermal Time model does not take dormancy into account. Thus, at t0 the chilling requirements of the species are assumed to have been met, or that the plant does not have chilling requirements. Five mechanistic phenology models take dormancy into account: (1) the Sequential model (Richardson et al. 1974; Sarvas 1974; Hänninen 1987; 1990b), (2) the Parallel model (Landsberg 1974; Sarvas 1974; Hänninen 1987, 1990b), (3) the Alternating model (Cannell and Smith 1983; Murray et al. 1989; Kramer 1994b), (4) the Deepening Rest model (Kobayashi et al. 1982), and (5) the Four Phases model (Vegis 1964; Hänninen 1990b). The Sequential model assumes that forcing (warm) temperatures are not effective unless chilling requirements are fulfilled, because plants have no competence to respond to forcing temperatures as long as chilling requirements are not met. The Parallel model assumes that forcing temperatures are active simultaneously with chilling temperatures. The Alternating model assumes a negative exponential relationship between the sum of forcing units required for completion of quiescence and the sum of chilling units received. The Four Phases model assumes three phases of dormancy (pre-rest, true-rest, and post-rest) before the phase of quiescence. This is formalized by an increasing temperature threshold for forcing during pre-rest and a decreasing temperature threshold for forcing during post-rest, and buds cannot respond to forcing temperature at all during true rest. In the different models, the developmental responses to temperature during dormancy and during quiescence have been described by various types of functions (Figure 1). Forcing units have commonly been formulated either as degree-days (Eq. 3) or as follows (Figure 1b-1d):

⎧0 ⎪ a Rf(xt) = ⎨ ⎪⎩ 1 + e b ( x t − c )

if x t < 0 if x t ≥ 0

with a = 28.4, b = -0.185, c = 18.4 (Sarvas 1972; Hänninen 1990b).

(4)

222

Phenology: An Integrative Environmental Science

Figure 4.1-1. Examples of responses to temperature used to calculate chilling (a, Eq. 5; c, Eq. 6) and forcing units (b, Eq. 3; d, Eq. 4). Tb1, Tb2, base temperatures, To, optimal temperature.

Chilling units have commonly been formulated as chilling days:

⎧1 Rc(xt) = ⎨ ⎩0

if x t < T b 2 if x t ≥ T b 2

(5)

where Tb2 the base temperature (Figure 1a), or as a more complex function;

⎧ ⎪ if x t ≤ T m or x t ≥ T M ⎪0 ⎪ xt − Tm if x t > T m Rc(xt) = ⎨ T − T m ⎪ o ⎪ x t −TM if x t < T M ⎪ ⎩ To − T M

(6)

where To is the optimal temperature, Tm = -3.4 and TM = 10.4 (Sarvas 1974; Hänninen 1990b) (Figure 1c). The variety of model assumptions and formulations called for a consistent notation and for attempts at unification. Hänninen (1990b) first unified the Sequential, Parallel, Deepening Rest, and Four phases models, but kept the form of chilling and forcing temperature responses unchanged.

Chapter 4.1: Plant Development Models

223

Later, Hänninen (1995) suggested an approach in which chilling, forcing, and growth competence (the way in which chilling affects forcing), can be chosen freely (Hänninen 1995). Earlier, Kramer (1994a, 1994b) had suggested a fitting procedure to select between different forms of the three sub-models.

Figure 4.1-2. (a) Comparison of Rf(xt) (Eq. 3) with Tb1 = 5°C and 28.4 x Rf(xt) (Eq. 8) with d = -0.15587, e = 19.56, (b) Comparison of Rc(x ( t) (Eq. 6) and 2 x Rc(x ( t) (Eq. 7) with a = 0.07613, b = 0.0000086, c = 4.2, (c) Comparison of Rc(x ( t) (Eq. 5) with Tb2 = 10°C and Rc(x ( t) (Eq. 7) with a = 0.5, b = 50, c = 10.

The fitting procedure was improved by Chuine et al. (1998, 1999) using the simulated annealing algorithm (Press et al. 1989). Chuine (2000) developed a fully Unified model (Chuine 2000) based on two general functions that describe the relationships between temperature and the rates of chilling and forcing development: Rc (xt) =

1 1+ e

a ( xt − c ) 2 + b ( xt − c )

(7)

224

Phenology: An Integrative Environmental Science Rf (xt) =

1 1 + e d ( xt − e )

(8)

where Rc(x ( t) is the rate of development during dormancy and Rf(xt,) is the rate of development during quiescence. Most of the mechanistic models described above are special cases of the Unified model, depending on the choice of the parameter values of Eq. 7 and 8 (Figure 2), and a few additional parameters (for details see Chuine 2000).

3.

METHODOLOGICAL CONSIDERATIONS

The critical problem with mechanistic phenology models is that the basic biochemistry and biophysics during dormancy is currently unknown. The dormancy release process remains a black box in mechanistic phenology models because we lack a method of direct measurement of the state of development during dormancy. Thus we cannot make direct measurements of the model parameter values. Two approaches are used to obtain estimates of parameter values: the experimental approach, that analyzes the temperature response of bud growth and development under controlled conditions; and the statistical approach, that estimates the models’ parameter values using statistical model-fitting techniques.

3.1

The Experimental Approach

The experimental approach consists of experiments carried out to analyze the underlying mechanisms of phenological responses, one mechanism at a time. Sarvas (1972) determined experimentally the temperature response of the rate of development during quiescence, using observation of meiosis in pollen mother cells of several forest tree species. He noticed that developmental time, i.e. the time required on average to go from one meiotic phase to the next, declines exponentially with increasing incubation temperature. Thus, the rate of developmentt increases sigmoidally with temperature. He tested the model with the timing of flowering in forest stands. Although the Thermal Time model with a +5oC threshold temperature approximates quite well the temperature response measured in the laboratory, Sarvas (1972) also found that ontogenetic development could take place between –3°C and +5oC. As the physiological processes of dormancy break were, and still are, unknown, Sarvas (1974) determined the progress of dormancy break indirectly using regrowth tests where seedlings

Chapter 4.1: Plant Development Models

225

were incubated in growth-promoting forcing conditions following a period under chilling conditions. Both the duration and temperatures during the chilling conditions were varied systematically. Sarvas found that the duration of chilling required for rest completion was shortest at +3.5oC, so the rate of development (rate of dormancy break) was highest at this temperature. These results led to the bell-shaped temperature response (peaking at 3.5°C) proposed for the rate of dormancy break (Figure 2b) An experimental approach was also used by Hänninen (1990b), who developed the first version of the unified model to compare various model assumptions concerning the effects of chilling on the response of buds to forcing temperatures. According to Hänninen’s results with Scots pine and Norway spruce seedlings, the effects fell between the assumptions of the Sequential and the Parallel model.

3.2

The Statistical Approach

The statistical approach estimates parameter values with statistical model-fitting techniques. In this approach, field or experimental data on the timing of the phenological event are related to air temperature data gathered at the same location before the event. Two techniques have been used, both estimating parameters using the least squared residuals method. The easiest method is to fix all but one parameter to a given value, and find the value of the free parameter that minimized the sum of squared residuals. All parameters are varied this way one after the other. This technique has several limits, most importantly (i) a finite number of parameter values can be tested, (ii) parameter values are estimated independently from each other although they are usually not independent, (iii) the least squares function may have several local optima and it is almost impossible to find the global optimum without a more systematic search. More efficient methods consist in estimating all parameters simultaneously using optimization algorithms. Traditional optimization algorithms, e.g. Downhill Simplex, Newton methods (Press et al. 1989) however rarely converge towards the global optimum (Kramer 1994b; Chuine et al. 1998, 1999). The simulated annealing method is more efficient in this respect (Chuine et al. 1998, 1999) because it is especially designed for functions with multiple optima. However, accurate parameter estimation is not sufficient, especially in the case of phenology models, where prediction accuracy is particularly important because these models must predict future phenology in the coming year (e.g., for orchard management) or over the next centuries (e.g., for global warming impact assessment). An adequate testing method is crossvalidation (Chatfield 1988). In cross-validation, the model is tested by

226

Phenology: An Integrative Environmental Science

comparing its predictions to observations not used in model fitting. However, this method is data-hungry and it is not always possible to split the dataset into two parts, one to fit the model, the other to test its prediction accuracy. The above discussion shows that parameter estimates of phenology models can be developed with two quite different approaches. The experimental approach uses detailed ecophysiological laboratory or greenhouse experiments. This is a time-consuming process. Because model parameters may be under genetic control, they should be measured for different populations. The statistical approach is much quicker, provided that sufficiently long phenological and temperature records are available, and that adequate statistical methods are used. However, this may be too rough an approach and a combination of the two (experimental and statistical) is probably the best solution to obtain accurate and realistic models.

4.

APPLICATIONS OF PLANT PHENOLOGY MODELS

Plant phenology models are important tools in a wide range of issues such as (1) prediction of the impact off global warming on the phenology of wild and cultivated species (Pouget 1963, 1966; Richardson et al. 1974; Ashcroft et al. 1977; Swartz and Powell 1981; Anderson et al. 1986; Osborne et al. 2000), (2) improvement of primary productivity models (Lieth 1970, 1971; Kramer and Mohren 1996), (3) prediction of the occurrence of pollen in the atmosphere, and thus the occurrence of pollen allergies (Frenguelli et al. 1991; Frenguelli et al. 1992; Marletto et al. 1992; Chuine et al. 1998; Frenguelli and Bricchi 1998; Chuine et al. 1999), and (4) support of foresters and farmers in management decisions, such as selection of provenances for reforestation and to prevent frost damage to fruit trees. In the following paragraphs we describe some of these applications.

4.1

Frost Hardiness Modeling

Phenology models of bud burst have been frequently applied to assess the risk of frost damage to perennial plants (Cannell et al. 1985; Cannell and Smith 1986; Hänninen 1991; Kramer 1994a; Linkosalo et al. 2000). Bud development and growth is highly correlated to dehardening. Frost hardiness gradually increases while dormancy sets-in and is gradually lost during quiescence once dormancy is broken. Thus, the risk of frost damage can be assessed by estimating minimum air temperatures around bud burst (Cannell

Chapter 4.1: Plant Development Models

227

et al. 1985; Murray et al. 1989; Hänninen 1991). Repeating the simulations over many years, a frequency distribution of minimum temperatures occurring around bud burst can be compiled, and from this the probability of frost damage (the proportion of years when air temperature drops below a genotype-specific damage threshold). Using this approach, both the suitability of different tree provenances and cultivars for the current climate (Cannell et al. 1985; Hänninen and Hari 1996) and the effects of climatic warming on native trees (Cannell and Smith 1986; Murray et al. 1989; Hänninen 1991; Kramer 1994a; Murray et al. 1994) have been assessed. More mechanistic models of hardiness and frost damage have been developed that simulate frost hardiness and frost damage over the whole year, and not only around bud burst (Kobayashi and Fuchigami 1983b; Repo et al. 1990; Kellomäki et al. 1995; Leinonen et al. 1995). Leinonen (1996) developed the most complex and probably most accurate such model. In this model the daily state of hardiness is regulated by air temperature and photoperiod, and the frost hardiness response to these environmental factors depends on the current state of ontogenic development. The minimum temperature that can be sustained with no damage varies during the annual cycle. In his model Leinonen introduced an index of injury that responds to temperature according to the current frost hardiness.

8

gC m-2 d-1

6 4 2 0 -2 0 -4

100

200 0

300

time (d)

Figure 4.1-3. Fluxes of CO2 measured (dots) and predicted by the model FORGRO (line) at a Pinus sylvestris forest in Hyytiala, Finland. Data kindly provided by Prof. T. Vesala. Reprinted from Mohren (1999), © SPB Academic Publishing, The Hague, The Netherlands, used with permission.

228

4.2

Phenology: An Integrative Environmental Science

Forest Growth and Climate Change

An important application of phenology models is their coupling with general models of forest growth to assess climate change impacts. FORGRO is a model that uses phenology models and frost hardiness models to simulate tree growth and productivity (Mohren 1987; Kramer 1995a; Kramer et al. 1996; Leinonen and Kramer 2002). The onset and end of the growing season can be observed either by recording the changes of the canopy such as bud burst, autumn coloration or loss of foliage, or by measuring gas exchanges between the vegetation and the atmosphere. An example of the latter is presented in Figure 3. The data were collected using the eddy-covariance method in a tower above the forest (Vesala et al. 1998). Part of the springtime CO2 flux is caused by the activity of the understory, which is not described by this model. The decline of the CO2 exchange from mid-summer to autumn is mainly the direct effect of decreasing light availability and temperature on photosynthesis. A rise in atmospheric CO2 concentration and temperature influences a multitude of processes in a tree and in a forest stand. FORGRO describes the direct effects of CO2 and temperature on photosynthesis, and the direct effect of temperature on both plant and soil respiration. The description of these processes can be found in Kramer et al. (1996) and Mohren (1987). Indirect effects of temperature include the duration of the growing season and the level of frost hardiness. Figure 4 shows thatt the duration of the growing season is extended considerably by an anticipated 6°C increase in minimum winter temperature (Bach 1987), and that the “safety margin” between minimum daily temperature and the level of frost hardiness is much reduced as a result.

4.3

Species Range Modeling

The seasonal coordination of phenology to local climate conditions has several major impacts on plant survival and reproduction (fitness), as well as on competitive relationships via vegetative and reproductive performances (Lechowicz and Koike 1995). Using a species range model (PHENOFIT) Chuine and Beaubien (2001) showed that phenology was a major determinant of species range. PHENOFIT estimates survival and reproductive success based on the match between annual plant development and local seasonal variations of climate. A mismatch between the two may result in frost injuries to flowers and leaves, but also in drought injuries should the vegetation period occur during the drought season, or in low fecundity should the period between flowering and fall be too short or too cold for fruit to mature (Pigott and Huntley 1981). These mismatch lead to decreasing primary productivity, survival and reproductive success.

Chapter 4.1: Plant Development Models

229

PHENOFIT has been calibrated for two North American species, quaking aspen, Populus tremuloides Michx. and sugar maple, Acer saccharum Marsh. Calibration consisted in fitting f phenology models for budburst, flowering and fruiting. As no leaf coloration model exists so far, the date of leaf coloration is estimated from a linear relationship with latitude from records taken at the northern and the southern boundaries of the species’ range. The geographical distribution where PHENOFIT estimated non-null

Figure 4.1-4. Duration of the growing season (grey horizontal bars) for Pinus sylvestris in Finland, and frost hardiness in current (A) and future (B), rise of the minimum daily temperature by 6°C. Reprinted from Kramer et al. (2000, Figures 4 a and b, p. 72), © Springer-Verlag, Heidelberg, Germany, used with permission.

fitness matched the observed species distributions very well. Sensitivity analysis showed that killing frost never happens and is thus not the factor determining the northern boundary of either species’ distribution. It also showed that phenology coupled to frost injury (both are intimately linked) and to a lesser extent drought injury, explained most of the sugar maple’s range. The distribution of quaking aspen was explained by drought injury, and to a lesser extent by phenology and by frost injury. The northern and southern boundaries of both species were determined by the inability to fully develop leaves or flowers (by lack of forcing temperature in the north, and lack of chilling temperature in the south). Elsewhere between these

230

Phenology: An Integrative Environmental Science

extremes, the distribution of sugar maple was limited by frost injury to leaves and flowers, and that of quaking aspen by frost injury to flowers and drought injury. This illustrates why phenology should be a major component of ecological studies, in particular when addressing global warming impacts.

ACKNOWLEDGEMENTS The authors thank the anonymous referee for valuable comments and revisions that improved this chapter. The financial support for Kramer by the EU-project DynaBeech (QLRT 1-CT99-1210) and the national research program: Societal and Environmental Importance of Nature, Forests, and Landscape (nr. 381) of the Dutch Ministry of Agriculture, Nature Management and Fisheries is gratefully acknowledged.

REFERENCES CITED Andersen, T. B., A model to predict the beginning of the pollen season, Grana, 30, 269-275, 1991. Anderson, J. L., C. D. Kesner, and E. A. Richardson, Validation of chill unit and flower bud phenology models for Montmorency sour cherry, Acta Hort, 184, 71-77, 1986. Ashcroft, G. L., E. A. Richardson, and S. D. Seeley, A statistical method of determining chill unit and growing degree hour requirements for deciduous fruit trees, Hort Sci., 12, 347348, 1977. Bach, W., Development of climatic scenarios from general circulation models, in The impact of climatic variations on agriculture, Vol. 1: Assessment on Cool Temperate and Cold Regions, edited by Parry, M. L., T. R. Carter and N. T. Konijn, pp. 125-157, Kluwer Academic Publishers, Dordrecht, 1987. Boyer, W. D., Air temperature, heat sums, and pollen shedding phenology of longleaf pine, Ecology, 54, 421-425, 1973. Campbell, R. K., and A. I. Sugano, Phenology of bud burst in Douglas-fir related to provenance, photoperiod, chilling and flushing temperature, Bot. Gaz., 136, 290-298, 1975. Cannell, M. G. R., Chilling, thermal time and the dates of flowering of trees, in Manipulation of fruiting, edited by C. J. Wright, pp. 99-113, Butterworth and Co, London, 1989. Cannell, M. G. R., M. B. Murray, and L. J. Sheppard, Frost avoidance by selection for late budburst in Picea sitchensis, J. Appl. Ecol., 22, 931-941, 1985. Cannell, M. G. R., and R. I. Smith, Thermal time, chill days and prediction of budburst in Picea sitchensis, J. Appl. Ecol., 20, 951-963, 1983. Cannell, M. G. R., and R. I. Smith, Climatic warming, spring budburst and frost damage on trees, J. Appl. Ecol., 23, 177-191, 1986. Chatfield, C., Problem solving: a statistician guide, Chapman and Hall, London, 261 pp., 1988.

Chapter 4.1: Plant Development Models

231

Chuine, I., A unified model for the budburst of trees, J. Theor. Biol., 207, 337-347, 2000. Chuine, I., and E. Beaubien, Phenology is a major determinant of temperate tree distributions, Ecol. Letters, 4, 500-510, 2001. Chuine, I., P. Cour, and D. D. Rousseau, Fitting models predicting dates of flowering of temperate-zone trees using simulated annealing, Plant, Cell and Env., 21, 455-466, 1998. Chuine, I., P. Cour, and D. D. Rousseau, Selecting models to predict the timing of flowering of temperate trees: implication for tree phenology modelling, Plant, Cell and Env., 22, 113, 1999. Ellis, R. H., E. H. Roberts, and R. J. Summerfield, Variation in the optimum temperature for rates of seedling emergence and progress towards flowering among six genotypes of faba bean (Vicia faba), Ann. Bot., 62, 119-126, 1988. Emberlin, J., J. Mullins, J. Corden, W. Millington, M. Brooke, M. Savage, and S. Jones, The trend to earlier Birch pollen season in the U. K.: a biotic response to changes in weather conditions?, Grana, 36, 29-33, 1997. Falusi, M., and R. Calamassi, Geographic variation and bud dormancy in beech seedlings (Fagus sylvatica L), Ann. Sci. For., 53, 967-979, 1996. Frenguelli, G., and E. Bricchi, The use of pheno-climatic model for forecasting the pollination of some arboreal taxa, Aerobiologia, 14, 39-44, 1998. Frenguelli, G., E. Bricchi, B. Romano, M. F. Ferranti, and E. Antognozzi, The role of air temperature in determining dormancy release and flowering of Corylus avellana L., Aerobiologia, 8, 415-418, 1992. Frenguelli, G., E. Bricchi, B. Romano, G. Mincigriucci, and F. T. M. Spieksma, A predictive study on the beginning of pollen season for Gramineae and Olea europaea L., Aerobiologia, 5, 64-70, 1989. Frenguelli, G., T. M. Spieksma, E. Bricchi, B. Romano, G. Mincigrucci, A. H. Nikkels, W. Dankaart, and F. Ferranti, The influence of air temperature on the starting dates of the pollen season of Alnus and Poplulus, Grana, 30, 196-200, 1991. Häkkinen, R., Statistical evaluation of bud development theories: application to bud burst of Betula pendula leaves, Tree Physiol., 19, 613-618, 1999. Häkkinen, R., T. Linkosalo, and P. Hari, Methods for combining phenological time series: application to bud burst in birch ((Betula pendula) in Central Finland for the period 18961955., Tree Physiol., 15, 721-736, 1995. Häkkinen, R., T. Linkosalo, and P. Hari, Effects of dormancy and environmental factors on timing of bud burst in Betula pendula, Tree Physiol., 18, 707-712, 1998. Hänninen, H., Effects of temperature on dormancy release in woody plants: implications of prevailing models., Silva Fenn., 21, 279-299, 1987. Hänninen, H., Modeling dormancy release in trees from cool and temperate regions, in Process modeling of forest growth responses to environmental stress, edited by R. K. Dixon, R. S. Meldahl, G. A. Ruark and W. G. Warren, pp. 159-165, Timber Press, Portland, 1990a. Hänninen, H., Modelling bud dormancy release in trees from cool and temperate regions., Acta Forest. Fenn., 213, 1-47, 1990b. Hänninen, H., Does climatic warming increase the risk of frost damage in northern trees?, Plant, Cell and Env., 14, 449-454, 1991.

232

Phenology: An Integrative Environmental Science

Hänninen, H., Effects of climatic change on trees from cool and temperate regions: an ecophysiological approach to modelling of budburst phenology, Can. J. Bot., 73, 183-199, 1995. Hänninen, H., and P. Hari, The implications of geographical variation in climate for differentiation of bud dormancy ecotypes in Scots pine, Acta Forest. Fenn., 254, 11-21, 1996. Hänninen, H., S. Kellomäki, K. Laitinen, B. Pajari and, T. Repo, Effect of increased winter temperature on the onset of height growth of Scots pine: a field test of a phenological model, Silva Fenn., 27, 251-257, 1993. Heide, O. M., Dormancy release in beech buds (Fagus ( sylvatica) requires both chilling and long days, Physio. Plant., 89, 187-191, 1993. Hunter, A. F., and M. J. Lechowicz, Predicting the timing of budburst in temperate trees, J. of Appl. Ecol., 29, 597-604, 1992. Kellomäki, S., H. Hänninen, and M. Kolström, Computations on frost damage to Scots pine under climatic warming in boreal conditions, Ecol. Appl., 5, 42-52, 1995. Kikuzawa, K., A cost-benefit analysis of leaf habit and leaf longevity of trees and their geographical pattern., Am. Nat., 138, 1250-1263, 1991. Kikuzawa, K., The basis for variation in leaf longevity of plants, Vegetatio, 121, 89-100, 1995a. Kikuzawa, K., Leaf phenology as an optimal strategy for carbon gain in plants, Can. J. Bot., 73, 158-163, 1995b. Kikuzawa, K., Geographical distribution of leaf life span and species diversity of trees simulated by a leaf-longevity model., Vegetatio, 122, 61-67, 1996. Kikuzawa, K., and G. Kudo, Effects of the length of the snow-free period on leaf longevity in alpine shrubs: a cost-benefit model, Oikos, 73, 214-220, 1995. Kobayashi, K. D., and L. H. Fuchigami, Modeling bud development during the quiescent phase in red-osier dogwood (Cornus sericea L.), Agr. Meteo., 28, 75-84, 1983a. Kobayashi, K. D., and L. H. Fuchigami, Modelling temperature effects in breaking rest in Red-osier Dogwood (Cornus sericea L.), Ann. Bot., 52, 205-215, 1983b. Kobayashi, K. D., L. H. Fuchigami, and M. J. English, Modelling temperature requirements for rest development in Cornus sericea, J. Am. Soc. Hor. Sci., 107, 914-918, 1982. Kramer, K., A modelling analysis of the effects of climatic warming on the probability of spring frost damage to tree species in The Netherlands and Germany, Plant, Cell and Env., 17, 367-377, 1994a. Kramer, K., Selecting a model to predict the onset of growth of Fagus sylvatica., J. Appl. Ecol., 31, 172-181, 1994b. Kramer, K., Modelling comparison to evaluate the importance of phenology for the effects of climate change in growth of temperate-zone deciduous trees, Clim. Res., 5, 119-130, 1995a. Kramer, K., Phenotypic plasticity of the phenology of seven European tree species in relation to climatic warming, Plant, Cell and Env., 18, 93-104, 1995b. Kramer, K., A. D. Friend, and I. Leinonen, Modelling comparison to evaluate the importance of phenology and spring frost damage for the effects of climate change on growth of mixed temperate-zone deciduous forests, Clim. Res., 7, 31-41, 1996.

Chapter 4.1: Plant Development Models

233

Kramer, K., I. Leinonen, and D. Loustau, The importance of phenology for the evaluation of impacts of climate change on growth of boreal, temperate and Mediterranean forests ecosystems: an overview, Int. J. Biometeorol., 44, 67-75, 2000. Kramer, K., and G. M. J. Mohren, Sensitivity of FORGRO to climatic change scenarios: a case study on Betula pubescens, Fagus sylvatica and Quercus roburr in the Netherlands, Clim. Change, 34, 231-237, 1996. Kupias, R. and Y. Mäkinen, Correlations of Alder pollen occurrence to climatic variables, First international conference on aerobiology, Munich, 1980. Lamb, R. C., Effects of temperature above and below freezing on the breaking of rest in the Latham raspberry, J. Am. Soc. Hort. Sci., 51, 313-315, 1948. Landsberg, J. J., Apple fruit bud development and growth; analysis and an empirical model., Ann. Bot., 38, 1013-1023, 1974. Lechowicz, M. J., and T. Koike, Phenology and seasonality of woody plants: An unappreciated element in global change research., Can. J. Bot., 73, 147-148, 1995. Leinonen, I., A simulation model for the annual frost hardiness and freeze damage of Scots Pine, Ann. Bot., 78, 687-693, 1996. Leinonen, I., and K. Kramer, Applications of phenological models to predict the future carbon sequestration potential of Boreal forests, Clim. Change, 55, 99-113, 2002. Leinonen, I., T. Repo, H. Hänninen, and K. Burr, A second-order dynamics model for the frost hardiness of trees., Ann. Bot., 76, 89-95, 1995. Lescourret, F., N. Blecher, R. Habib, J. Chadboeuf, D. Agostini, O. Paliiy, B. Vaissière, and I. Poggi, Development of a simulation model for studying kiwi fruit orchard management, Agr. Syst., 59, 215-239, 1999. Lieth, H., Phenology in productivity studies, in Analysis of temperate forest ecosystems, 1, edited by D. E. Reichle, pp. 29-55, Springer Verlag, Heidelberg, 1970. Lieth, H., The phenological viewpoint in productivity studies, in Productivity of forest ecosystems. Proceedings of the Brussels Symposium by UNESCO., edited by P. Duvigneaud, pp 71-83, UNESCO, Paris, 1971. Linkosalo, T., Regularities and patterns in the spring phenology of some boreal trees, Silva Fenn., 33, 237-245, 1999. Linkosalo, T., T. Carter, R. Häkkinen, and P. Hari, Predicting spring phenology and frost damage risk of Betula spp. under climatic warming: a comparison of two models, Tree Physiol., 20, 1175-1182, 2000. Linkosalo, T., R. Häkkinen, and P. Hari, Improving the reliability of a combined phenological times series by analyzing observation quality, Tree Physiol., 16, 661-664, 1996. Marletto, V., G. P. Branzi, and M. Sirotti, Forecasting flowering dates of lawn species with air temperature: application boundaries of the linear approach, Aerobiologia, 8, 75-83, 1992. Menzel, A, and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999. Mohren, G. M. J., Simulation of forest growth, applied to Douglas fir stands in the Netherlands, Wageningen Agricultural University, Wageningen, The Netherlands, 184 pp., 1987.

234

Phenology: An Integrative Environmental Science

Mohren, G. M. J., H. H. Bartelink, K. Kramer, F. Magnani, S. Sabaté and D. Loustau, Modelling long-term effects of CO2 increase and climate change on European forests, with emphasis on ecosystem carbon budgets, in Forest ecosystem modelling, upscaling and remote sensing, edited by R. J. M. Ceulemans, F. Veroustreate, V. Gond, and J. B. H. F. V. Rensbergen, pp. 179-192, SPB Academic Publishing, The Hague, 1999. Murray, M. B., G. R. Cannell, and R. I. Smith, Date of budburst of fifteen tree species in Britain following climatic warming., J. Appl. Ecol., 26, 693-700, 1989. Murray, M. B., R. I. Smith, I. D. Leith, D. Fowler, H. S. Lee, A. D. Friend, and P. G. Jarvis, Effects of elevated CO2, nutrition and climatic warming on bud phenology in Sitka spruce ((Picea sitchensis) and their impact on the risk of frost damage, Tree Physiol., 14, 691-706, 1994. Nizinski, J. J., and B. Saugier, A model of leaf budding and development for a mature Quercus forest., J. Appl. Ecol., 25, 643-652, 1988. Oliveira, M., Calculation of budbreak and flowering base temperatures for Vitis vinifera cv. Touriga Francesa in the Douro region of Portugal, Am. J. Enol. Vitic., 49, 74-78, 1998. Osborne, C. P., I. Chuine, D. Viner, and F. I. Woodward, Olive phenology as a sensitive indicator of future climatic warming in the Mediterranean, Plant, Cell and Env., 23, 701710, 2000. Phillipp, M., J. Böcher, O. Mattson, and S. L. J. Woodell, A quantitative approach to the sexual reproductive biology and population structure in some Arctic flowering plants: Dryas integrifolia, Silene acaulis and Ranunculus nivalis, Medd Grönl Biosciences, 34, 160, 1990. Pigott, C. D., and J. P. Huntley, Factors controlling the distribution of Tilia cordata at the Northern limits of its geographical range. III Nature and cause of seed sterility, New Phytol.,87, 817-839, 1981. Pipper, E. L., K. L. Boote, J. W. Jones, and S. S. Grimm, Comparison of two phenology models for predicting flowering and maturity date of soybean, Crop Sci., 36, 1606-1614, 1996. Pouget, R., Recherches physiologiques sur le repos végétatifs de la vigne (Vitis vinifera L;): la dormance des bourgeons et le mécanisme de sa disparition, INRA, Paris, 1963. Pouget, R., Etude du rythme végétatif: caractères physiologiques liés à la précocité de débourrement chez la vigne, Annales de l'amélioration des plantes, 16, 6-100, 1966. Press, W. H., B. P. Flannery, S. A. Teukolsky, and W. T. Vetterling, Numerical recipes in Pascal, Cambridge University Press, Cambridge, 759 pp., 1989. Reaumur, R. A. F. de, Observations du thermomètre, faites à Paris pendant l'année 1735, comparées avec celles qui ont été faites sous la ligne, à l'isle de France, à Alger et quelques unes de nos isles de l'Amérique., Memoires de l'Académie des Sciences de Paris, 1735. Reich, P. B., Phenology of tropical forests: patterns, causes, and consequences, Can. J. Bot., 73, 164-174, 1994. Repo, T., A. Mäkelä, and H. Hänninen, Modelling frost resistance of trees, Silva Carelica, 15, 61-74, 1990. Richardson, E. A., S. D. Seeley, and D. R. Walker, A model for estimating the completion of rest for 'Redhaven' and 'Elberta' peach trees, Hort. Science, 9, 331-332, 1974.

Chapter 4.1: Plant Development Models

235

Roberts, E. H., R. J. Summerfiled, R. H. Ellis, and K. A. Stewart, Photothermal time for flowering in lentils ((Lens culinaris) and the analysis of potential vernalization responses, Ann. Bot., 61, 23-39, 1988. Robertson, G. W., A biometeorological time scale for a cereal crop involving day and night temperatures and photoperiod, Int. J. Biometeorol., 12, 191-223, 1968. Sarvas, R., Investigations on the annual cycle of development on forest trees active period, Communicationes Instituti Forestalis Fenniae, 76, 110, 1972. Sarvas, R., Investigations on the annual cycle of development of forest trees: Autumn dormancy and winter dormancy, Communicationes Instituti Forestalis Fenniae, 84, 1101, 1974. Schwartz, M. D., Spring index models: an approach to connecting satellite and surface phenology, in Phenology in seasonal climates II, edited by H. Lieth and M. D. Schwartz, pp. 23-38, Backhuys Publishers, Leiden, 1997. Schwartz, M. D., Green-wave phenology, Nature, 394, 839-840, 1998. Schwartz, M. D., Advancing to full bloom: planning phenological research for the 21st century, Int. J. Biometeorol., 42, 113-118, 1999. Schwartz, M. D., and G. A. Marotz, An approach to examining regional atmosphere-plant interactions with phenological data, J. Biogeograph., 13, 551-560, 1986. Schwartz, M. D., and G. A. Marotz, Synoptic events and spring phenology, Phys. Geog., 9, 151-161, 1988. Sinclair, T. R., S. Kitani, J. Bruniard and T. Horide, Soybean flowering date: linear and logistic models based on temperature and photoperiod, Crop Sci., 31, 786-790, 1991. Spieksma, F. T. H., J. Emberlin, M. Hjelmroos, S. Jäger, and R.M. Leuschner, Atmospheric birch ((Betula) pollen in Europe: trends and fluctuations in annual quantities and the starting dates of the seasons, Grana, 34, 51-57, 1995. Swartz, H. J., and L.E. Powell, The effect of long chilling requirement on time of bud break in apple, Acta Horticulturae, 120, 173-177, 1981. Thorhallsdottir, T. E., Flowering phenology in the central highland of Iceland and implications for climatic warming in the Arctic, Oecologia, 114, 43-49, 1998. Vegis, A., Dormancy in higher plants, Annual review of plant physiology, 15, 185-224, 1964. Vesala, T., J. Haataja, P. Aalto, N. Altimir, G. Buzorius, E. Garam, K. Hämeri, H. Ilvesniemi, V. Jokinen, P. Keronen, T. Lahti, T. Markkanen, J.M. Mäkelä, E. Nikinmaa, S. Palmroth, L. Palva, T. Pohja, J. Pumpanen, Ü. Rannik, E. Siivola, H. Ylitalo, P. Hari, and M. Kulmala, Long-term field measurements of atmosphere-surface interactions in boreal forest combining forest ecology, micrometeorology, aerosol physics and atmospheric chemistry, Trends in Heat, Mass and Momentum Transfer, 4, 17-35, 1998. Winter, F., A simulation model of phenology and corresponding frost resistance in 'Golden delicious' apple, Acta Horticulturae, 184, 103-107, 1986.

Chapter 4.2 ANIMAL LIFE CYCLE MODELS Jacques Régnière1 and Jesse A. Logan2 1 Natural Resources Canada, Canadian Forest Service, Quebec, Canada; 2USDA Forest Service, Logan, Utah, USA

Key words:

1.

Animals, Models, Landscape, Entomology, Temperature

INTRODUCTION

Maintaining an appropriate seasonality is a basic ecological requisite for all organisms. Critical life-cycle events must be keyed to the appropriate seasonal cycles, whether it be the wet-season/dry-season cycle of the tropics or the summer/winter cycle of temperate zones. The more pronounced the seasonal climatic signal, the stronger the need for an appropriate timing maintained through phenology. Although adaptive seasonal timing is no less important for warm-blooded animals than for poikilotherms (e.g. appropriately timed reproduction, hibernation), phenology models have been applied mostly to plants and cold-blooded animals. The phenological habitat can conveniently be thought of as a hyperspace with temperature, moisture, nutrition, and photoperiod constituting the defining axes. The potential effect of moisture on phenology is expressed as a mortality differential, especially at extremes. Photoperiod is a consistent seasonal variable in the temperate zone that serves as an important environmental cue for the initiation of critical events such as diapause (a physiological hibernation or aestivation state in insects). Diapause is a basic physiological process that typically serves to synchronize and reset the seasonal cycle of phenological events in the face of desynchronizing variability in seasonal temperatures. Nutrition can affect phenology in the Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 237-254 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

238

Phenology: An Integrative Environmental Science

same way as moisture through mortality, or by altering development rates. It can also serve as an environmental cue for the initiation of diapause. Temperature is the strongest determinant of poikilotherm phenology, and is also the best understood. Thus, the focus of phenology modeling efforts has been to relate temperature to resulting phenological events. There is a large body of scientific literature relating the phenological responses of poikilotherms to temperature. Much of this knowledge pertains to insects or their close relatives. Our focus will therefore be on modeling the seasonal life cycles of insects. The underlying physiological mechanisms of response to temperature are similar enough that the methods we describe are applicable to all poikilotherms.

2.

TEMPERATURE-DEPENDENT MODELS

Relating temperature to the development of insects, requires differentiating age from stage. Although both are related to time, age is strictly chronological in nature, while stage is a developmental concept typically defined by distinct morphological characteristics, often requiring a molt for transition from one stage to the next. Another time-related concept, development rate, is the temporal progression through an instar or stage and is dependent on temperature in a predictable fashion. Assuming that development rate is a constant function r(T) T of temperature T within a stage, it is the first derivative of the state of developmentt relative to time; in practice it is the inverse (1/x / ) of the time required to complete that life stage. For variable temperatures T( T t), where t is time, we formally define the state of development a (also referred to as physiological age), as:

d aj ( ) dt

[

( )];

t

j

(

j

) 0;

j

()

∫ [ j

( )] ;

j

(

j

) 1 (1)

t j −1

Life stage j begins at tj-1, which is the time of completion of the previous life stage (ttj-1 as indicated by the initial condition of the differential equation T t)] t Δt, during short above). Numerically, developmental increments Δaj = r[T( time steps Δt are summed over time until aj = 1, which indicates the completion of life stage j and defines tj, the time at the end of life stage j. These relationships underlie almost all models of insect phenology (see Logan and Powell 2001). Once the mathematical relationship between temperature, time, and physiological age is defined, there remains the issue

Chapter 4.2: Animal Life Cycle Models

239

of finding an appropriate functional relationship between temperature and the development rate r(T). T

2.1

Development Rate Functions

The earliest functional form used to describe the relationship between temperature and rate of development was the linear or day-degree model, a concept that dates back to the 1700s (Wang 1960), and it has been extensively used to model both animal and plant phenology. Parameters in the day-degree model can be estimated from either field or laboratory data. Laboratory data typically consist of experimentally measured development rates over a range of fixed temperatures. Field estimates are obtained by assuming a reasonable threshold temperature and then summing (integrating) temperature in excess of the threshold over time (often in days) until “heat units” accumulate to a specific total. Day-degree models often work well if the temperatures of ecological interest do not fall outside of the linear region of the organism’s thermal response (Figure 1a). Their advantage is their simplicity in estimating parameters and making phenological predictions. Their disadvantage is that they constitute a more or less adequate approximation (Wang 1960). The observation that development rates are nonlinear was made over 60 years ago, at least (Janisch 1932). It was not, however, until the mid 1970s that the use of nonlinear rate functions became widely practiced, mostly in response to the widespread availability of digital computers that provided methods for parameter estimation and convenient numerical solution of equation (1). Stinner et al. (1975), Logan et al. (1976), and Sharpe and DeMichele (1977) described nonlinear functions for insect temperature dependent development rates that have been widely applied since their introduction. This large body of literature supports the general shape of the development rate function as an exponential phase at low temperatures increasing to an optimum, and then a precipitous decline from the optimum temperature to the lethal thermal maximum (Figure 1a). Parameter estimation for nonlinear development rate models is more complex than for linear-based day-degree models. Procedures for estimating parameters of nonlinear development rate functions have been automated for a reasonable suite of equations (Wagner et al. 1984; Logan 1988). A nonlinear rate model is required whenever simulations must cover temperatures over the full range of physiological activity. Life cycle phenomena that involve temperature extremes (diapause for example) also require nonlinear representation. The widely reported acceleration of development under the variable temperature regimes that occur in nature is the natural consequence of nonlinearity (e.g., Ruel and Ayres 1999). Given the availability of digital

240

Phenology: An Integrative Environmental Science

computers for both parameter estimation and simulation, there is little reason not to use nonlinear rate functions.

Figure 4.2-1. (A) Typical shape of poikilotherm thermal response (development rate, in days-1). (B) Typical distribution of individual development rates relative to the median at a given temperature.

2.2

Modeling the Life Cycle

Once appropriate functional forms of r(T) T for each life stage have been established, phenology prediction becomes a question of solving equation (1) for the value of tj for each life stage j (for a graphical description of modeling phenology with equation (1), see Logan and Powell 2001). Daydegree models are easily evaluated. For models with nonlinear rate curves, the easiest approach is to use numerical integration for direct solution of equation (1). Procedures are available for automated construction and application of nonlinear rate based phenology models as well (e.g., Logan 1988). Prediction of phenology and particularly of the entire life cycle is best represented in mathematical terms as a circle map, relating the date of occurrence of a phenological event in one generation to its date of occurrence in the next. This circle map can then be used to determine the stability of seasonality (Powell et al. 2000; Logan and Powell 2001; Régnière and Nealis 2002). Logan and Powell (2001) call this circle map the G-function. A stable G-function is one that exhibits an attracting date (date towards which a phenological event converges within a few iterations of the annual temperature cycle, regardless of the starting date in the first generation). This convergence is usually very rapid (only a few generations), even in the absence of diapause or other bio-fix mechanisms. This is an indication of the capability of insects to adapt to changing climatic conditions.

Chapter 4.2: Animal Life Cycle Models

2.3

241

Including Developmental Variability

Ecological and biological phenomena often depend as much on variation in the population (Figure 1b) as on the average response. It is therefore often useful to include variability in development rates in phenology models. A common way of doing this is through time-varying distributed delays, a technique borrowed from engineering (Forrester 1961; Manetsch 1976; Vansickle 1977). An advantage of distributed delays is the availability of computational algorithms for implementation due to their wide application in engineering and other areas of applied mathematics. The major disadvantage is lack of flexibility in representing the distribution of development times. Cohort models have often been used to include variability in development rates (Curry et al. 1978). This modeling approach uses computed physiological age as the independent variable for the cumulative probability density function of stage completion. The approach assumes that the normalized distribution of development rates does not vary with temperature. Once this assumption is made, simulation is a simple matter of computing the physiological age of all cohorts (a cohort is typically defined as the individuals that entered a life stage during one time-step of the model) that comprise a life stage. Development and implementation of these types of models have been automated (Logan 1988). The advantage of cohort approaches to modeling phenology is their flexibility in choice of distribution models. The primary disadvantage is the “same shape” assumption, which is questionable at temperature extremes. With the relentless advance in computing power, it is becoming more practical to simulate population processes through sub-populations (Régnière 1984) or individual-based models. In the latter approach, a sample population is drawn from an observed or assumed distribution of development rates (Figure 1b). Simulation is performed by solving equation (1) for each individual in the population. The advantage of individual-based models is their complete flexibility in selection of both development rate functions and probability density distribution. The major disadvantage is computational overhead. Depending on the type of application, this method may be too computationally intensive. However, object-oriented programming techniques have greatly enhanced the efficiency and usefulness of individual-based phenology modeling (e.g., Cooke and Régnière 1996).

242

3.

Phenology: An Integrative Environmental Science

MODELING LANDSCAPE INFLUENCES

One of the areas of central interest in animal ecology is the influence of landscapes on the outcome of ecological processes as environmental conditions, plant communities and movement influence them (Haila 2002; McGarigal and Cushman 2002). The ability to model these influences at the landscape level is a key to improving our predictive understanding of these outcomes so as to improve area-wide management of pests, resources and ecosystems (Boutin and Hébert 2002; Ryszkowski 2002). In the ecology of poikilotherms, landscapes play a key role in determining patterns of abundance through their influence on local climate (Chen et al. 1999). A system called BioSIM© was developed by Régnière et al. (1995) to perform most of the functions described below.

3.1

Main Influences of Landscape on Climate

Whiteman (2000) lists four factors that determine local climate over a given landscape: latitude, altitude above sea level, continentality (distance from the sea) and exposure to regional circulation (wind and ocean currents). Other factors that can also influence local climate are cold-air drainage and terrain shading; both are particularly important in steep landscapes (Bolstad et al. 1998). Gas physics, moisture content, and solar radiation explain much of the effects of latitude and elevation on air temperature over the landscape. Usually, these influences are modeled with thermal gradients. A dry (unsaturated) parcel of air cools at a rate of 0.98°C per 100 m elevation (the adiabatic lapse rate). In addition, air temperature drops by about 1°C per degree of latitude away from the equator. Actual thermal gradients are usually smaller than the adiabatic lapse rate; they vary with location, time of year, and even time of day (Régnière and Bolstad 1994), largely in response to the amount of moisture in the air, its general temperature, air circulation patterns and the proximity of large water bodies. For example, minimum (nighttime) and maximum (daytime) air temperature elevation gradients on the Pacific coast of British Columbia are very different from each other. Minimum temperature gradients average less than –0.5°C per 100 m. Maximum temperature gradients are actually inverted (positive) in summer months as a result of cold-water maritime influence (Figure 2a). Latitudinal gradients in the area are strongly patterned seasonally, with temperatures cooling down northward in winter, and actually warming northward in summer (Figure 2b), once again due to the presence of cold waters in the southern part of the area. Longitudinal gradients are also strongly patterned over the year (Figure 2c).

Chapter 4.2: Animal Life Cycle Models

243

Figure 4.2-2. Monthly thermal gradients from 1961 to 1990 averages for Vancouver Island and the Pacific coast of British Columbia (Canada). (a) elevation, (b) latitudinal, and (c) longitudinal gradients.

Several methods have been devised to interpolate climatic variables from a number of punctual data sources over a surrounding landscape. Two are of particular interest because of their relative simplicity and because of the general availability of their algorithms: GIDS (Gradients with Inverse Distance Squared weighting, Nalder and Wein 1998) and ANUSPLIN (thinplate smoothing splines, Hutchinson 1991). While their performance is very similar (Price et al. 2000), GIDS is most attractive because of its simplicity. It uses multiple linear regression fitted to data from a number (we have used 20) of nearby weather stations:

Y

a mE E mN N

mW W

(2)

where Y is a climate value (e.g., minimum air temperature), E is elevation, N is latitude and W is longitude of the region’s weather stations; a is an intercept constant, and mE, mN and mW are regional thermal gradients for elevation, latitude and longitude. These are applied to differences in latitude (ΔN ΔN), longitude (ΔW) W and elevation (ΔE ΔE) between the unsampled locations and a number (we have used 4) of nearby weather stations. The inverse of the squared distances (1/dd2) between these n nearby stations and unsampled locations are used as weights in the estimation of the Yu): climate datum (Y n

Yu =

∑ d1 ( i =1

i

)

2

(3)

n

∑ i =1

1 di 2

244

Phenology: An Integrative Environmental Science

This process (Figure 3) can be applied equally well to monthly climate statistics (normals) or to daily records to obtain air temperature and precipitation information for any number of unsampled points across a given landscape. Difficulties in application of this methodology sometimes arise when weather stations are exceedingly sparse or do not cover the range of elevations found in the landscape (multiple regression models are notoriously poor at extrapolation). For that reason, our implementation of the GIDS algorithm doubles, then triples, the number of stations used to estimate gradients whenever the elevation of the unsampled location is outside ± 10% of the range found among the set of nearby stations.

3.2

Daily Weather Generators

Daily weather generators play an important role in the investigation of climatic influences on animal ecology because of the cumulative nature of daily or even hourly conditions. This level of detail is especially important in understanding the ecology of fast-developing cold-blooded animals, such as insects. As discussed in earlier sections of this chapter, thermal responses are strongly nonlinear, even when described by degree-day approximations. Thus, average outcomes cannot be obtained from average inputs. Thirty days of monthly average temperature does not have the same effect as 30 days of variable temperature with the same average. This has been called the Kaufmann effect (Worner 1992). When general questions concerning past climate are asked, answers can often be obtained by providing models with actual past weather records. However, such an approach has several limitations. First, past weather records usually cover a limited period att any given location (especially in North America). Second, it is never clear just how “general” a conclusion actually is about a given ecological process because of the limited amount of historical data available, especially in view of the extreme variability of weather conditions. Daily weather generation provides a general approach that can be applied equally well to past, present and future conditions under climate change. Several daily weather generators have been developed (Richardson 1981; Richardson and Wright 1984; Racsko et al. 1991; Hutchinson 1995; Wilks 1999), but many require considerable amounts of input information and often must be re-parameterized for application in specific geographical areas. Régnière and Bolstad (1994) developed a generally applicable algorithm (TempGen) for simulation of daily minimum and maximum air temperature using monthly normals (long-term average and extreme minimum and maximum temperatures). This generator is being expanded to generate realistic daily rainfall and solar radiation as well, and to mimic

Chapter 4.2: Animal Life Cycle Models

245

natural variation in mean monthly temperature and precipitation to simulate extreme events such as drought, untimely frost, heat waves, etc. It is also being validated for application in North America and Europe (unpublished).

Figure 4.2-3. Digital elevation model of British Columbia and Alberta (Canada). Climate information for unsampled locations (for example Ƒ, centered in larger circle) can be obtained with the GIDS method (Nalder and Wein 1998) in which data from several nearby weather stations (ż) are used to estimate regional thermal gradients by multiple linear regression. These gradients are then applied to differences in latitude, longitude, and elevation between station data and unsampled locations (an inverse distance-squared weighted average).

Because TempGen uses monthly averages as input, it is quite well suited to accept the climate-change scenarios generated by Global Circulation Models (two examples that have become widely used are the CGCM1 model developed by the Canadian Centre for Climate Modeling and Analysis, and another by the Hadley Center for Climate Prediction and Research). Output from TempGen, based on climate change normals, can therefore be used readily to simulate the impact of global warming on ecological processes modeled from daily climate inputs.

3.3

Spatial Interpolation Methodology

Running simulation models of animal development that use daily weather inputs can be demanding even for relatively fast computers. This means that it may be prohibitively time consuming to produce model output for each unit (pixel or raster) of a landscape (output map), except with the most

246

Phenology: An Integrative Environmental Science

simple degree-day models (e.g., Russo et al. 1993). A solution is to run models for a relatively small number (a few hundred) of randomly located points across a landscape, and to use a spatial interpolation method to estimate model output at other locations on the landscape. This approach was first proposed by Schaub et al. (1995), who used a linear regression between elevation and phenological target event (which they called a tfunction) to transform a digital terrain model of the landscape algebraically into a phenology map. Régnière (1996) expanded this t-function concept from a simple regression with elevation to a spatial regression using latitude, longitude, elevation, slope and aspect, their squares and interactions as predictors. Régnière and Sharov (1999) used universal kriging with elevation as an external drift variable as an interpolation method (see Isaaks and Srivastava 1989 for methodological details). Although general rules are difficult to provide, it has been our experience that spatial regression works better over smaller and steep landscapes, where elevation, slope and aspect are the main determinant of climatic variation; over larger areas, or where topography is less pronounced, kriging often works somewhat better (Gignac 2000). Other interpolation techniques exist, for example the GIDS method described earlier (which is a local regression technique) or inverse-distance weighted averaging. The choice of the most appropriate interpolation method can be made on the basis of cross-validation. This procedure consists of successively removing and replacing the known model output values at simulation points, estimating them with the chosen interpolation method and comparing the two sets (model output and the interpolated values). Simple coefficients of determination can be calculated to provide an objective comparison criterion between the interpolation methods tested (Figure 4). In the following example from spruce budworm (see below), the best interpolation method is spatial regression (R² = 0.812).

4.

EXAMPLES FROM ENTOMOLOGY

4.1

Spruce Budworm

Spruce budworm, Choristoneura fumiferana (Clem.), is a tortricid moth whose larvae defoliate conifer trees (firs and spruces of the Abies and Picea genera) of northeastern North America on a somewhat regular cycle of 3040 year (Royama 1984). It is an insect with obligate larval winter diapause, which means that by the end of winter the entire population is synchronized in the early stages of post-diapause development; spring emergence of larvae

Chapter 4.2: Animal Life Cycle Models

247

Figure 4.2-4. Cross-validation of defoliation by spruce budworm as predicted by Cooke’s model and estimated by four spatial interpolation methods. (A) universal kriging; (B) GIDS; (C) spatial regression; and (D) inverse-distance squared weighted average. Lines are the equality diagonal.

occurs within a week, two at the most, as soon as sufficient developmentinducing warmth has occurred. Cooke and Régnière (1996) developed an object-oriented, individual-based model to simulate interactions between spruce budworm, host trees, and the bacterium Bacillus thuringiensis (B.t. ( ) used as a bio-pesticide. The resulting so-called Cooke’s model has been extensively validated (Régnière and Cooke 1998). Cooke’s model can be used as a landscape-level pest management tool to assist in optimizing the delivery of pest management operations such as sampling, pheromone trap deployment, and pesticide applications. The efficacy of aerial sprays of B.t. is nonlinearly dependent on the stage of development of the target insect. Obviously, a pesticide application made prior to the emergence of larvae, before the onset of feeding, will not be efficacious. An application made too late, after much of the feeding damage has been done, may kill insects, but cannot protect foliage that has already been consumed. Thus, there is an optimal timing of applications that is based on topography, climate, insect thermal responses, the feeding ecology of the target stages, and their specific sensitivity to the pesticide. Over complex terrain, spruce budworm development can cover a wide range (Plate 1). The nonlinear nature of the influence of timing on B.t. efficacy is apparent from the output of Cooke’s model (Plate 1). Applications that are made too early (when the insect is still overwintering and has just started feeding) are the least efficacious as predicted in the mountains to the north and to the southeast of Quebec City. The most efficacious treatments occur

248

Phenology: An Integrative Environmental Science

when the insect is in instars 3 and 4. Treatments applied too late, past the fourth instar, are less efficacious as seen in the low-elevation area to the south and west of Trois-Rivières. This makes adequate timing of control operations using B.t. against spruce budworm quite critical to their success in protecting conifer foliage.

4.2

Gypsy Moth

Gypsy moth, Lymantria disparr (L.), is a lymantriid that was accidentally introduced in eastern North America in 1869, and has since spread gradually to the north, west and south (Liebhold et al. 1992). It is a periodic pest of deciduous trees, especially oaks (Quercus), maples (Acer ( ), birches (Betula ( ) and poplars ((Populus) (Montgommery 1990). Hypotheses about the determinants of its rate of spread and eventual range on the continent, especially to the north, have focused on egg mortality due to low winter temperatures, or on forest susceptibility (Sharov et al. 1999). Limitations that the insect will encounter in establishing a to the west and south of its current distribution are less well understood (e.g., Allen et al. 1993). A detailed model of gypsy moth phenology was assembled by Régnière and Sharov (1998) from components found in the literature for the various life stages. The egg hatch component of this model was replaced by the detailed model of gypsy moth egg diapause built by Gray et al. (2001). The resulting model can simulate the entire life cycle of the insect through successive generations in any climate. It was first used in the context of timing of an eradication program in British Columbia, Canada, in 2000 (Nealis et al. 2001). Régnière and Nealis (2002) used the model to determine the areas of southern British Columbia that were most likely to support the establishment of this exotic insect, on the basis of local climate. This analysis was based on whether or not the model predicted a biologically feasible life cycle for the insect in a given location under normal climatic conditions. If peak oviposition was predicted to occur no later than the end of October (a time when temperatures are too cold for eggs to enter diapause successfully) for 20 successive generations, gypsy moth was presumed to have the potential of establishing itself in that location. This analysis is mathematically identical to the G-function developed by Logan and Powell (2001). The same approach was used to produce a map of the probability of establishment of gypsy moth throughout the North American continent, north of Mexico. A series of 27,360 simulation points was located randomly across the continent. Twenty-generation model runs were made for each point using daily weather data generated from local normals (using the GIDS local gradient method described earlier). For each model run, the outcome

Chapter 4.2: Animal Life Cycle Models

249

was rated as 0 (seasonality did not remain viable for 20 generations) or 1 (seasonality did remain viable). Each run was replicated 30 times (30 stochastically different daily weather traces), and the average outcome for each point was used as an estimate of the probability of gypsy moth establishment at that location:

Pi

1 (1 n+2

n

∑

ij

)

(4)

j =1

where pij is the simulation outcome for point i and replicate j, n = 30 is the number of replicates. Equation (4) was formulated so that 0.02 ≤ Pi ≤ 0.98 to ensure that the logit transformation, used for linearization, did not lead to undefined results:

g ( P)

l [ ln

]

(5)

and was then interpolated spatially by universal kriging using elevation as a drift variable, over a digital elevation model of North America at 30 arc second (≈1 km) resolution. The resulting map (Plate 2) was back transformed to a probability scale by inverting equation (5).

4.3

Mountain Pine Beetle

Mountain pine beetle, Dendrochtonus ponderosae Hopkins, is a bark beetle (Coleoptera: Scolytidae) that has a large impact on ponderosa pine forests throughout western North America all the way from northern Mexico to central British Columbia (Logan and Powell 2001). There are two critical factors determining the ability of mountain pine beetle to overcome the defenses of its host tree, to kill it and successfully reproduce: adequate seasonal timing and simultaneous attack by large numbers of beetles. Thus, an adaptive seasonality for this insect implies that critical events in its life cycle must be timed adequately, and that the development of the population must not be so spread out (variable) that large numbers of adults are not available for synchronous attack within a given summer. In most insects, winter diapause serves to halt development during the cold season and maintain the population in the early stages of post-diapause until temperature warms up in the spring, a process that resets the biological calendar and synchronizes the entire population. In mountain pine beetle, however, there is no evidence of diapause in any life stage. In this species, seasonality seems entirely determined by the seasonal patterns of weather

250

Phenology: An Integrative Environmental Science

(Logan and Bentz 1999). Direct control of the insect’s seasonality by weather patterns is intriguing ecologically and evolutionarily. It would seem that very complex spatial and temporal patterns of abundance could result from the detail of regional, even local annual temperature fluctuations. A detailed phenology model is an ideal tool to address this type of question. Such a model is available for mountain pine beetle (Bentz et al. 1991, citations in Logan and Powell 2001). It is a distributed, nonlinear description of the thermal responses of all stages of the insect’s life cycle. Logan and Bentz (1999), Powell et al. (2000) and Logan and Powell (2001) studied the model’s behavior and provided insight as to how temperature regimes alone could synchronize an insect population without having recourse to diapause. There are three basic conditions for the seasonality of mountain pine beetle to be adaptive under a given annual temperature regime (as depicted by this phenology model): it must be univoltine, oviposition dates from generation to generation must converge to a nearconstant time of year, and this date must fall between biologically realistic bounds (early July to late August). We generated maps depicting the likelihood of mountain pine beetle establishment and thriving, on the basis of the probability of its achieving adaptive seasonality as defined by the three criteria above. We applied climate change scenarios, defined by deviations in monthly mean minimum and maximum air temperature, to normals used by BioSIM©. These scenarios were obtained from the Canadian Centre for Climate Modeling and Analysis (http://www.cccma.bc.ec.gc.ca/), gridded at finer spatial resolution (http://www.cics.uvic.ca/scenarios/index.cgi?Scenarios). Climate change normals (30-year averages) were calculated at 10-year intervals from the period 1971-2000 to 2041-2070. Actual normals were used for the first four decades. Simulations (20 successive generations) were run for 500 randomly located points in British Columbia and Alberta (Figure 3), and each simulation was replicated 30 times for each set of normals. The logistic transformation (equations 4 and 5) was used prior to applying universal kriging (with elevation as drift variable) to the probability of adaptive seasonality. The resulting series of probability maps covers a 120-year time span (Plate 3). The model predicts a gradual restriction of the insect’s most suitable range in British Columbia and Alberta towards more northern locations and to higher elevations as the climate of the region warms, disrupting the univoltinism requirement for adaptive seasonality in mountain pine beetle. Currently, mountain pine beetle thrives in south and central British Columbia. It has also occasionally caused damage in the Cypress Hills area of southeastern Alberta. Prairies and the high elevations of the Canadian Rocky Mountains have confined it there. The northern part of Alberta is

Chapter 4.2: Animal Life Cycle Models

251

forested by jack pine, Pinus banksiana, a species that is only now coming in contact with mountain pine beetle in the mountain passes between central BC and Alberta (Logan and Powell 2001, personal communication, A. Carroll, Canadian Forest Service, Victoria, BC). However, it seems quite likely that mountain pine beetle can spill over the natural barrier of the high mountains, given the increasing suitability of northern latitude and higher elevations as global warming proceeds.

5.

CONCLUSIONS

The development of phenology-modeling methodologies in the field of animal ecology has tended to precede that of computing technology in the past 30 years, since the introduction of nonlinear and distributed models of poikilotherm thermal responses. These models are becoming increasingly sophisticated, detailed and accurate, and the study of their behavior is teaching us about the evolution of seasonality and the effects of temperature on the distribution and population stability of poikilotherms. It is also becoming increasingly feasible to investigate the outcomes of phenological processes through models that make predictions over large, climatically and topographically complex areas. These technologies allow us to use landscape-wide phenological projections in the conduct of area-wide Integrated Pest Management activities. They also provide us with the ability to study and better understand the ecology and distribution of indigenous species based on comparison of observations with model predictions. We can also use these tools to analyze the probable reactions of these indigenous species to changing environments, most importantly climate change, but also changes in the distribution of host plants resulting from human activity. Finally, these tools can be used to predict the probable distribution and thriving of invasive species, such as gypsy moth, as soon as we gain sufficiently detailed knowledge of their thermal responses. Using Geographical Information Systems technology, it is also becoming relatively simple to merge the outcome of our detailed understanding of developmental processes with other geographically critical information such as the distribution of susceptible plants, soils, water, as well as insect or disease survey data. This convergence of information constitutes the basis for investigation of more complex ecological issues that are always related to seasonality and phenology, but not always directly or simply.

252

Phenology: An Integrative Environmental Science

REFERENCES CITED Allen, J. C., J. L. Foltz, W. N. Dixon, A. M. Liebhold, J. J. Colbert, J. Régnière, D. R. Gray, J. W. Wilder, and I. Christie, Will the gypsy moth become a pest in Florida?, Florida Entomologist, 76, 102-113, 1993. Bentz, B. J., J. A. Logan, and G. D. Amman, Temperature-dependent development of the mountain pine beetle (Coleoptera: Scolytidae) and simulation of its phenology, Can. Entomologist, 123, 1083-1094, 1991. Bolstad, P. V., L. Swift, F. Collins, and J. Régnière, Measured and predicted air temperatures at basin to regional scales in the southern Appalachian mountains, Agricult. Forest Meteorol., 91, 161-176, 1998. Boutin, S., and D. Hebert, Landscape ecology and forest management: developing an effective partnership, Ecol. Appl., 12, 390-397, 2002. Chen, J., S. C. Saunders, T. R. Crow, R. J. Naiman, K. D. Brosofske, G. D. Mroz, B. L. Brookshire, and J. F. Franklin, Microclimate in forest ecosystem and landscape ecology: variations in local climate can be used to monitor and compare the effects of different management regimes, Bioscience, 49, 288-297, 1999. Cooke, B. J., and J. Régnière, An object-oriented, process-based stochastic simulation model of Bacillus thuringiensis efficacy against spruce budworm, Choristoneura fumiferana (Lepidoptera: Tortricidae), Int. J. Pest Management, 42, 291-306, 1996. Curry, G. L., R. M. Feldman, and K. C. Smith, A stochastic model of temperature-dependent population, Theoretical Population Biology, 13, 197-213, 1978. Forrester, J. W., Industrial dynamics, MIT Press, Cambridge, MA 464 pp., 1961. Gignac, M., Comparaison de la régression spatiale et du krigeage avec dérive pour interpoler des extrants de modèles de simulation de développement d'insectes au Québec en fonction de l'échelle, de la topographie et de l'influence maritime, M.Sc. Thesis, Faculty of Forestry and Geomatics, Université Laval, Sainte-Foy, Quebec, Canada, 2000. Gray, D. R., F. W. Ravlin, and J. A. Braine, Diapause in the gypsy moth: a model of inhibition and development, J. Insect Physiology, 47, 173-184, 2001. Haila, Y., A conceptual genealogy of fragmentation research: from island biogeography to landscape ecology, Ecol. Appl., 12, 321-334, 2002. Hutchinson, M. F., The application of thin-plate smoothing splines to continent-wide data assimilation, in Data assimilation systems, edited by J. D. Jasper, pp. 104-113, BMRC Res. Rep. No. 27, Melbourne Bureau of Meteorology, 1991. Hutchinson, M. F., Stochastic space-time weather models from ground-based data, Agricult. Forest Meteorol., 73, 237-264, 1995. Isaaks, E. H., and R. M. Srivastava, An introduction to applied geostatistics, Oxford University Press, New York, 561 pp., 1989. Janisch, E., The influence of temperature on the life-history of insects, Trans. Entomolological Society of London, 80, 137-168, 1932. Liebhold. A. M., J. A. Halverson, and G. A. Elmes, Gypsy moth invasion in North America: a quantitative analysis, J. Biogeography, 19, 513-520, 1992. Logan, J. A., Toward an expert system for development of pest simulation models, Environ. Entomology, 17, 359-376, 1988. Logan, J. A., and B. J. Bentz, Model analysis of mountain pine beetle (Coleoptera: Scolytidae) seasonality, Environ. Entomology, 28, 924-934, 1999. Logan, J. A., and J. A. Powell, Ghost forests, global warming, and the mountain pine beetle, Amer. Entomologist, 47, 160-173, 2001.

Chapter 4.2: Animal Life Cycle Models

253

Logan, J. A., D. J. Wolkind, S. C. Hoyt, and L. K. Tanigoshi, An analytic model for description of temperature dependent rate phenomena in arthropods, Environ. Entomology, 5, 1133-1140, 1976. Manetsch, T. J., Time varying distributed delays and their use in aggregative models of large systems, IEEE Transactions in Systems, Man, and Cybernetics, 6, 547-553, 1976. McGarigal, K., and S. A. Cushman, Comparative evaluation of experimental approaches to the study of habitat fragmentation effects, Ecol. Appl., 12, 335-345, 2002. Montgommery, M. E., Variation in the suitability of tree species for the gypsy moth, in Proceedings U.S. Department of Agriculture Interagency Gypsy Moth Research Review, edited by K.W Gottschalk, M. J. Tivery, and S. I. Smith, pp. 1-13, USDA Forest Service General Technical Report NE 146, 1990. Nalder, I. A., and R.W. Wein, Spatial interpolation of climatic normals: test of a new method in the Canadian boreal forest, Agricult. and Forest Meteorol., 92, 211-225, 1998. Nealis, V. G., J. Régnière, and D. R. Gray, Modeling seasonal development of Gypsy moth in a novel environment for decision-support of an eradication program, in Integrated management and dynamics of forest defoliating insects, edited by A. M. Liebhold, M. L. McManus, I. S. Otvos, and S. L. C. Fosbroke, pp. 124-132, USDA Forest Service General Technical Report NE-277, 2001. Powell, J. A., J. Jenkins, J. A. Logan, and B. J. Bentz, Seasonal temperature alone can synchronize life cycles, Bull. Mathematical Biology, 62, 977-998, 2000. Price, D. T., D. W. McKenney, I. A. Nalder, M. F. Hutchinson, and J. L. Kesteven, A comparison of two statistical methods for spatial interpolation of Canadian monthly mean climate data, Agricult. Forest Meteorol., 101, 81-94, 2000. Racsko, P., L. Szeidl, and M. Semonov, A serial approach to local stochastic weather models, Ecological Modelling, 57, 27-41, 1991. Régnière, J., A method of describing and using variability in development rates for the simulation of insect phenology, Can. Entomologist, 116, 1367-1376, 1984. Régnière, J., Generalized approach to landscape-wide seasonal forecasting with temperaturedriven simulation models, Environ. Entomology, 25, 869-881, 1996. Régnière, J., and P. Bolstad, Statistical simulation of daily air temperature patterns in eastern North America to forecast seasonal events in insect pest management, Environ. Entomology, 23, 1368-1380, 1994. Régnière, J., and B. Cooke, Validation of a process-oriented model of Bacillus thuringiensis variety kurstaki efficacy against spruce budworm (Lepidoptera: Tortricidae), Environ. Entomology, 27, 801-811, 1998. Régnière, J., and V. Nealis, Modelling seasonality of gypsy moth, Lymantria dispar (Lepidoptera: Lymantriidae), to evaluate probability of its persistence in novel environments, Can. Entomologist, 134, 805-824, 2002. Régnière, J., and A. Sharov, Phenology of Lymantria disparr (Lepidoptera: Lymantriidae), male flight and the effect of moth dispersal in heterogeneous landscapes, Int. J. Biometeorol., 41, 161-168, 1998. Régnière, J., and A. Sharov, Simulating temperature-dependent ecological processes at the sub-continental scale: male gypsy moth flight phenology as an example, Int. J. Biometeorol., 42, 146-152, 1999. Régnière, J., B. Cooke, and V. Bergeron, BioSIM: a computer-based decision support tool for seasonal planning of pest management activities, User's Manual, Canadian Forest Service Information Report LAU-X-116, 35 pp., 1995. Richardson, C. W., Stochastic simulation of daily precipitation, temperature and solar radiation, Water Resources Res., 17, 182-190, 1981.

254

Phenology: An Integrative Environmental Science

Richardson, C. W., and D. A. Wright, WGEN: A model for generating daily weather variables, US Department of Agriculture, Washington, DC, Agricultural Research Service 8 pp., 1984. Royama, T., Population dynamics of the spruce budworm Choristoneura fumiferana, Ecol. Monographs, 54, 429-462, 1984. Ruel, J. J., and M. P. Ayres, Jensen’s inequality predicts effects of environmental variation, TREE, 14, 361-366, 1999. Russo, J. M., A. M. Liebhold, and J. G. W. Kelley, Mesoscale weather data as input to a gypsy moth (Lepidoptera: Lymantriidae) phenology model, J. Economic Entomology, 86, 838-844, 1993. Ryszkowski, L., Landscape ecology in agroecosystems management: Advances in agroecology, CRC Press, Boca Raton, Fla., 366 pp., 2002. Schaub, L. P., F. W. Ravlin, D. R. Gray, and J. A. Logan, A landscape framework to predict phenological events for gypsy moth (Lepidoptera: Lymantriidae) management programs, Environ. Entomology, 24, 10-18, 1995. Sharov, A. A., B. C. Pijanowski, A. M. Liebhold, and S. H. Gage, What affects the rate of gypsy moth (Lepidoptera: Lymantriidae) spread: winter temperature or forest susceptibility?, Agricult. Forest Entomology, 1, 37-45, 1999. Sharpe, P. J. H., and D. W. DeMichele, Reaction kinetics of poikilotherm development, J. Theoretical Biology, 64, 649-670, 1977. Stinner, R. E., G. D. Butler Jr., J. S. Bacheler, and C. Tuttle, Simulation of temperaturedependent development in population dynamics models, Can. Entomologist, 107, 11671174, 1975. Vansickle, J., Attrition in distributed delay models, IEEE Transactions in Systems, Man and Cybernetics, 7, 635-638, 1977. Wagner, T. L., H.-I. Wu, P. J. H. Sharpe, R. M. Schoolfield, and R. N. Coulson, Modeling insect development rates: a literature review and application of a biophysical model, Annals Entomological Soc. Amer., 77, 208-225, 1984. Wang, J. Y., A critique of the heat unit approach to plant response studies, Ecology, 41, 785790, 1960. Whiteman, C. D., Mountain Meteorology: Fundamentals and applications, Oxford University Press, New York, 355 pp., 2000. Wilks, D. S., Simultaneous stochastic simulation of daily precipitation, temperature and solar radiation at multiple sites in complex terrain, Agricult. Forest Meteorol., 96, 85-101, 1999. Worner, S. P., Performance of phenological models under variable temperature regimes: consequences of the Kaufmann or rate summation effect, Environ. Entomology, 21, 689699, 1992.

Chapter 4.3 PHENOLOGICAL VARIATION OF FOREST TREES Robert Brügger1, Matthias Dobbertin2, and Norbert Kräuchi2 1

PHENOTOP, Institute of Geography of the University of Berne, Berne, Switzerland; 2WSL Swiss Federal Institute for Forest, Snow, and Landscape Research, Forest Ecosystems and Ecological Risks Division, Birmensdorf, Switzerland

Key words:

1.

Beech, Growth, Variation, Development stages, PEI

INTRODUCTION

Plant phenological observations are typically made on a few (undefined number) individuals from a range of species, chosen to represent the population of this species or the collective behavior of crop fields, meadows, and forests ecosystems. Observations of individuals from the same plant species are rare. Nevertheless, the focus on phenological events, and in particular their power to express the impact of climate change, has increased over the past decade. New questions concerning biological life cycles, seasonal growth, winter dormancy, and their manifestation in the phenology of the plant have emerged. Against this backdrop, observations of multiple individuals from single plant species, growing in the same local area, are poised to play an important role in answering these questions.

Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 255-267 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

256

Phenology: An Integrative Environmental Science

2.

PHENOPHASES, PHENOLOGICAL DEVELOPMENT STAGES, AND PEI

Phenological development is illustrated by the Phenological Emergence Index (PEI), which is derived from observation of multiple phenological development stages. The relation between phenological development stages, phenophases, and the PEI is shown in Figure 1.

2.1

Phenological Development Stages

A phenological development stage is a visually perceptible event concerning a specific morphological part of the plant during its growth continuum or seasonal development, which is given an exact description (e.g., “green tip of the new leaf is visible in the opening bud”).

2.2

Phenophase (Phytophenology)

Focusing on the whole plant, each phenological development stage can be observed with a specific frequency. This frequency can be a defined number (e.g., “the green tips of the first three leafs”), or a defined quota (e.g., “the green tips of 50% of all leaves”). For each phenophase, both the specific phenological development stage and the number (or quota, respectively) have to be defined.

2.3

The Phenological Emergence Index (PEI)

The Phenological Emergence Index (PEI) is a dynamic variable that describes specific time windows in the seasonal growth of a single plant. It is derived from the observation of phenological development stages. An index is evaluated for both the leaf unfolding and leaf coloring phenological processes. Figure 1 represents the timeline of the phenological development stages of a spring event. The stages (in this example three) are labeled E1 to E3. The PEI is calculated from the data of these three stages according to Equation 1.

PEI(t) = Ei k

k i=1

i*Ei (t)

/k

= %-quota of phenological development stage i = total number of phenological development stages used

(1)

Chapter 4.3: Phenological Variation of Forest Trees

3.

257

PHENOLOGICAL DEVELOPMENT FROM SPRING TO AUTUMN

Between 1990 and 1999, 34 Beech trees (Fagus ( silvatica L.) and (initially) 67 Spruces ((Picea abies L. Karst.) were observed at different sites in the Canton of Berne, Switzerland. The observations were taken based on miscellaneous field observation techniques (Flekinger 1965; Innes et al. 1994). Investigations were made of seasonal progress in leaf (needle)

Figure 4.3-1. Phenological Emergence Index (PEI). The figure shows the cumulative frequency of the three phenological development stages of leaf unfolding, E1, E2 and E3, the PEI and the phenophase “leafing of Beech”(E350) of a single Beech ((Fagus silvatica L.) at Eymatt (altitude 530 m, near Berne) in 1995. The stages are defined in the text. E350 represents the phenophase “50% of all leaves have reached E3 (are unfolded)”. In this example the phenophase was observed on 1 May.

unfolding and leaf coloring, with an irregular sampling frequency. In the case of leaf unfolding, the selected sampling frequency was once in three days, and once a week for leaf coloring. Both processes were recorded over the time span of several phenological development stages. At each tree, each phenological stage was quantitatively estimated according to classes at 10% intervals. Because of their special biological interest, the initial occurrence of the phenological stages was reported in more detail (Brügger 1998). In 1998, a new program of forest observations was implemented, which covers single-tree observations as well as group observations (Vassella and Brügger 2001).

258

3.1

Phenology: An Integrative Environmental Science

Selected Phenological Development Stages of Beech

The phenological development of the Beech tree is divided into two processes: leaf unfolding and leaf coloring. Leaf unfolding occurs in three development stages (see Figure 2) and leaf coloring in four stages. The four stages of leaf coloring of Fagus silvatica are: 1. The leaf blade, or part of it, is light green (i.e., no longer the dark green it was in summer) 2. The leaf blade, or part of it, is yellow. 3. The leaf blade, or part of it, is orange, red, or brown 4. Leaf abscising. The quota of leaves shed is obtained from the difference between the current crown transparency and the condition of summer crown transparency.

Phenological development stages of leaf unfolding for Fagus silvatica L.. Stage 1 (A): The bud is elongated and the first green tip of the leaf is visible. Stage 2 (B): The bud is wide open and the leaf blade is clearly visible, but not in its entirety. The leaf base and the stalk are still invisible. Stage 3 (C): The entire leaf blade and the leaf stalk are visible.

3.2

The Phenological Development of a Beech Tree

We can use the PEI timeline to follow the development of phenological events. The observed time-segment of development in spring is about four weeks, and approximately fourteen weeks in autumn. In spring, the stages are consecutively visible on the tree and last for a short period only. However, in autumn the different stages are visible simultaneously over a

Chapter 4.3: Phenological Variation of Forest Trees

259

period of several weeks, even if the quota of each stage rises and falls several times. While many yellow or brown leaves may be observed in the tree foliage on one day, all these colored leaves will have disappeared a few days later, i.e. they have been shed and the foliage looks green again (see Figure 3). The two patterns reflect different time scales, but also the different physiological background of the development processes. Temperature is the main determinant for bud burst (e.g., Ducousso et al. 1996; Linkosalo 2000). Leaf coloring and shedding are influenced by temperature, photoperiod, precipitation, and drought or wind, although “the relationship between simple meteorological factors and leaf/leaf fall is faint” (Menzel 2002).

Figure 4.3-3. Development of autumn coloring. The quota of the phenological development stages of leaf coloring and the PEI of a Beech ((Fagus silvatica L.) at Leissigen (altitude 685 m) in 1995. The different shadings represent the different quota of each stage to the PEI in the timeline. The stages are defined in the text of this chapter.

4.

VARIABILITY OF PHENOLOGICAL DEVELOPMENT AT ONE SITE

The phenological development of the trees at one site is within the specific range of the first and last tree. As mentioned by Waggoner (1974), this range grows with the number of observed trees. Nevertheless, it is

260

Phenology: An Integrative Environmental Science

limited by climatic determinants, e.g. frost, local soil, or competition conditions, and the genetic plasticity of each species. Two ways to display this variability are presented in Figure 4 (with the PEI) and in Figure 5 (with dates of phenophases, calculated based on observation of phenological development stages). Within the collective, the state of phenological development differs from about one to nineteen days for leaf unfolding, and about six to thirty-six days for leaf coloring. The disparities between the levels of development of individual trees within a collective are often greater than those between collectives. In spring, however, the chronological order of phenological development is quite regular. It runs along an imaginary transect of “temperature and solar radiation gradient”(altitude, exposure and slope) and

Figure 4.3-4. Autumn-PEI of 10 Beech trees (Fagus ( silvatica L.) in a mixed Beech stand. Variations in the phenological development at one site in Switzerland during two typical years. The autumnal development is normal in 1998 (light gray) and retarded in 1997 (dark gray). The early-developing trees are nevertheless more advanced in 1997 than the latedeveloping trees in 1998, at least until mid-October.

Chapter 4.3: Phenological Variation of Forest Trees

261

Table 4.3-1. Variation of phenophases of Beech trees ((Fagus silvatica L.) in mixed Beech stands. Differences (in days) of the onset-dates between the first and the last tree, leaf unfolding (mean values 1991-1999) and leaf coloring (mean values 1990-1999). Site (n = num. of trees)

Altitude (m.a.s.l.)

Exposure

Leaf unfolding [Days]

Leaf coloring [Days]

Leissigen n = 8 Magglingen n = 10 Eymatt n = 11 Vingelzberg n = 5

685 925 530 630

N SE NW SE

14.3 7.1 9.4 8.5

20.6 18.6 16.9 20.9

Figure 4.3-5. Variation of two phenophases of Beech trees ((Fagus silvatica L.) in mixed Beech stands. The phenophases “E350 leaf unfolding” and “E250 leaf coloring” at four sites in the Canton of Berne, Switzerland. The Julian date is interpolated (linear) from irregular time series, integrating the observation of the specific phenological development stages E3 in spring, E2 in autumn, respectively. E350 in spring is the date when 50% of the leaves reach stage 3 (“the entire leaf and the leaf stalk are visible”, Figure A) and in autumn, E250 is the date when 50% of the leaf area of the whole tree reach stage 2 (“the leaf blade, or part of it, is yellow”, Figure B). The earliest date of leaf unfolding is at Vingelzberg (1992) on April 22 (Julian 113), the latest date is Leissigen (1991) on May 19 (Julian 139). In autumn, the earliest date of leaf coloring is Vingelzberg (1999) on September 10 (Julian 254), and the latest date is Magglingen (1991) on November 8 (Julian 312). Mean values of the differences in days within the sites are shown in Table 1.

262

Phenology: An Integrative Environmental Science

once more reflects the close relationship between temperature and bud burst. This imaginary transect hardly appears in the leaf coloring data.

4.1

Varying Phenological Development in a Mixed Beech Stand

Investigations of Beech tree collectives show regularities in the pattern of phenological development. Generally, it is the same trees that open buds first (or last) in spring, or shed their leaves early (or late) in autumn. Hence the trees were able to be classified at each site according to their

25 20

25 20

0

10 20 30 days

0

0

.8 .6 .0

.2

.4

.0 .2 .4 .6 0 4 8 10 20 30 days

Rfrc [FU/day]

.8 15 12 0

15

20 0 .8 6 .6 4 .0 0

.2

2

.4

PEI

early middle late Rfrc

Sfrc [FU]

30

35 1 100

1996

30

32 2 100 24

early middle late Sfrc

28

20 40 60 80 100

1994

0

PEI

site: Vingelzberg

10 20 days

30

Figure 4.3-6. PEI and Forcing Status in Spring. The phenological development of different development classes of Beech trees (Fagus ( silvatica L., dotted lines) compared with a temperature-driven forcing variable (straight line) in spring. The upper three figures (A) show the cumulative frequencies and the lower three figures (B) the daily rates. The day count-up starts with the event of the first observed buds in the phenological development stage 1 (see Figure 2). In 1994, an early warm period induces some growing effects in the class of “early” Beeches. After that, a period with low rates interrupts or halts the growth of all classes for about ten days. Following this period, the development rate of the two classes “early” and “middle” increases with the first warm spell, and the rate of the “late” class follows with the second period of high forcing rates. In 1995, the peaks of forcing and development rates often coincide. Again, the “early” class reacts first. In 1996, a high forcing input occurs for a period of about two weeks, during which the buds of all development classes open virtually at the same time.

Chapter 4.3: Phenological Variation of Forest Trees

263

phenological growth pattern as “early”, “middle”, “late”, or “indifferent”, both in terms of leaf unfolding and leaf coloring (the “indifferent” trees don’t show a clear pattern and are not presented here, methods in Brügger 1998).

PEI -

-

Sfrc [FU]

site: Eymatt

0.8 0.4

Rfrc [FU/day]

-1

0.0

-3

-2

PEI

0-

-

-

Sfrc Rfrc

31.7.

30.8.

29.9.

29.10.

28.11.

date [day.month.] Figure 4.3-7. PEI and Forcing Status in Autumn. The phenological development of different development classes of Beech trees (Fagus ( silvatica L., dotted lines) compared with a temperature-driven forcing variable (straight line) in autumn. The upper figure (A) shows the cumulative frequencies (the phenological development is presented as a negative development) and the lower figure (B) the rates. Declining forcing rates at the end of August indicate the phenological development in autumn. The development itself (the coloring of the leaves, i.e. the chemical degradation of chlorophyll and other pigments in the leaves) dictates higher forcing rates (period with high Rfrc-rates in September). The class of “early” Beech trees react first, followed by the classes “middle” and “late”. More than one development wave was observed.

Much of the variation within a collective can be explained by this different behavior. The phenological development rate follows the forcing

264

Phenology: An Integrative Environmental Science

rate with a specific time delay, but the “early” Beech trees “react” to earlier forcing peaks, and “late” trees react to later peaks. This different behavior reflects different physiological exploitation of temperature as a growingforce. In Figure 6, three different years are presented for spring and Figure 7 shows one example in autumn (for details see Brügger 1998).

5.

PHENOLOGY IN FOREST GROWTH STUDIES

Most transient changes in the structure of forest ecosystems, such as the decline of certain tree species, are driven by a combination of climatic and anthropogenic changes, and are modified by local, biological interactions acting on temporal scales ranging from months to centuries. Because mankind is unable to sense slow changes directly or to interpret cause-effect relations for these changes, processes acting over decades are hidden in the “invisible present” (Magnuson et al. 1983). Recognition of changes in ecosystem functioning, and development of assessment strategies to deal with consequences due to these changes, are only possible if enough data are available, and interrelationships between different ecosystem parameters are sufficiently understood. Long-term research helps to identify the events of the invisible present and allows identification of the potential ecological risks. By identifying the ecological consequences of human actions, the

Figure 4.3-8. Temperature and stem radius. Hourly temperature and circumference data to detect winter shrinkage and on-set of growth of Abies alba Miller (silver fir).

Chapter 4.3: Phenological Variation of Forest Trees

265

Swiss Long-term Forest Ecosystem Research network LWF (Thimonier et al. 2001) provides an essential foundation for assessing ecological risks. In this context, phenological processes play an important role. An improved insight in potential phenological changes may contribute to a better understanding of abiotic ecosystem processes like climate-induced changes in nutrients cycles, water fluxes, and carbon sequestration. During the late 1970s and beginning of the 1980s, several cases of severe forest decline were reported in selected areas of Europe. Air pollution was largely incriminated originally, which triggered the setting up in 1985 of the International Co-operative Program on the Assessment and Monitoring of Air Pollution Effects on Forests (ICP-Forests), under the Convention on Long-range Transboundary Air Pollution (CLTRAP) of the UN/ECE. Switzerland is one of 36 countries participating in this so-called level II program. Phenological data play an important role in forest growth studies at the Swiss Level II plots. Without automatic monitoring of stem circumference/radius, shoot elongation, and destructive sampling (cambium marking, micro cores) it is not possible to determine the start and end of the growing season (see Figure 8 and Figure 9). Even with automatic measurements, it is often very difficult to determine the onset and offset of growth (Dobbertin 2001). Phenological data may help to more easily determine the vegetation periods (Figure 10), however for conifers it will be difficult to determine the end of the growing seasons.

Figure 4.3-9. Stem radius changes recorded on an hourly basis: Silver fir ((Abies alba Miller) and Norway spruce ((Picea abies L. Karst.).

266

Phenology: An Integrative Environmental Science

Level II site Bettlachstock Switzerland PAR-ratio openland/inside forest

60

Year 2000 50 40 30

Phenology: 50% foliage

Phenology: 50% discolored

20 10 0 3/5

4/2

4/30 5/28 6/25 7/23 8/20 9/17 10/15 11/12 12/10

month/day Figure 4.3-10. Vegetation Period. Using PAR-measurements and phenological observations to determine the vegetation period in a mixed Beech Stand in Switzerland (LWF-site Bettlachstock).

6.

CONCLUSIONS

Observations of Beech trees in Switzerland show relevant differences in the phenology of the single individual. Within the collective of Beech trees, the state of phenological development differs from about one to nineteen days for leaf unfolding, and about six to thirty-six days for leaf coloring. Data on the occurrence of specific phenophases and their absolute differences within a collective are important aids to estimating the potential of climate damage in threshold situations (e.g., frost damage in spring), or the potential of degradation or drift within a population in relation to climate change. Mapping the observed phenological development of a single individual onto daily weather indices also provides a closer understanding of the different indications of natural growth processes. Phenology has established itself as a valuable instrument for monitoring the impact of climate change. Since such data also allows observation of natural development and variations, it is also an extremely useful way of obtaining additional biological information.

Chapter 4.3: Phenological Variation of Forest Trees

267

ACKNOWLEDGEMENTS This project was supported by the Swiss Federal Agency of Environment, Forest and Landscape (SAEFL). Thanks are also due to F. Jeanneret, A. Vassella, R. Volz, H. Wanner, S. Braun, M. Sieber, C. Defila, and R. Häsler.

REFERENCES CITED Brügger, R., Die phänologische Entwicklung von Buche und Fichte, Beobachtungen, Variabilität, Darstellung und deren Nachvollzug in einem Modell, Geographica Bernensia G49, 186 pp., 1998. Dobbertin, M., Hourly recorded changes in tree stem diameter - Can we distinguish stem growth from contraction and expansion of the bark?, in L'arbre 2000 The Tree, papers presented at the 4th International Symposium on The Tree, edited by M. Labrecque, pp. 263-267, 21-25 August 2000, Montreal, Canada, Isabelle Quentin, 2001. Ducousso, A., J. P. Guyon, and A. Kremer, Latitudinal and altitudinal variation of bud burst in western populations of sessile oak (Quercus petraea (Matt.) Liebl.), Ann. Sci. For., 53, 775-782, 1996. Fleckinger, J., Phénologie. Phénologie et Arboriculture Fruitière, Bon Jardinier 1(2, part. C), 362-372, 1965. Innes, J. L., J. Böhm, J. B. Bucher, M. Dobbertin, E. Jansen, P. Kull, A. Rigling, L. Walthert, S. Zimmermann, and S. Sanasilva-Bericht, Der Zustand des Schweizer Waldes, Ber. Eidgenöss. Forsch. anst. Wald Schnee Landsch., 339, 60 pp., 1994. Linkosalo, T., Analyses of the spring phenology of boreal trees and its response to climate change, Univ. Helsinki, Dep. Forest Ecology Publications, 22, 1-55, 2000. Magnuson J. J., C. J. Bowser, A. L. Beckel, The invisible present: Long-term ecological research on lakes, L&S Magazine, 3, University of Wisconsin-Madison, Fall, 1983. Menzel, A., Phenology: Its importance to the global change community, Climatic Change, 54, 379-385, 2002. Thimonier, A., M. Schmitt, P. Cherubini, and N. Kräuchi, Monitoring the Swiss Forests – building a research platform, in Monitoraggio ambientale metodologie ed applicazioni, edited by T. Andofillo and V. Carraro, pp. 121-132, Atti del XXXVIII Corso di Cultura in Ecologia, 2001. Vassella A., and R. Brügger, Impulsprojekt für das phänologische Monitoring im Wald und für den Einbezug in die Langfristige Waldökosystemforschung (LWF), SAEFL Report Ökol07/97, 62 pp., Bern, 2001. Waggoner, P. E., Modeling Seasonality, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 301-327, Springer, New York, 1974.

Chapter 4.4 PHENOLOGICAL GROWTH STAGES Mono- and Dicotyledonous Plants Uwe Meier Federal Biological Research Center for Agriculture and Forestry, Braunschweig, Germany

Key words:

1.

Plant growth stages, BBCH scale, Pome fruit, Agriculture, Standardization

HISTORY OF THE DESCRIPTION OF PLANT GROWTH STAGES

Numerous authors have published descriptive development-stage scales over the past 70 years, covering various plant species. Troitzki (1925) examined connections between occurrence/control of the apple blossom weevil (Anthonomus ( pomorum) and phenological development of flower buds. He divided apple bud formation into three stages and twelve phases. This initial development scale for woody fruit plants was also used by Klemm (1937) and Soenen (1951, 1952) in their papers about plant protection, and Fleckinger (1948) to describe phenological development stages of apples and pears. He assigned capital letters from A-I to the main stages, and differentiated them by numbers from 1 to 4. This classification was widely used in science and industry until 1994. Based on Feekes’ (1941) scale, Large (1954) published the first numeric code for cereal. Zadoks et al. (1974) developed an adjusted and refined scale for cereals and rice that is still distributed and applied internationally. By the end of the 1980s, increasing interconnections in international crop research, and electronic data processing of results, demanded an international system with a standardized description and coding of plant development stages. Such a system was also needed to increase Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 269-283 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

270

Phenology: An Integrative Environmental Science

communication among agricultural sciences, other disciplines like botany, agricultural insurance agencies, and phenological observations networks. In response, two different working groups in Germany (West Germany and East Germany, divided at that time) simultaneously developed new schemes. The Federal Biological Research Center for Agriculture and Forestry (BBA) published leaflet series Number 27 (Meier 1985) and members of the Academy of Agricultural Sciences (Buhtz et al. 1990) produced a coordinated decimal code (KDC) for numerous crops. Yet, the goal of a uniform coding was still not realized. The second group consisted of staff members from four companies that have been conducting agricultural field research for decades. Publication of BBCH codes for some specific crops (Bleiholder et al. 1989) generated considerable interest and they were quickly adopted internationally, for example in Brazil (Bleiholder et al. 1991a), Spain (Bleiholder et al. 1991b), and England (Lancashire et al. 1991). Hack et al. (1992) first published the principles of a truly uniform scale, and description of development stages for the most important specific crops was completed with the book “BBCH-Monograph, Growth Stages of Plants” (Meier 1997), at least for now. The authors1 formed a working group consisting of government authorities, science, and industry, which subsequently published 27 descriptions of these development stages of crops and wild plants in four languages (English, German, Spanish, French). This documentation contributed considerably to the worldwide distribution of the BBCH system in plant research and administration, and even solved many interdisciplinary communication problems.

2.

THE BBCH SCALE

The BBCH scale is a system for uniform coding of phenologically similar growth stages of all mono- and dicotyledonous plant species. It was developed through cooperation among the German Federal Biological Research Center for Agriculture and Forestry (BBA), the German Federal Office of Plant Varieties (BSA), the German Agrochemical Association (IVA)1, the agrochemical industry, the Ministry of Agriculture of SchleswigHolstein1 and the Institute for Vegetables and Ornamentals in Grossbeeren/Erfurt, Germany (IGZ)1. The decimal code is based on the earlier well-known cereal code developed by Zadoks et al. (1974), in order to avoid major departures from this widely used phenological key. The abbreviation BBCH derives from Biologische Bundesanstalt, Bundessortenamt, and CHemical industry.

Chapter 4.4: Phenological Growth Stages

2.1

271

The Basic Principles of the Scale

– The general scale forms the framework within which the individual scales are developed. It can also be used for plant species for which no special scale is currently available. – Similar phenological stages of each plant species are given the same code. – For each code, a description is given, and for some important stages, drawings are included. – For the description of the phenological development stages, clear and easily recognized (external) morphological characteristics are used. – Except where otherwise stated, only the development of the main stem is taken into consideration. – The growth stages refer to representative individual plants within the crop stand. Crop stand characteristics may also be considered. – Relative values relating to species (and/or variety) specific ultimate sizes are used for the indication of sizes. – The secondary growth stages 0 to 8 correspond to the respective ordinal numbers or percentage values. For example, stage 3 could represent 3rd true leaf, 3rd tiller, 3rd node, 30% of the final length or size typical of the species, or 30% of the flowers open. – Post harvest or storage treatment is coded 99. – Seed treatment before planting is coded 00.

2.2

Organization of the Scale

The entire plant developmental cycle is subdivided into ten (clearly recognizable and distinguishable) longer-lasting developmental phases. These principal growth stages are described using numbers from 0 to 9 in ascending order. The principal growth stages are described in Table 1. In order to accommodate many different plant species, shifts in the course of development are allowed, and certain stages may even be omitted. The principal growth stages need not proceed in the strict sequence defined by the ascending order of the figures, but can occasionally also proceed in parallel. If two or more principal growth stages proceed in parallel, both can be indicated by using a diagonal stroke (example 16/22). If only one stage is to be indicated, either the more advanced growth stage, or the principal growth stage of particular interest must be chosen, depending upon plant species. The principal growth stages alone are not sufficient to define exactly application or evaluation dates, since they always describe time spans in the course of the plant development.

272

Phenology: An Integrative Environmental Science

Table 4.4-1. Principal growth stages Stage Description 0 Germination / sprouting / bud development 1 Leaf development (main shoot) 2 Formation of side shoots / tillering 3 Stem elongation or rosette growth / shoot development (main shoot) 4 Development of harvestable vegetative plant parts or vegetatively propagated organs / booting (main shoot) 5 Inflorescence emergence (main shoot) / heading 6 Flowering (main shoot) 7 Development of fruit 8 Ripening or maturity of fruit and seed 9 Senescence, beginning of dormancy

Secondary stages are used if points of time or steps in the plant development must be indicated more precisely. In contrast to the principal growth stages, these are defined as short developmental steps characteristic of the respective plant species, which are passed successively during the respective principal growth stage. They are also coded by using the figures 0 to 9. The combination of figures for the principal and the secondary stages, results in a two-digit code scale that allows precise definition of all phenological growth stages for the majority of plant species. Only in the case of some plant species (e.g., cucumber, onion, potato, tomato) is further subdivision necessary within a principal growth stage beyond that possible using the secondary stages from 0 to 9. For these cases, a three-digit scale is presented alongside the two-digit one. This involves the inclusion of the socalled “mesostage” between the principal and the secondary stage, which provides a further subdivision with figures 0 and 1 describing the development on the main stem, and figures 2 to 9 that of the side shoots (2nd to 9th order). In this way up to 19 leaves can be counted on the main stem, or the branching can be described. The BBCH scales allow comparison of individual codes only within one principal growth stage, since an arithmetically greater code indicates a plant at a later growth stage. Sorting codes into numerical order therefore allows a listing in order of the stage of plant development. The time span of certain developmental phases of a plant can be exactly defined and coded by indicating two stages. For this purpose, two codes are connected with a hyphen. Thus, for instance, the code 51 - 69 describes the developmental phase from the appearance of the first inflorescence or flower buds until the end of flowering. For a uniform coding that covers the maximum number of plant species, it is necessary to use primarily phenological criteria rather than homologous or analogous stages. Thus, for instance, germination of plants from true seed and sprouting from buds are classified in one principal growth stage, the

Chapter 4.4: Phenological Growth Stages

273

principal growth stage 0, even though they are completely different biological processes. For the BBCH scales, the descriptions are based on the actual characteristic features of the individual plant. If the scales are used for the definition of the development stage of a plant stand, the description should apply to at least 50% of the plants. Greater differences in the course of the development of different plant groups have to be taken into consideration for the description of the general scale. This problem is dealt with by offering several definitions for one specific stage, wherever the formulation of a uniform text is impossible. The following letters show to which plant group the respective definitions refer: D = Dicotyledons G = Gramineae M = Monocotyledons P = Perennial plants V = Development from vegetative parts or propagated organs. No code letter is used if the description applies to all groups of plants. 2.2.1 2.2.1.1 – 00 – 01 – 02 – 03 – 04 – 05

– 06 – 07

– 08 – 09

The extended BBCH scale, general Principal growth stage 0: germination, sprouting, bud development Dry seed (seed dressing takes place at stage 00) P, V Winter dormancy or resting period Beginning of seed imbibition; P, V Beginning of bud swelling Seed imbibition complete; P, V End of bud swelling Radicle (root) emerged from seed; P, V Perennating organs (e. g. bulbs, rhizomes and tubers) forming roots Elongation of radicle, formation of root hairs and/or lateral roots G Coleoptile emerged from caryopsis; D, M Hypocotyl with cotyledons or shoot breaking through seed coat; P, V Beginning of sprouting or bud breaking D Hypocotyl with cotyledons growing towards soil surface; P, V Shoot growing towards soil surface G Emergence: Coleoptile breaks through soil surface D, M Emergence: Cotyledons break through soil surface (except hypogeal germination);

274

Phenology: An Integrative Environmental Science D, V P

Emergence: Shoot/leaf breaks through soil surface; Bud shows green tips

2.2.1.2 Principal growth stage 1: leaf development (main shoot) – 10 G First true leaf emerged from coleoptile; D, M Cotyledons completely unfolded; P First leaves separated – 11 First true leaf, leaf pair or whorl unfolded; P First leaves unfolded – 12 2 true leaves, leaf pairs or whorls unfolded – 1. Stages continuous till … – 19 9 or more true leaves, leaf pairs or whorls unfolded 2.2.1.3 – 20 – 21 G – 22 G – 23 G – 2. – 29 G

Principal growth stage 2: formation of side shoots/tillering First side shoot visible; First tiller visible 2 side shoots visible; 2 tillers visible 3 side shoots visible; 3 tillers visible Stages continuous till … 9 or more side shoots visible; 9 or more tillers visible

2.2.1.4

Principal growth stage 3: stem elongation or rosette growth, shoot development (main shoot) – 30 – 31 Stem (rosette) 10% of final length (diameter); G 1 node detectable – 32 Stem (rosette) 20% of final length (diameter); G 2 nodes detectable – 3. Stages continuous till … – 39 Maximum stem length or rosette diameter reached; G 9 or more nodes detectable 2.2.1.5

Principal growth stage 4: development of harvestable vegetative plant parts or vegetatively propagated organs/booting (main shoot) – 40 Harvestable vegetative plant parts or vegetatively propagated organs begin to develop – 41 G Flag leaf sheath extending

Chapter 4.4: Phenological Growth Stages – 42 – 43 G – 44 – 45 G – 46 – 47 G – 48 – 49 G

275

Harvestable vegetative plant parts or vegetatively propagated organs have reached 30% of final size; Flag leaf sheath just visibly swollen (mid-boot) Harvestable vegetative plant parts or vegetatively propagated organs have reached 50% of final size; Flag leaf sheath swollen (late-boot) Harvestable vegetative plant parts or vegetatively propagated organs have reached 70% of final size; Flag leaf sheath opening Harvestable vegetative plant parts or vegetatively propagated Organs have reached final size; First awns visible

Principal growth stage 5: inflorescence emergence (main shoot)/heading – 50 – 51 Inflorescence or flower buds visible; G Beginning of heading – 52 – 53 – 54 – 55 First individual flowers visible (still closed); G Half of inflorescence emerged (middle of heading) – 56 – 57 – 58 – 59 First flower petals visible (in petalled forms); G Inflorescence fully emerged (end of heading) 2.2.1.6

2.2.1.7 – 60 – 61 – 62 – 63 – 64 – 65 – 66 – 67

Principal growth stage 6: flowering (main shoot) First flowers open (sporadically) Beginning of flowering: 10% of flowers open 20% of flowers open 30% of flowers open 40% of flowers open Full flowering: 50% of flowers open, first petals may be fallen Flowering finishing: majority of petals fallen or dry

276

Phenology: An Integrative Environmental Science

– 68 – 69

End of flowering: fruit set visible

2.2.1.8 – 70 – 71 G – 72 – 73 G – 74 – 75 G – 76 – 77 G – 78 – 79 2.2.1.9 – 80 – 81 – 82 – 83 – 84 – 85

Principal growth stage 7: development of fruit 10% of fruits have reached final size or fruit has reached 10% of final size;2 Caryopsis watery ripe 20% of fruits have reached final size or fruit has reached 20% of final size2 30% of fruits have reached final size or fruit has reached 30% of final size;2 Early milk 40% of fruits have reached final size or fruit has reached 40% of final size2 50% of fruits have reached final size or fruit has reached 50% of final size;2 Milky ripe, medium milk 60% of fruits have reached final size or fruit has reached 60% of final size2 70% of fruits have reached final size or fruit has reached 70% of final size;2 Late milk 80% of fruits have reached final size or fruit has reached 80% of final size2 Nearly all fruits have reached final size2

Principal growth stage 8: ripening or maturity of fruit/seed Beginning of ripening or fruit coloration Advanced ripening or fruit coloration; G Dough stage – 86 – 87 Fruit begins to soften (species with fleshy fruit) – 88 – 89 Fully ripe: fruit shows fully-ripe color, beginning of fruit abscission

Chapter 4.4: Phenological Growth Stages

277

2.2.1.10 Principal growth stage 9: senescence, beginning of dormancy – 90 – 91 P Shoot development completed, foliage still green – 92 – 93 Beginning of leaf-fall – 94 – 95 50% of leaves fallen – 96 – 97 End of leaf fall, plants or above ground parts dead or dormant; P Plant resting or dormant – 98 – 99 Harvested product (post-harvest or storage treatment is applied at stage 99)

2.3

Specific Phenological Growth Stages: Pome Fruit

Like all special scales, the scale for apple, pear, quince, currant, and strawberry has been based on a decimal code and subdivided into 10 principal growth stages (the stages of development) in line with the general BBCH scale (Meier et al. 1994). If two stages of development (principal growth stages) run approximately parallel or overlap, as might happen during flowering (principal growth stage 6) and leaf development (principal growth stage 1), the more advanced stage must be stated. However, the stages can be coded parallel to each other, or just the more interesting stage. Unlike many other crops, part of the generative development of pome fruit takes place before the vegetative development. This means that the stages of inflorescence (bud swelling and flowering, principal growth stages 5 and 6) occur before leaf and shoot development (principal growth stages 1 and 3). Since the general BBCH scale system (Hack et al. 1992) established that principal growth stages (with their ascending numerical order) don’t have to be in succession (in a strictly hierarchical sense), a non-ascending sequence of numbers does not present a contradiction. 2.3.1

Principal growth stage 0 – bud development

Dormancy was divided by Larcherr (1976) into pre-dormancy, main dormancy and after dormancy. These dormant periods are morphologically indistinguishable and are therefore combined in one stage, the dormancy (stage 00). With rising temperatures and increasing daylight hours, preparations for bud break have already started during after dormancy. With swelling of the buds, the inner bud scales elongate, showing light, colored

278

Phenology: An Integrative Environmental Science

patches at their base (stage 01-03). For pome fruit, the beginning of bud break is characterized by the first green leaf tips becoming visible (stage 07). 2.3.2

Principal growth stage 1 – leaf development

Leaf development of pome fruit starts with the first leaves emerging and separating (stage 10). This stage is called “mouse-ear stage.” When further leaves unfold, the first developing shoot becomes visible (stage 11). The leaves have not yet fully expanded. This happens only at the end of leaf development (stage 19). 2.3.3

Principal growth stage 3 – shoot development

Shoot development (stages 31 to 35) from the terminal bud of pome fruit starts while the first leaves unfold. According to Kobel (1954) pome fruit is divided into three sections: the main shoot (starting at the beginning of spring); the “Lammas shoot” which develops from the terminal bud of the young shoot; and the autumn shoot, that develops in connection with culture methods, weather, and other environmental conditions. 2.3.4

Principal growth stage 5 – inflorescence emergence

The development of the inflorescence bud begins with bud swelling after winter dormancy (stage 51). During bud burst of pome fruit, green leaf tips enclose the flowers. As soon as the green leaf tips are above the bud scales and separate, this stage (54) can be called “mouse-ear stage” for pome fruit as well. The green bud stage of pome fruit has been reached when the single flowers start separating, for cherries, plums, peaches and apricots when the sepals are open and the flower petals elongate (stage 56). Later, most flowers of pome and stone fruit form a hollow ball with their petals, although all single flowers are still closed (stage 59). 2.3.5

Principal growth stage 6 – flowering

The time when about 10% of the flowers are open is called “beginning of flowering” (stage 61), if 50% of the flowers are open it is labeled “full flowering.” 2.3.6

Principal growth stage 7 – development of fruit

Roemer (1962) divides of pome fruit development into the following stages:

Chapter 4.4: Phenological Growth Stages

279

– from flowering up to fruit diameter of 10 mm – fruit diameter from 10 to 110 mm – fruit diameter from 110 mm up to time when fruit is ripe for picking and/or beginning of harvest. It is more favorable to state the development of fruit in mm-fruit size, as this is also a criterion for chemical thinning (stage 71-74). Natural fruit fall is also mentioned (stage 73). For further stages of fruit development, the fruit size is stated in percent, assuming that the typical size of the respective fruit is well known. 2.3.7

Principal growth stage 8 – maturity of fruit and seed

The maturity of fruit and seed runs mainly parallel to fruit development. An indication of the beginning of maturity of pome fruit is the change in fruit color, (e.g. from grass-green to greenish-yellow to yellow). In many varieties, yellow pigments become more evident by chlorophyll degradation and additionally red pigments are newly formed for the development of the overcolor. For pome fruit, the determination of the degree of ripeness on the basis of changes in the ground fruit color has proved to be useful. The fruit is ripe for picking (stage 87), which means in the case of pome fruit that the fruit is morphologically and biochemically already sufficiently developed. This stage can usually not be determined by using one characteristic only. Several fruit characteristics have to be referred to, like the fruit flesh firmness (penetrometer reading), starch breakdown value (iodine-starch test) and sugar content (refractometer reading). In doing this, fixed cultivarspecific minimum values shall be achieved. Depending of the variety, the point of time when the fruit is ripe for consumption (stage 89) can coincide exactly with the time for picking. 2.3.8

Principal growth stage 9 – senescence, beginning of dormancy

The vegetative development of pome fruit is completed when the growth of the terminal bud has finished (stage 91). At this time the foliage can still be green. Depending on culture methods, the weather and other factors, the development of undesired autumn shoots is possible. Further development is characterised by the beginning of leaf discoloring (stage 92) and the beginning of leaf fall, until all leaves have fallen (stage 93 and 97).

280 2.3.9

Phenology: An Integrative Environmental Science The specific BBCH scale for pome fruit

2.3.9.1 Principal growth stage 0: sprouting/bud development – 00 Dormancy: leaf buds and the thicker inflorescence buds closed and covered by dark brown scales – 01 Beginning of leaf bud swelling: buds visibly swollen, bud scales elongated, with light colored patches – 03 End of leaf bud swelling: bud scales light colored with some parts densely covered by hairs – 07 Beginning of bud break: first green leaf tips just visible – 09 Green leaf tips about 5 mm above bud scales 2.3.9.2 – 10 – 11 – 15 – 19 2.3.9.3 2.3.9.4 – 31 – 32 – 3. – 39 2.3.9.5

Principal growth stage 1: leaf development Green leaf tips 10 mm above the bud scales; first leaves separating First leaves unfolded (others still unfolding) More leaves unfolded, not yet at full size First leaves fully expanded Principal growth stage 2: Principal growth stage 3: shoot development3 Beginning of shoot growth: axes of developing shoots visible Shoots about 20% of final length Stages continuous till ... Shoots about 90% of final length Principal growth stage 4: -

2.3.9.6 Principal growth stage 5: inflorescence emergence – 51 Inflorescence buds swelling: bud scales elongated, with light colored patches – 52 End of bud swelling: light colored bud scales visible with parts densely covered by hairs – 53 Bud burst: green leaf tips enclosing flowers visible – 54 Mouse-ear stage: green leaf tips 10 mm above bud scales; first leaves separating – 55 Flower buds visible (still closed) – 56 Green bud stage: single flowers separating (still closed) – 57 Pink bud stage: flower petals elongating; sepals slightly open; petals just visible – 59 Most flowers with petals forming a hollow ball

Chapter 4.4: Phenological Growth Stages

281

2.3.9.7 – 60 – 61 – 62 – 63 – 64 – 65 – 67 – 69

Principal growth stage 6: flowering First flowers open Beginning of flowering: about 10% of flowers open About 20% of flowers open About 30% of flowers open About 40% of flowers open Full flowering: at least 50% of flowers open, first petals falling Flowers fading: majority of petals fallen End of flowering: all petals fallen

2.3.9.8 – 71 – 72 – 73 – 74

Principal growth stage 7: development of fruit Fruit size up to 10 mm; fruit fall after flowering Fruit size up to 20 mm Second fruit fall Fruit diameter up to 40 mm; fruit erect (T-stage: underside of fruit and stalk forming a T) Fruit about half final size Fruit about 60% final size Fruit about 70% final size Fruit about 80% final size Fruit about 90% final size

– – – – –

75 76 77 78 79

2.3.9.9 – 81 – 85 – 87 – 89

Principal growth stage 8: maturity of fruit and seed Beginning of ripening: first appearance of cultivar-specific color Advanced ripening: increase in intensity of cultivar-specific color Fruit ripe for picking Fruit ripe for consumption: fruit have typical taste and firmness

2.3.9.10 Principal growth stage 9: senescence, beginning of dormancy – 91 Shoot growth completed; terminal bud developed; foliage still fully green – 92 Leaves begin to discolor – 95 50% of leaves discolored – 97 All leaves fallen – 99 Harvested product

282

Phenology: An Integrative Environmental Science

NOTES 1

Members of the BBCH working group: L. Buhr and U. Meier (BBA), R. Klose (BSA), C. Feller (IGZ), H. Bleiholder and E. Weber (BASF), H. Hack (IVA), M. Hess, T. van den Boom and P. D. Lancashire (BAYER CROP SCIENCE), R. Stauss (Ministry of Agriculture, Schleswig-Holstein). 2 This stage is not used, if the main fruit growth happens in principal growth stage 8. 3 From terminal bud.

REFERENCES CITED Bleiholder, H., T. van den Boom, P. Langelüddecke, and R. Stauss, Einheitliche Codierung der phänologischen Stadien bei Kultur- und Schadpflanzen, Gesunde Pflanzen, 41, 381384, 1989. Bleiholder, H., H. Kirfel, P. Langelüddecke, and R. Stauss, Codificação unificada dos estádios fenológicos de culturas e ervas daninhas, Pesq. Agropec. Bras., 26, 1423-1429, 1991a. Bleiholder, H., T. van den Boom, P. Langelüddecke, and R. Stauss, Codificación uniforme para los estadios fenológicos de las plantas cultivadas y de las malas hierbas, Phytoma España, 28, 54-56, 1991b. Buhtz, E., L. Boese, C. Grunert and W. Hamann, Koordinierter Dezimalcode (KDC) der phänologischen Entwicklung für landwirtschaftliche Kulturpflanzen, Gemüse, Obst und Sonderkulturen, Feldversuchswesen 7/1, 189 pp., VEB-Verlag, Berlin, 1990. Feekes, W., De tarwe en haar milieu. Versl. techn. Tarwe Comm., 12, 523-888, 1941. Fleckinger, J., Les stades vegétatifs des arbres fruitiers, en rapport avec le traitements, Pomologie Française, Supplément, 81-93, 1948. Hack, H., H. Bleiholder, L. Buhr, U. Meier, U. Schnock-Fricke, E. Weber and A. Witzenberger, Einheitliche Codierung der phänologischen Entwicklungsstadien monound dikotyler Pflanzen - Erweiterte BBCH-Skala, -Allgemein -, Nachrichtenbl. Deut. Pflanzenschutzd., 44, 265-270, 1992. Klemm, M., Der gegenwärtige Stand der Frage über die Schädlichkeit des Apfelblütenstechers ((Anthonomus pomorum L.). Z. Angew, Entomologie, 23, 223-264, 1937. Kobel, F., Lehrbuch des Obstbaus auf physiologischer Grundlage, Springer-Verlag, Heidelberg, 422 pp., 1954. Lancashire, P. D., H. Bleiholder, P. Langelüddecke, R. Stauss, T. van den Boom, E. Weber, and A. Witzenberger, An uniform decimal code for growth stages of crops and weeds, Ann. Appl. Biol., 119, 561-601, 1991. Larcher, W., Ökologie der Pflanzen, Verlag Eugen Ulmer, Stuttgart, 322 pp., 1976. Large, E. C., Growth stages in cereals, Illustrations of the Feekes scale, Plant Pathol., 3, 128129, 1954. Meier, U., H. Graf, H. Hack, M. Hess, R. Kennel, R. Klose, D. Mappes, D. Seipp, R. Stauss, J. Streif, and T. van den Boom, Phänologische Entwicklungsstadien des Kernobstes ((Malus domestica Borkh. und Pyrus communis L.), des Steinobstes (Prunus ( -Arten), der Johannisbeere (Ribes-Arten) und der Erdbeere (Fragaria x ananassa Duch.), Nachrichtenbl. Deut. Pflanzenschutzd., 46, 141-153, 1994.

Chapter 4.4: Phenological Growth Stages

283

Meier, U., Die Merkblattserie 27 “Entwicklungsstadien von Pflanzen” der Biologischen Bundesanstalt für Land- und Forstwirtschaft, Nachrichtenbl. Deut. Pflanzenschutzd., 37, 76-77, 1985. Meier, U., BBCH-Monograph, Growth stages of plants / Entwicklungsstadien von Pflanzen / Estadios de las plantas / Stades dedéveloppement des plantes, Blackwell WissenschaftsVerlag, Berlin, 622 pp., 1997. Roemer, K., Untersuchungen über den Einfluss der Temperatur auf das Wachstum von Apfelfrüchten, Mitt. Obstbauversuchsring Altes Land, 2 pp., 1962. Soenen, A., Les bases de l’avertissement en culture fruitière. Le développement du bourgeon floral, Comptes rendus de Recherches, IRSIA, 5 pp., 1951. Soenen, A., Les bases de l’avertissement en culture fruitièr, Dissertation, Louvin, Institut Agronomique, Memoires V., 1952. Troitzky, N. N., Vorläufige Untersuchungsmittel der experimentell-biologischen Station für angewandte Entomologie. Leningrad, 1925, republished in Jahreszeitlicher Verlauf der Entwicklungsstadien bei Obstarten in Beziehung zu Jahreswitterung und Pflanzenschutzmassnahmen, edited by W. Kolbe, pp. 97-163, Pflanzenschutz-Nachrichten Bayer, 32, 1979. Zadoks, J. C., T. T. Chang, C. F. Konzak, A decimal code for the growth stages of cereals, Weed Research, 14, 415-421, 1974.

Chapter 4.5 ASSESSING PHENOLOGY AT THE BIOME LEVEL Xiaoqiu Chen Department of Geography, College of Environmental Sciences, Peking University, Beijing, China

Key words:

1.

Biome, Plant community, Remote sensing, Growing season, Regression

INTRODUCTION

Seasonal biome dynamics are controlled by recurrent variations of environmental conditions, especially climate. Therefore, in temperate areas with distinct climatic rhythmicity, seasonal biome features are rich and colorful. The most obvious natural stages during a year are growing season and rest period. Further, several characteristic sub-stages can be identified within growing season and rest period through careful observation. Because plant phenological events are integrative indicators of seasonal changes, using occurrence dates of phenophases to determine seasons (especially the growing season) is an effective way to reveal the overall characteristics of nature’s seasonality. This information is useful for doing agricultural tasks in the right season, making seasonal landscape designs, preventing plant diseases and insect pests, directing sightseeing (Yang and Chen 1980), and validating remote sensing phenology (Chen et al. 2000). Traditional methods use representative phenophases at a location to divide seasons and define the growing season. For example, in Germany, flowering of Forsythia suspensa is a symbol of the beginning of early spring, while flowering of Sambucus nigra indicates the beginning of early summer. This approach to dividing seasons is definitive, and provides an Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 285-300 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

286

Phenology: An Integrative Environmental Science

easy way to identify beginning dates of each season. However, it may not represent overall characteristics of seasonality at the biome level. An early attempt to define the seasons of local plant communities was implemented and modified by Yang and Chen (1980, 1998). They defined thresholds of cumulative frequencies of specific plant phenophases as seasonal indices, according to the typical aspect of each season, established by usage. For example, the typical indicator of early spring is germination of woody and herbaceous plants, so 10% cumulative frequency of “bud-burst” of all observed plants was used as that season’s index. Summer is characterized by the green crown canopy, so 90% cumulative frequency of “50% leaf unfolding” of all plants was used as the index; autumn begins with a few colored leaves on the canopy (5% cumulative frequency of “first leaf coloration”); and winter arrives when almost all leaves fall to earth (50% cumulative frequency of “the end of defoliation”). This method avoids the randomness of using a single phenophase to determine a season, and integrates the overall characteristics of each season. However, different aspects may represent the same season in different climate zones and plant communities. For example, in subtropical areas of China, early spring may begin with leaf replacing in some species, while summer may begin with fruit maturing. Hence, a new procedure called “phenological frequency distribution pattern” was developed for determining seasons and growing season at the biome level (Chen and Cao 1999).

2.

MEASURING PLANT COMMUNITY SEASONALITY IN THE BEIJING AREA

2.1

Data and Methods

The Beijing Botanical Garden was selected as a sample site to determine seasons of the Beijing Plain. The site is 20 km northwest from the city center, and lies on the top of a diluvial fan in front of the “West Hill”. The altitude at the site is about 100 m. Since 1979, my colleagues and I have been taking phenological observations of 70 plant species of trees, shrubs and herbs, according to the uniform observation criteria (Institute of Geography 1965) at this site, based on which the “Phenological Calendar” of the Beijing Botanical Garden was compiled (Yang and Chen 1995). Phenological data of deciduous trees and shrubs for measuring seasons were acquired from this “Phenological Calendar”. This mixed data set is composed of 707 average occurrence dates of all observed phenophases, including bud-burst, first leaf unfolding, 50% leaf unfolding, first bloom,

Chapter 4.5: Assessing Phenology at the Biome Level

287

50% bloom, the end of blooming, fruit or seed maturing, fruit or seed shedding, first leaf coloration, full leaf coloration, first defoliation, the end of defoliation, etc. The original data extended from 1979 to 1987. Sequential and overlapping occurrences of phenophases represent average seasonal succession of the local plant community. Using these data, the frequency and cumulative frequency of phenophases in every five-day period (pentad) throughout the year were calculated. Figure 1 describes phenological characteristics of the plant community in different stages of the year (Chen and Cao 1999). Four turning points are detectable on the empirical cumulative frequency curve (Figure 1). These turning points divide the curve into four main stages that correspond to spring, summer, autumn, and winter, respectively. Further, several secondary stages within each primary season can be identified on the frequency curve (Figure 1). According to the changing rates of phenological cumulative frequency between pentads and fluctuation patterns of phenological frequency, 12 phenological seasons were identified (Table 1).

Figure 4.5-1. Frequency (black, left scale) and cumulative frequency curve (gray, right scale) of phenophases at the Beijing Botanical Garden, data source Chen and Cao (1999).

2.2

Results

Spring begins in the 12th pentad and ends in the 27th pentad, during which the cumulative frequency increases very rapidly. This kind of changing

288

Phenology: An Integrative Environmental Science

pattern describes a fast developing course from sapless winter conditions to flourishing features of bud-burst, bloom, and leaf unfolding for plant canopies. About 49% of all phenophases occurs in spring. The period with maximum phenological frequency appears between the 17th and 24th pentads, in which more than 20 phenophases occur in each pentad. On the frequency curve, this period shows a significant wave crest. Hereby, spring is divided into three periods: early spring (12th-16th pentad), mid-spring (17th-24th pentad), and late spring (25th-27th pentad). Summer is the period of time from the 28th to 49th pentad. At the beginning of summer, the cumulative frequency increases slowly and then, tends to maintain a steady-state with a slight changing rate. This type of trend reflects stability in features of the local plant community. Only 12% of all phenophases appear over these 110 days. The canopy is covered with dark green leaves. Therefore, phenological summer corresponds to Table 4.5-1. Phenological seasons in the foot area of the West Hill of Beijing (1979-87). Season Beginning date Pentad Days Early spring Mid-spring Late spring Early summer Mid-summer Late summer Early autumn Mid-autumn Late autumn Early winter

2.25 3.22 5.1 5.16 6.15 8.9 9.3 9.28 10.18 11.22

12-16 17-24 25-27 28-33 34-44 45-49 50-54 55-58 59-65 66-68

25 40 15 30 55 25 25 20 35 15

Mid-winter

12.7

69-7*

60

Late winter 2.5* * date and pentad of the next year

8-11*

20

the period of maximum photosynthetic activity. Three periods are identified: early summer with a relatively high frequency (28th-33th pentad), midsummer with a decreasing frequency until the lowest level (34th-44th pentad), and late summer with a slightly increasing frequency (45th-49th pentad). Autumn (from the 50th to 65th pentad) is the second most rapid developing stage in phenological cumulative frequency, compared to spring. About 38% of all phenophases occur in autumn, which reflects the changing process from exuberant dark green leaves to yellow or red, to (at the end of this season) shedding of almost all leaves on plant canopies. The frequency curve shows three wave crests, based on which three periods are delimited:

Chapter 4.5: Assessing Phenology at the Biome Level

289

early autumn (50th-54th pentad), mid-autumn (55th-58th pentad), and late autumn (59th-65th pentad). Winter begins in the 66th pentad and ends in the 11th pentad of the next year. It is a period with few phenological changes, or even a pause in biological activity. Only 1% of all phenophases appear in winter, including “the end of defoliation” of a few trees at the beginning of this season and “bud-burst” of several species at the end. Deciduous plants enter a state of dormancy and show bare branches in winter. Like other seasons, winter is divided also into three periods: early winter with the end of defoliation (66th68th pentad), mid-winter with deep dormancy (69th pentad-7th pentad of Table 4.5-2. Percent occurrence frequencies and total cases of selected phenophases in each season. MidSeasons Early Late Early MidLate spring spring spring summer summer summer Bb Lu-first Lu-50% B-first B-50% B-end Fm Fs Cases Seasons

93.9

3.0 3.0

33 Early autumn

14.9 24.1 13.3 20.0 14.9 12.3 0.5 195 Midautumn

3.7 7.4 22.2 14.8 40.7 11.1 27 Late autumn

4.3 41.3 26.1 23.9 2.2 2.2 46 Early winter

36.4 18.2 31.8 13.6

14.3

22 Midwinter

7 Late winter 100

42.9 42.9

Bb Lu-first Lu-50% B-first 3.9 1.1 0.8 B-50% 3.9 1.1 9.8 1.5 B-end Fm 25.5 1.1 Fs 11.8 2.2 Lc-first 29.4 51.7 8.4 Lc-full 2.0 6.7 28.2 D-first 13.7 31.5 13.7 4.5 47.3 100 D-end Cases 51 89 131 6 0 3 Abbreviations: Bb: bud-burst; Lu-first: first leaf unfolding; Lu-50%: 50% leaf unfolding; Bfirst: first bloom; B-50%: 50% bloom; B-end: the end of blooming; Fm: fruit or seed maturing; Fs: fruit or seed shedding; Lc-first: first leaf coloration; Lc-full: full leaf coloration; D-first: first defoliation; D-end: the end of defoliation.

290

Phenology: An Integrative Environmental Science

the next year), and late winter with resuming growth (8th-11th pentad of the next year).

2.3

Characteristic Aspects of Phenological Seasons

In order to understand landscape-scale aspects of phenology and their changes in each season, my colleague and I analyzed the combination and structure of plant morphology and colors (Chen and Cao 1999). The plant morphological combination can be described by occurrence frequencies of different phenophases in each season (Table 2). The primary phenological event in early spring is “bud-burst” that accounts for 94% of all events (“bloom” is second). The species with earliest leafing include: Ulmus pumila, Populus tomentosa, Prunus davidiana, and Jasminum nudiflorum. The earliest bloom appears on a shrub called Chimonanthus praecox. Mid-spring is a season with abundant flowers. About 47% of events are related to bloom, and another 37% are related to leaf unfolding. In late spring, bloom accounts for 78% of all events, but the total numbers of phenophases are much smaller than midspring. Green leaves have covered canopies of most trees and shrubs by this time. In addition, several plants have even begun to spread their seeds, such as Salix matsudana, Populus alba, and Taraxacum mongolicum. Early summer is another flowering season second only to mid-spring. The frequency of bloom reaches 91%. From early summer to late summer, flowering trees and shrubs become less frequent. In addition to the flowering plants of spring, some summer species have a very long florescence with several “high tides”, such as Ziziphus jujuba, Sorbaria kirilowii, Albizzia julibrissin, Sophora japonica, Lagerstroemia indica, etc. Coming into early autumn, landscape features are obviously different than in late summer. “Fruit maturing and shedding” (accounting for 37% of all events) becomes the dominant phenophase. “Leaf coloration” accounts for 31% of all events. Mid-autumn is a multicolored season. About 58% of all phenophases are associated with “leaf coloration”, whereas another 36% deal with “defoliation”. During late autumn, most deciduous trees and shrubs enter into “full leaf coloration” and “the end of defoliation.” The frequency of both phenophases reaches 76%. In early winter, only a few species of trees and shrubs, such as Salix babylonica, Malus micromalus, and Chimonanthus praecox, retain a small quantity of leaves, and soon afterwards all leaves fall to earth. There are almost no changes in plant morphology during mid-winter. Other than the silent mid-winter, late winter reveals a little spring breath, as the buds of several plants, such as Chimonanthus praecox, Salix matsudana, and Ulmus pumila, begin to expand.

Chapter 4.5: Assessing Phenology at the Biome Level

291

Flowers, fruits, and autumn leaves represent the colorful characteristics of landscape-scale phenology, so the color combination in each season is portrayed by occurrence frequency of flowers, fruits, and autumn leaves with different colors. Here, only the flowers, fruits, and leaves with remarkable and vivid colors (except green) are included in the statistics (Table 3). Table 4.5-3. Percent occurrence frequencies and total cases of phenophases associated with colorful flowers, fruits, and leaves in each season. Season Sp-1* Sp-2 Sp-3 Su-1 Su-2 Su-3 Au-1 Au-2 Au-3 Flowers: Yellow 100 42.1 36.8 31.0 52.6 22.2 100 33.3 Red 2.6 5.3 10.5 11.1 White 13.2 21.1 40.5 15.8 50.0 11.1 Pink 26.3 10.5 9.5 10.5 11.1 Purple 15.8 26.3 19.0 10.5 50.0 44.4 66.7 Cases 2 76 19 42 19 4 9 2 3 Fruits: Yellow 33.3 38.5 Red 100 33.3 38.5 Brown 66.7 66.7 23.1 100 Cases 1 3 3 13 1 Leaves: Yellow 100 75.0 67.3 72.9 Red 12.5 17.3 14.6 Brown 12.5 15.4 12.5 Cases 1 16 52 48 * Sp-1: early spring, Sp-2: mid-spring, Sp-3: late spring; Su-1: early summer, Su-2: midsummer, Su-3: late summer; Au-1: early autumn, Au-2: mid-autumn, Au-3: late autumn.

In terms of the color combination, spring is a season with multiple flowers of various colors. In mid-spring, plants with yellow and pink flowers account for 42% and 26% of all flowering plants, respectively, whereas in late spring yellow, purple, and white flowers are dominant (but the number of flowering plants is only one fourth of that in mid-spring). It is interesting that few red flowers appear in spring. Summer flowers are characterized as yellow and white color. In early and mid-summer, plants with yellow and white flowers account for about 70%, and next those with purple, pink, and red flowers. In addition, several fruits begin to mature and show brown, red, and yellow colors in summer. The continuous green canopy of the plant community forms the background of colorful flowers and fruits. The color of early autumn is interspersed mainly by yellow and red fruits and leaves. Abundant leaves of various colors characterize midautumn and late autumn. The famous plants with red leaves include Rhus

292

Phenology: An Integrative Environmental Science

typhina, Acer truncatum, Acer mono, and Cotinus coggygria var. cinerea. Entering winter, as all autumn leaves fall to earth, the landscape features only gray and brown colors. The seasonal color combination of a plant community in a quasi-natural landscape provides a basis in garden design for showing the seasonal features at the location.

2.4

Phenological Seasons and Climate Conditions

Climate conditions are the main causes of plant phenology. In order to reveal the relationships among seasonal succession of cumulative frequency, frequency of phenophases, and seasonality of heat and moisture regimes, Chen and Cao (1999) calculated the pentad mean air temperature, accumulated daily mean air temperature (GDD) above 0°C, pentad mean precipitation, and accumulated precipitation in each season. Table 4 shows that seasonal phenological progress coincides with seasonal climate fluctuations. During phenological spring, the pentad mean temperature increases rapidly from 2.4°C to 18.4°C. On average, the rate of temperature increase is 1°C per pentad. The growing degree-day accumulation (GDD) in spring is 811.2°C days, accounting for 18% of annual total GDD. At the same time, pentad mean precipitation increases slightly and accumulated precipitation accounts for only 6.4% (41.3 mm) of annual total precipitation. Phenological summer has the most abundant heat and moisture resources. Pentad mean temperatures exceed 22°C, with peak value in mid-summer. The GDD reaches 2657.7°C days, accounting for 59% of annual total GDD. Pentad precipitation in mid-summer also reaches its highest level of the year and the accumulated precipitation accounts for 79.4% (512.3 mm) of annual total precipitation. In contrast to spring, pentad mean temperature decreases sharply from 18.8°C to 6.2°C in autumn. On average, the rate of decrease is 0.8°C per pentad. GDD during autumn is noticeably larger than in spring, accounting for 23% (1034.3°C days) of annual total GDD. Similarly, pentad mean Table 4.5-4. Distribution of temperature and precipitation in each phenological season. Seasons Pentad mean Pentad mean GDD above Accumulated temperature 0°C (°C day) precipitation precipitation (°C) (mm) (mm) Early spring Mid-spring Late spring Early summer Mid-summer

2.4 11.6 18.4 22.0 25.6

67.1 467.7 276.4 656.2 1405.5

1.7 2.7 3.6 8.1 29.1

8.7 21.9 10.7 48.8 320.5

Chapter 4.5: Assessing Phenology at the Biome Level

293

Seasons

Pentad mean temperature (°C)

GDD above 0°C (°C day)

Pentad mean precipitation (mm)

Accumulated precipitation (mm)

Late summer Early autumn Mid-autumn Late autumn Early winter Mid-winter Late winter

23.5 18.8 13.9 6.2 -0.4 -4.1 -2.1

596.0 487.5 289.2 257.6 9.1 0 0

28.6 8.8 4.8 2.3 0.8 0.5 1.2

143.0 44.1 19.2 16.0 2.4 5.5 4.6

precipitation decreases from 8.8 mm to 2.3 mm, with accumulated precipitation accounting for 12.3% (79.3 mm, about 1.9 times as large as that in spring) of annual total precipitation. Phenological winter is the period with the lowest temperature and the least precipitation. Pentad mean temperatures are lower than 0°C and the GDD is extremely small (only 9.1°C days) compared with the annual total value of 4512.3°C days. The minimum pentad mean precipitation appears in mid-winter with 0.5 mm, and the accumulated precipitation in winter is also very small, only 12.5 mm.

3.

MEASURING PLANT COMMUNITY GROWING SEASON IN EASTERN CHINA

The phenological cumulative frequency curve of a plant community can not only be used to determine seasons but also to measure the growing season. Since growing season is commonly defined as the number of days during a year in which plants (crops) can grow (Wang 1963, p. 108), the growing season of a plant community can therefore be understood as the time interval between the beginning of phenological spring and the end of phenological autumn. Chen et al. (2000, 2001) analyzed plant communities in eastern China using this approach, and their methods and results are summarized in the following sections.

3.1

Data and Methods

Phenological data were acquired from Yearbooks of Chinese Animal and Plant Phenological Observation (Institute of Geography at Chinese Academy of Sciences, 1988, 1989a, b, 1992) and from an unpublished database. The study period extended from 1982 to 1993. Mudanjiang (44°26’N, 129°40’E, 300m), Beijing (40°01’N, 116°20’E, 50 m) and Luoyang (34°40’N,

294

Phenology: An Integrative Environmental Science

112°25’E, 155 m) were selected as sample stations at which phenological data cover above 40 plant species and 10 phenophases per species. Mudanjiang is located in the middle temperate and humid climate zone with natural vegetation of deciduous broad-leaved and coniferous mixed forests, whereas Beijing and Luoyang lie in the warm temperate and sub-humid climate zone with natural vegetation of deciduous broad-leaved forest. The topography at the stations is nearly flat. Although the plant community compositions at the three locations have different characteristics, all plant species observed were selected according to the Chinese phenological observation guide (Institute of Geography at Chinese Academy of Sciences, 1965) to assure spatial comparability. The corresponding meteorological stations at Mudanjiang and Beijing are near the phenological stations, but there is not a meteorological station at Luoyang. Therefore, Mengjin (34°50’N, 112°26’E, 324.6 m) located 18 km north of Luoyang was substituted as the meteorological station for that site. Climate data contain daily mean air temperature and daily precipitation, based on which monthly and annual mean air temperatures, monthly and annual precipitations, and growing degree-days (GDDs) above 5°C were calculated. The climate variables chosen generally represent energy, heat, and moisture regimes affecting the phenological growing season of plant communities. In order to determine the growing season at a site, a mixed data set composed of all phenological occurrence dates of deciduous trees and shrubs was established for each year. Then, the frequency and cumulative frequency of the phenological occurrence dates in every 10-day period throughout each year were calculated. The precise thresholds for determining the growing season were arbitrarily set as the dates on which the phenological cumulative frequency reached 5% and 10% (for the beginning date), and 95% and 90% (for the end date). Then, the temporal characteristics of growing season parameters were presented. Further, correlation analyses were performed between time series of growing season parameters and climate variables for the three stations to examine statistical relationships. As a last step, simple and multiple linear regression models were created to quantitatively describe responses of the growing season to climate change.

3.2

Temporal Characteristics of the Growing Season

The beginning and end dates of the growing season varied as expected with latitude, namely, stations at higher latitude had later beginning and earlier end dates to the growing season, and a smaller standard deviation (SD), whereas those at lower latitude had earlier beginning and later end

Chapter 4.5: Assessing Phenology at the Biome Level

295

dates to the growing season, and a larger SD. So, the length of the growing season and its SD increase from higher latitude to lower latitude (Table 5). Regarding interannual fluctuations, a similar pattern was found between the beginning dates, the end dates, and the lengths of the two growing seasons (see definitions in Table 5) at each station. The 3-year moving average was calculated (not shown) to detect the linear trend in the length of the growing season. The results show that a significant linear trend appeared only in Beijing from 1982 to 1993 (Figure 2). For growing season 1, the lengthening was 1.43 days per year, whereas for the growing season 2, the lengthening was 0.68 days per year. Thus it can be seen that the trends in the length of the growing season show an inconsistent pattern in different Table 4.5-5. Mean and standard deviation of growing season parameters at each station. Station Measure BGS1 BGS2 EGS1 EGS2 LGS1 LGS2 Mudanjiang

Mean (day 96 104 286 277 191 173 of year or days) SD (days) 5 4 3 3 6 5 72 82 314 305 242 223 Beijing Mean (day of year or days) SD (days) 6 6 7 5 10 8 Luoyang Mean (day 61 73 322 310 262 237 of year or days) SD (days) 10 8 7 9 11 11 Note: BGS1-beginning of growing season 1 (5%); BGS2-beginning of growing season 2 (10%); EGS1-end of growing season 1 (95%); EGS2-end of growing season 2 (90%); LGS1-length of growing season 1 (5%-95%); LGS2-length of growing season 2 (10%90%).

parts of the research region, which is perhaps due to differences in the composition of local plant communities and in topoclimate conditions. Accordingly, significant linear trends in the length of the growing season at specific sites can generally not be used as a simple indicator of regional vegetation dynamics.

3.3

Growing Season and Climate Factors

In order to explore phenology-climate relationships at the plant community level, a correlation analysis was undertaken among the beginning of the growing season (BGS), the end of the growing season (EGS), the length of the growing season (LGS), and seasonal climate

296

Phenology: An Integrative Environmental Science

variables (Chen and Pan 2002). Table 6 shows that the BGS is mainly influenced by mean air temperature and GDD above 5°C from January to March at Beijing and Luoyang, and from February to April at Mudanjiang. In addition, some monthly mean temperatures also correlate highly with the BGS at Mudanjiang and Beijing. The overall negative correlation indicates that higher mean temperatures and GDD totals in late winter and spring may induce an earlier onset of the phenological growing seasons of local plant communities. However, there is no significant correlation between BGS and seasonal precipitation.

Figure 4.5-2. Growing season 1 (black) and 2 (gray) duration at Beijing, data source Chen et al. (2001).

In contrast to the BGS, a significant correlation between EGS and climate variables was only detectable at Luoyang. The positive correlation between EGS and September-November mean temperature, and between EGS and September-November GDD above 5°C, indicates that higher mean temperatures and GDD totals in autumn result in a later end of the growing season at Luoyang. The negative correlation between the end date of growing season 2 and September-November precipitation suggests that a

Chapter 4.5: Assessing Phenology at the Biome Level

297

drier autumn with less precipitation delays the end of growing season 2 at Luoyang (Table 7). Eight LGS models for all three sites were constructed between growing season duration (as the dependent variable), and mean temperature during late winter and spring, annual mean temperature, and annual total GDD above 5°C (as independent variables), respectively. All three independent Table 4.5-6. Correlation between BGS and climate variables at individual sample stations, * significance level 0.05, ** significance level 0.01, B=Beijing, L=Luoyang, M=Mudanjiang Variables M M B B L L BGS 1 BGS 2 BGS 1 BGS 2 BGS 1 BGS 2 T1 (2) -0.604* -0.675* -0.531 -0.344 -0.307 -0.170 T2 (3) -0.764** -0.728** -0.514 -0.456 -0.298 -0.273 T3 (4) -0.645* -0.734** -0.463 -0.683* -0.316 -0.487 T1-3 (2-4) -0.824** -0.868** -0.635* -0.662* -0.584* -0.589* GDD1-3 (2-4) -0.640* -0.743** -0.537 -0.712** -0.610* -0.733** P1-3 (2-4) -0.148 0.042 -0.052 -0.009 0.235 0.113 T1 (2): January (Beijing and Luoyang) or February (Mudanjiang) mean temperature T2 (3): February (Beijing and Luoyang) or March (Mudanjiang) mean temperature T3 (4): March (Beijing and Luoyang) or April (Mudanjiang) mean temperature T1-3 (2-4): January-March (Beijing and Luoyang) or February-April (Mudanjiang) mean temperature GDD1-3 (2-4): January-March (Beijing and Luoyang) or February-April (Mudanjiang) accumulated growing degree-days above 5°C P1-3 (2-4): January-March (Beijing and Luoyang) or February-April (Mudanjiang) accumulated precipitation Table 4.5-7. Correlation between EGS and climate variables at individual sample stations, * significance level 0.05, ** significance level 0.01, B=Beijing, L=Luoyang, M=Mudanjiang Variables M M B B L L EGS 1 EGS 2 EGS 1 EGS 2 EGS 1 EGS 2 T9 (8) -0.150 0.052 0.521 0.507 0.611* 0.562 T10 (9) 0.213 0.222 -0.136 -0.050 0.524 0.569 T11 (10): 0.229 0.208 -0.112 0.027 0.195 0.008 T9-11 (8-10) 0.145 0.234 0.053 0.160 0.737** 0.630* GDD9-11 (8-10) 0.128 0.218 0.129 0.162 0.659* 0.603* P9-11 (8-10) 0.259 0.071 0.186 0.022 -0.530 -0.583* T9 (8): September (Beijing and Luoyang) or August (Mudanjiang) mean temperature T10 (9): October (Beijing and Luoyang) or September (Mudanjiang) mean temperature T11 (10): November (Beijing and Luoyang) or October (Mudanjiang) mean temperature T9-11 (8-10): September-November (Beijing and Luoyang) or August-October (Mudanjiang) mean temperature GDD9-11 (8-10): September-November (Beijing and Luoyang) or August-October (Mudanjiang) accumulated growing degree-days above 5 °C P9-11 (8-10): September-November (Beijing and Luoyang) or August-October (Mudanjiang) accumulated precipitation

298

Phenology: An Integrative Environmental Science

variables correlate positively with the length of growing seasons 1 and 2 ( <0.001, Table 8). In contrast, the correlation between growing season (p duration and annual total precipitation was very low, which means that annual precipitation was not the limiting factor on growing season duration in the study area. Because there are close dependencies among the three climate variables, no multiple linear models could be constructed. Table 4.5-8. Summary statistics of LGS models for all three stations. Y X r Sig. Intercept

Slope

SE

LGS1 T1-3 (2-4) 0.934 0.000 227.33 11.52 LGS2 T1-3 (2-4) 0.925 0.000 207.08 10.39 LGS1 AT 0.961 0.000 157.52 7.24 AT 0.969 0.000 142.86 6.65 LGS2 LGS1 GDD 0.947 0.000 93.50 0.05 LGS2 GDD 0.961 0.000 83.40 0.04 AT: annual mean temperature; GDD: total growing degree-days in a year

11.5 11.1 8.9 7.2 10.3 8.1

On average, if annual mean temperature and mean temperature during late winter and spring increase by 1°C, the growing seasons would lengthen 6.7-7.2 days and 10.4-11.5 days, respectively, while if annual total GDD increases 100 degree days, the growing seasons would lengthen 4-5 days. Since the LGS models were based on the spatial-temporal series of all three sample stations, they should not only be useful to estimate the growing season at the sample stations but also at other adjacent sites with similar vegetation and climate conditions.

4.

CONCLUSIONS AND PERSPECTIVES

Other than the traditional methods for determining phenological seasons, the approach of “phenological frequency distribution pattern” is quantitative, integrative, broadly applicable, and can be used for assessing phenology and seasonality at the biome level. The major seasons and sub-seasons reveal significant stages of seasonal feature of local plant communities in relation to ecosystem dynamics and aesthetics. The Beijing Botanical Garden has applied the results in seasonal landscape design, daily garden management, and seasonal sightseeing service. In order to define phenological seasons using this approach at the biome level, relatively abundant phenological data are needed. Generally speaking, the more the plant species and phenophases are observed, the more precise the phenological seasons are defined. The phenological cumulative frequency of plant communities also provides a sensible and real metric for determining the growing season at the

Chapter 4.5: Assessing Phenology at the Biome Level

299

biome level. The growing season-climate relationships established by LGS models indicate that higher mean air temperature during late winter and spring, annual mean air temperature, and annual GDD totals can induce longer growing season duration. Since the phenological growing season represents approximately the period of photosynthesis, it becomes an important parameter for global terrestrial carbon-cycle modeling and net primary productivity modeling. In this context, the growing season duration of plant communities may be used to assess the feedback of biomes to the seasonal CO2 cycle and climate change. In addition, growing season parameters at the biome level and seasonal metrics of normalized difference vegetation index (NDVI) derived from satellite sensors can also be used to extrapolate the phenological growing season at regional scales (Chen et al. 2001). Further studies may verify the accuracy of the spatial extrapolation of the growing season, and explore the possibility of temporal extrapolation of the growing season using surface and satellite sensor-derived phenological data.

REFERENCES CITED Chen X., and Z. Cao, Frequency distribution pattern of plant phenophases and its application to season determination (in Chinese), Scientia Geographica Sinica, 19 (1), 21-27, 1999. Chen X., and W. Pan, Relationships among phenological growing season, time-integrated Normalized Difference Vegetation Index and climate forcing in the temperate region of Eastern China, Int. J. Climatology, 22(14), 1781-1792, 2002. Chen X., Z. Tan, M. D. Schwartz, and C. Xu, Determining the growing season of land vegetation on the basis of plant phenology and satellite data in Northern China, Int. J. Biometeorol., 44(2), 97-101, 2000. Chen X., C. Xu, and Z. Tan, An analysis of relationships among plant community phenology and seasonal metrics of Normalized Difference Vegetation Index in the northern part of the monsoon region of China, Int. J. Biometeorol., 45(4), 170-177, 2001. Institute of Geography at Chinese Academy of Sciences, Yearbook of Chinese Animal and Plant Phenological Observation No. 1, Science Press, Beijing, 122 pp., 1965. Institute of Geography at Chinese Academy of Sciences, Yearbook of Chinese Animal and Plant Phenological Observation No. 8, Geology Press, Beijing, 247 pp., 1988. Institute of Geography at Chinese Academy of Sciences, Yearbook of Chinese Animal and Plant Phenological Observation No. 9, Geology Press, Beijing, 268 pp., 1989a. Institute of Geography at Chinese Academy of Sciences, Yearbook of Chinese Animal and Plant Phenological Observation No. 10, Survey and Drawing Press, Beijing, 281 pp., 1989b. Institute of Geography at Chinese Academy of Sciences, Yearbook of Chinese Animal and Plant Phenological Observation No. 11, Chinese Science and Technology Press, Beijing, 347 pp., 1992. Wang J., Agricultural Meteorology, Pacemaker Press, Milwaukee, 693 pp., 1963. Yang G., and X. Chen, A preliminary study on phenological seasons in the Beijing area (in Chinese), Journal of Beijing Teacher’s College, 1(2), 110-119, 1980.

300

Phenology: An Integrative Environmental Science

Yang G., and X. Chen, Phenological Calendars and their Applications in the Beijing area, Capital Normal University Press (in Chinese), Beijing, 309 pp., 1995. Yang G., and X. Chen, On the seasonal rhythm of natural landscapes (in Chinese), Acta Ecologica Sinica, 18(3), 233-240, 1998.

Chapter 4.6 DEVELOPING COMPARATIVE PHENOLOGICAL CALENDARS Rein Ahas and Anto Aasa Institute of Geography, University of Tartu, Tartu, Estonia

Key words:

1.

Calendars, Seasonal rhythms, Comparisons, Averages, Linear changes

INTRODUCTION

Phenological calendars describe the beginning, duration, and mutual relationships among seasonal natural phenomena. They can be used to help describe seasonality of environmental conditions, ecosystems, and individual species. Calendars consist of various types of seasonal data, such as phenological phases of living organisms, seasonal phases of environmental conditions, and instrumentally measured climatic parameters. Phenological calendars are an integrative method for studying those seasonal phenomena and representing seasonality graphically. Series of phenological phases show variations in climate and natural processes in an integrated way, and therefore are a useful approach to climate change studies. At the same time, impacts on organisms and ecosystems are an important outcome of changes in climate (Penuelas and Filella 2001). Therefore, connecting phenology with long time-series has become crucial for studying climate change, and phenological calendars can help examine its variations in space and time. The aim of this chapter is to present an overview of theoretical and practical methods for composing phenological calendars, using example data from eight European phenological observation sites, over the 1951-1998 period. They were collected from different national observation networks as part of the European Fifth framework program POSITIVE (Figure 1, Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 301-318 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

302

Phenology: An Integrative Environmental Science

Table 1, and Menzel et al. 2001). Collectively, more than 400 phases of flora, fauna, and the physical environment have been observed at these stations. The stations and phases presented here were chosen according to length of time-series, quality of data, and number of matching species. Commonly, most phenological data come from the spring-summer period; the role of autumn-winter phases is also minimal in these calendars. Climate parameters represented in the calendars are derived from daily mean temperatures.

Figure 4.6-1. Beginning of birch leaf unfolding in Europe, with observation sites used in this chapter.

Table 4.6-1. Coordinates of stations used in this chapter. Station Country Longitude Trier-Petrisberg GER 6.65 Schleswig GER 9.57 Viechtach GER 12.88 Weiz AUT 15.63 Pärnu EST 24.52 Zhitomir UKR 28.67 Borok RUS 38.13 Kuznetskoye RUS 60.50

Latitude 49.75 54.52 49.08 47.22 58.37 50.33 57.88 55.50

Elevation (m) 260 18 430 465 3 221 111 274

Chapter 4.6: Developing Comparative Phenological Calendars

2.

303

THEORETICAL OVERVIEW

Phenological calendars (used for describing seasonality of places, years, or individual species) consist of developmental phases of living organisms (phenophases), phases of physical environmental conditions, and descriptive climate parameters. Typical phenophases among plants include beginning of sap rising, budding, sprouting, leaf unfolding, blooming, ripening of fruits, leaf coloring and fall, and end of growth. Phases of the physical environment are snow and ice phenomena, ground condition phenomena, Descriptive climate parameters most hydrological phenomena, etc. commonly used in phenological calendars (based on temperature, precipitation, radiation and wind measurements) consist of climatic seasons, sums (of active, effective, and negative temperatures), frost dates, duration of rain, drought, and growing seasons. Climatic seasons are homogeneous stages of climatic cycles within the year that are characterized by a common source direction and genesis for climate processes and phenomena. They can be determined from daily average permanent transitions of temperature during the warm seasons, and by snow phenomena in winter (Galahov 1959; Raik 1963; Jaagus and Ahas 2001). Many other seasonal phenomena can be expressed in phenological calendars, for example widespread characteristics of primary sectors like tilling of the soil, beekeeper’s calendars, and fisherman’s calendars. Phenological phases in the calendars are expressed as dates. In addition to these dates, the intervals between phases are called durations, and used for describing periods with homogenous conditions. Phenophases and other phenomena that occur every year are rhythmical. However, considered collectively, they describe an annual sequence of cycles and periods, which are not strictly regular in timing or length. Phenophase series are not exactly periodical (time intervals between them are different each year) nor cyclic (they do not return to exactly the same development phase at the same time each year). Thus, phenological calendars are often based on long-term averages of the seasonal rhythms of different phenomena (Ahas 1999; Ahas et al. 2000). Calendars are divided into two main types: descriptive calendars that show the occurrence and sequence of certain phases; and relational calendars that compare phases and their ecological implications. A dendrophenological calendar is an example of descriptive calendars that show gardeners the average blooming period and its variability in a specific area. Agri-phenological calendars belong to the group of relational calendars, as they combine temperature and the condition of ground, and developmental phases of plants, with optimal times for tilling soil, or calendars to predict/prevent/protect against plant pests. Phenological calendars can be

304

Phenology: An Integrative Environmental Science

produced either by dividing phases according to similar climatic seasons or determining seasons (periods with homogeneous conditions and similar variability) based on the phases. Developing and naming phenological seasons is usually based on a local tradition, depending on the characteristics of the species involved. Phenological calendars are often produced from descriptive phenology. Tables describing seasonal phenomena or results of observations in the region are typical of all countries with seasonal climates. Among the many phenologists who have worked out methods to compose and design phenological calendars are authors like Hopkins and Murray (1933) in the USA, Ihne (1895) and Schnelle (1955) in Germany, Schultz (1981) and the Russian Geographical Society (Hydrometeorological Printing House 1965) in the former Soviet Union, Defila (1992) in Switzerland, Klaveness and Wielgolaski (1996) in Norway, Ahas (2001) in Estonia, and many others. When composing a robust phenological calendar (describing the whole annual circle), all seasons must be equally represented with indicative or typical data. In order to describe the phenology of the ecosystem, key species and phases of all important life forms are critical, as well as indicative species to determine seasons and sub-seasons. For example, in Estonia, the indicative phase for beginning of early spring is birch sap rising, for spring, maple full bloom, and for summer, pollination of rye. In this region, tree species are the dominant source of data, as they are easiest to observe and compare. The most important consideration in calendar development is length and quality of the observation series used for composition. Time-series that are observed in one place year after year provide the most reliable information. The length of time-series needed to create a calendar depends on data characteristics and objectives of the work. According to the literature, the minimal length of observation series used for determining mean beginning dates should be 25 years (Gornik 1994), but it is also depends on trends and variability of data. We have used a minimal period of 20 years, as it is difficult (even in older observation programs) to find series that have been observed for at least 25 years at the same place. In order to conduct a proper regression analysis we still recommend a time-series of more than 35 years. A typical data series (Zhitomir, Ukraine) and associated linear trends are shown in Figure 2. Linear trends of long time-series are generally reliable, but we recommend using statistics from the most recent 10 years to determine average dates, since phenophase statistics from the first (1950s) and last (1990s) decades of our observation period can differ by 2-3 weeks. Phenological data usually have a normal distribution, but if the standard deviation is too high or there are phenological interceptions (overlap of

Chapter 4.6: Developing Comparative Phenological Calendars

305

phase timing), transformation of the data may be necessary before computing statistics. The disadvantages of all the phenological observation programs are the differences among phases/species observed, and gaps in the data. One aim of phenological calendars is to provide a method for standardizing timeseries. Comparing observations of the same phase with neighboring stations, or even similar phases at the same station can facilitate filling spatial and temporal gaps. The methods of interpolation when using calendars are different than those often employed with standard data, because phenological phases are related with each other within the calendar and among different calendars at neighboring stations. Typically with calendars, the average, standard deviation, and trends are compared at one observation site (or neighboring stations) and then similar phases are found based on that information. The next step is correlation analysis and determining the most similar phases with highest correlation. These then serve as a basis for interpolating or comparing station time-series. In cases with 20 or more years of observations, we have arbitrarily chosen a correlation coefficient of over 0.7 as the threshold for interpolation, but this value can be different for each location or region.

3.

PHENOLOGICAL CALENDARS

3.1

Descriptive Statistics and Graphic Design

Phenological calendars can be produced with a wide variety of statistical parameters and layouts. The study objectives, data, and methods determine which characteristics and graphic devices will be used to describe annual cycles or seasons. In this chapter, we present typical statistics and methods in our European phenological calendar, which includes a selection of phases from Pärnu, Estonia and statistics for Trier-Petrisberg, Germany. 3.1.1

The average date – calculated for study period

Usually the sequence of phases at an observation site or region is stable, and they are referred to as the phenological “stripe-code” of a region or species. Phenological interception (exchange of phase order) occurs during exceptional years among phases already close to each other in time. The sequence of phases also may vary in different geographical regions. There are many ways to present average dates graphically. Two of the most common are the phenological time-series with trends (Figure 2), and annual circles (Figure 3). If there is more than one phase for each species, then data

306

Phenology: An Integrative Environmental Science

are presented in phenospectric tables. Figure 4 shows a part of the mycophenological calendar of Estonia (Kalamees 2001), where parameters of the beginning, end, and duration of different species are highlighted. 3.1.2

Count – number of observations

Knowing the number of observations complements other characteristics and is important in phenology, since observation series often have gaps or are periodical. Presenting dates of important phases may be worth doing even if the number of observations is less than the critical level, because in certain cases calendar goals can be obtained with less than optimal amounts of information. 3.1.3

Standard deviation – describing data variability

In Figure 5, the standard deviations of two observation stations are compared. Temperate zone deviations are typically greater in early spring, and decrease by the beginning of summer. Variability starts rising again in late summer and reaches another peak during autumn phases. The large

Figure 4.6-2. (left) Temporal variability of phenological time series recorded in Zhitomir (Ukraine) from 1951-1998. 1) pollination of Hazel; 2) birch leaf unfolding; 3) apple full bloom. Figure 4.6-3. (right) Annual circle of climatic seasons in Vietach (inner circle) and Kuznetskoye (outer circle).

variations during spring and autumn are caused by the dependence of those phases on many factors and occasional impacts. Variety of climatic seasons is greater during spring and autumn as well. General data variety is often presented through quartiles. A graphic calendar with median and quartiles offers an excellent overview of a phase’s temporal dynamics (Figure 6).

Chapter 4.6: Developing Comparative Phenological Calendars 3.1.4

307

Minimal and maximal values and year(s) of their occurrence – describing the distribution and range of extremes

Extremes enhance standard deviation information and enable better estimates of general variety, and extreme years may show repetitive oscillations. In addition to extremes, the results of early-late years analysis, (years beyond the ±1 standard deviation value level) are often presented. The distribution of early and late years may help describe interannual variability in greater detail, these data are especially interesting due to their disproportionate influences on the direction of trends and cycles. The most reliable method for classifying years or seasons as early or late uses percentage thresholds of early and late values from a phenological calendar (Ahas 1999).

Figure 4.6-4. The general table of mycophenological spectra (Kalamees 2001).

Figure 4.6-5. Distribution of standard deviations at two observation stations.

308 3.1.5

Phenology: An Integrative Environmental Science Slope and p-values – describing linear changes and their confidence levels

Assessment of trends and slopes are an important part of climate change studies. Further, spatial and temporal distribution of slopes help to describe local phenological variability and distinguish differences. In Figure 7, there is a selection of slope values from linear trends at the Viechtach and Pärnu observation stations. In addition to describing important climatic changes, linear trends are also essential for checking the reliability of a calendar. If the slope of 50-year time-series trend is -0.30 days/year, then it has become 15 days earlier during the observation period. This would mean that average values from the calendar would not be appropriate for describing the phenomena in the most recent times, which must be taken into consideration when deciding how to show the parameters. For example, the Pärnu

Figure 4.6-6. Phases with quartiles, based on the Schleswig station.

Chapter 4.6: Developing Comparative Phenological Calendars

309

Figure 4.6-7. Temporal distribution of slope values in phenological calendars at Viechtach and Pärnu. Significant values are marked with filled dots.

phenological calendar has significant trends for most of phenophases, with a tendency to shift earlier. Figure 8 shows averages from Pärnu calendars in the 1950s and 1990s. Differences are clearly visible, with spring becoming 1-3 weeks earlier between the periods. 3.1.6

Periods and amplitude of quasi-oscillations in long time series (important parameter, recommended minimum of 50 observation years).

Many climatic parameters and phenological phases have irregular oscillation cycles, which are usually described by presenting one or two of the most frequent oscillation periods. Oscillation amplitude is normally described with extreme years or lists of early-late values. Rhythm and length of oscillations can be diverse, and in temperate Europe, the most frequent periods are 2.5, 4, 8, 11, 14, 18, 22 years. Periods of those oscillations are related to different temporal cycles in the geosphere and solar system. As phenological time series available for study are usually only 30-50 years long, oscillations longer than 16-years cannot be studied. Figure 9 demonstrates periods and amplitudes of quasi-oscillations in phenological time series in two time series. The optimal choice of statistics and graphical solutions for presentation depends on the aims and objectives of the particular work, its research methods, data, and type of presentation.

310

Phenology: An Integrative Environmental Science

Figure 4.6-8. The difference in averages between 1950s and 1990s show the statistically reliable changes described by a linear trend.

Figure 4.6-9. Periods and amplitudes of quasi-oscillations in phenological time series. A) Periodogram of oscillations of blooming dates of Tussilago farfvara in Schleswig; B) modeled sinusoid for 1st principal component of Petula pendula leaf unfolding. This sinusoid describes 11.0% of variability and has high correlation with German stations.

Chapter 4.6: Developing Comparative Phenological Calendars

3.2

311

Calendars of Place

The most commonly used technique for analyzing seasonal differences among places and regions is cartographic. All values from phenological calendars can be mapped with rasters, isolines, or dots. In order to compile a map, homogenous basic data is needed that covers the same period, and is gathered with comparable methodologies. For larger territories, acquiring such data can be a complicated matter. Figure 1 was generated from dates representing the beginning of birch leaf unfolding, and Figure 10 uses beginning of lilac bloom dates in Europe. The direction of phase progression, and intervals between beginning dates are generalized by the maps, which give a rather good overview of the spatial sequence of coming phases and seasons, as well as regional differences. The ground-based observations used in these phenological maps cover an extensive area of Europe, and are well suited for mapping. Relatively flat terrain is more easily portrayed with less data, as areas with more complex relief require an elevation model for accurate interpolations (Malosheva 1968; White et al. 1997; Ahas and Aasa 2001). Satellite imagery is also helpful for compiling phenological maps, especially for larger areas or local differences that cannot be represented by existing ground data. Still, the data-series obtained from satellite observations are not long enough for complex time-series analysis, and the number and precision of phases detectable from satellite images is limited (Schwartz 1997; Schwartz et al. 2002).

Figure 4.6-10. The beginning of lilac bloom in Europe, 1951-1998.

312

Phenology: An Integrative Environmental Science

Besides cartographic presentation, length, width, and height gradients are also used to describe the “spatial spread” of phases. However, these methods are only useful in rather homogeneous conditions, like vast plains (Hopkins 1938; Schnelle 1955; Beruchashvili 1994). Besides gradients, phenological spatial relationships can be studied relative to degrees of longitude and latitude. Scatter plots of the beginning of birch leaf unfolding compared to longitude are given in Figure 11. Correlation with latitudinal distribution is higher (r = 0.78) than longitudinal (r = 0.61). This suggests that the relationships between space and beginning dates are not linear in all directions, but still highly interrelated in Central Europe. Comparing phenological calendars geographically offers more opportunities to study the spatial peculiarities of seasonality. Phenological spectra are an effective approach for such analyses, as they allow comparisons of beginning dates and intervals (duration). Spectra of climatic seasons at the observation stations are displayed in Figure 12a, and spectra of spring phenophases at the same stations are shown in Figure 12b. Spring phenomena begin earliest in northwestern Germany (beginning of coltsfoot bloom at Trier-Petrisberg is March 12th) and latest in eastern portions of the East-European plains (beginning of coltsfoot bloom at Kuznetskoye is April 19th). By the beginning of summer, inter-station phase timing differences decrease, shortening to only 5 days for the previously mentioned stations by the time linden trees starts blooming, and air temperature permanently crosses upward over 13°C. The Zhitomir region (hot and dry summer) and Trier-Petrisberg differ from other stations the most. Phases at the beginning of autumn and coloring of leaves are more stable than phases of late The growing period of autumn and beginning of winter. birch

Figure 4.6-11. The relation between the beginning date of birch leaf unfolding and latitude and longitude, based on European data from 1951-1998 (mapped in Figure 1).

Chapter 4.6: Developing Comparative Phenological Calendars

313

Figure 4.6-12. A) Beginning spectra of climatic seasons; B) beginning spectra of spring phenophases at all observation stations from 1951-1995.

(between leaf unfolding and end of leaf fall is rather heterogeneous among the stations studied. The shortest growing period of 106 days was found at the northernmost station (Borok), and the longest period (175 days) occurred at Trier-Petrisberg. Therefore, the difference in growing season length is 69 days, which is huge considering that the difference in latitude between the sites is only 6 degrees. Another method for studying geographical variation of seasonal phenomena is to present the curve of differences. In order to do this, the data from each observation site are standardized, with results presented on the graph as departures from the average. The curves of differences of climatic seasons for our eight observation sites are shown in Figure 13. The

Figure 4.6-13. The difference of climatic seasons from the average (standardized values). A) a group of stations in the west; B) a group of stations in the east; C) a transition group.

Observation sites are grouped according to three curve types: a) westernspring phases are earlier than the average and autumn phases are later than the average (at Schleswig and Weiz summer comes late); b) eastern - spring comes later and autumn earlier; and c) central - at Pärnu spring comes later and autumn at an average time, while at Zhitomir, spring comes at an

314

Phenology: An Integrative Environmental Science

average time, summer comes much earlier, and autumn at an average time. Such methods are very helpful for enhancing and grouping seasonal differences within regions. The differences are shown most clearly when as many phases (spread out over a long period) are used as possible.

3.3

Calendars of Years

There are a number of methods for comparing the seasonal rhythm of years. First, it must be stressed that phenological years are never identical, and there are relatively few years with completely comparable data. The most widespread method for comparing years (if they have similar data) is comparison of their phenological spectra. Averages of the early spring of 1989, late spring of 1955, and the mean values from 1951-1998 at Türi are shown in Figure 14a. From the comparison of those early and late years, it is clear that the rhythm of seasonality within a year can be quite variable. Summers and winters are more stable because of near constant solar radiation receipts; most seasonal phenomena during those periods have little interannual variability. In contrast, variety within the transition seasons is huge, with the highest standard deviations coming in early spring and late autumn. The inertia of environmental conditions also impacts future seasons. Seasonality of years can be classified through various methods. Classification into early or late season based on mean phase values and the resulting year “type” has been widely employed. The number of values greater than ±1 standard deviation is the simplest criteria for determining the phenological type of years. This value, presented as a percentage, is crucial for generalizing from phenological calendars, and determining what percentage of the total list of phases were earlier or later than ±1 standard deviation. Lack of identical data (differences in the list of phases and gaps) does not facilitate comparing years and places with simple methods. When cases with values that are smaller or larger than ±1 standard deviation are presented in percentages, many problems that result from listing the phases can be solved. The percentile distribution of extremes at Pärnu is displayed in Figure 14b. Cluster analysis is another common way to compare and group phenological calendars by years. Taking into consideration a relatively large number of observed years, lack of phases, and gaps (all of which are typical for phenological data), results must be interpreted carefully. When using cluster analysis to group years, it is reasonable to define a rather small number of clusters. The types of years can also be influenced by changes in large-scale quasi-oscillation cycles. Central European phenological time series commonly have 8±1 years cycles, consistent with the North Atlantic Oscillation (NAO) rhythm (Aasa 2002).

Chapter 4.6: Developing Comparative Phenological Calendars

315

Figure 4.6-14. A) 1955, 1989, and long-term average phenological spectra at Türi; B) The percentage of values larger or smaller than ± 1 standard deviation in Pärnu’s phenological calendar.

3.4

Seasons

An important issue when studying seasonal phenomena is determining and analyzing seasons. Seasons are periods with homogeneous conditions and characteristics, and their beginning and end are determined by certain phases. More generally, seasons are the intervals between two phases. Usually regional traditions or practices stemming from the research topic are used to determine seasons. In temperate zone phenology, there are two main seasons (summer and winter), and two transitional seasons (spring and autumn). In addition to these major types, there are two kinds of subdivisions. The first group of subseasons have certain characteristics, and usually occur every year (for example early spring and pre-winter). The second are dependent on qualitative parameters, and determined by characteristics that do not necessarily happen every year (for example midsummer and Indian summer). Seasons and their duration are important phenological measures that help to compare years and their variability. The most commonly used method of determining seasons is distinguishing according to a growing season temperature threshold, such as the period with air temperatures higher than 5ºC. A second method is to determine the growing season based on the period of active plant growth. For example, it is possible to declare the growing season of birch trees as the period from leaf unfolding to the end of leaf fall. In many Nordic countries, the agricultural growing season is often linked to the start and end of growth in autumn sown rye. Besides vegetation period, the lengths of spring and autumn are often characterized, due to their importance for determining the type of climate and features of the entire annual circle. For our Central European phenological calendar, we used temperature thresholds of 0ºC, 5ºC, and 13ºC for separating climatic seasons. Early spring begins when the

316

Phenology: An Integrative Environmental Science

daily mean temperature permanently increases above 0ºC, spring 5ºC, and summer 13ºC. The start of autumn corresponds to temperature falling below 13ºC, late autumn +5º, and winter 0ºC. In snow climates, the dates of formation and disappearing of permanent snow cover often determine winter seasons. Description of the techniques for estimating permanent crossings above or below certain temperature limits was presented in a previous study (Jaagus and Ahas 2000). Therefore, in this chapter we do not examine this issue in any detail. Further, the varied cultural traditions for defining seasons in different parts of Europe make it difficult to propose new ways of classifying these phenomena across the continent.

4.

CONCLUSIONS AND PERSPECTIVES

Phenological calendars allow description and comparison of the seasonal rhythm of observed species, places, and years. In order to do this, they employ different types of statistics, methods of analysis, and graphical presentation techniques. The main goal when compiling calendars is to present basic data with as much details as possible, so that multiple users can fit them to required applications. When using phenological calendars it is necessary to apply and develop methods that can help overcome the problems resulting from irregularities and data gaps. In this chapter, we have analyzed phenological data from the POSITIVE program’s European database. The results allow us to split the study’s stations into four groups: 1) Germany and surrounding stations, that have the earliest beginning dates, longest growing season, longest intermediate seasons, largest trends (up to 5 weeks earlier over last 50 years) and strong 8-year oscillation cycles; 2) the Baltic region, with maritime climate influences (late summer, late fall), strong trends, and visible 8-year cycle; 3) Ukraine, that has the earliest mean spring, longest summer; a late fall, and low standard deviations; and 4) the Russian plain, with the latest beginning dates, shortest growing season, continental climate with long winter and short summer, short intermediate seasons, no trends in eastern and southeast sub-regions, and fragments of 8- and 13-year oscillation cycles. The methods presented in this chapter were used in compiling phenological calendars of Estonia and Central Europe. There seems to be no need to standardize the format of phenological calendars, as the identity and approachability of basic data is most important. However, time-series enable researchers to gain the most information and present statistics of their own. Unfortunately, strict limitations resulting from data protection in many counties make basic data for public use rather limited. Even if available, it is seldom presented in detail due to publication space limitations.

Chapter 4.6: Developing Comparative Phenological Calendars

317

Issues of climate change have raised interest in phenological monitoring. Therefore, as new networks become operational and more researchers start to use this information, it is critical that observation program methods be standardized, so that data gathered from different regions and continents can be properly compared. Phenological calendars are one method that can be used to efficiently compile and compare these data. Therefore, it would be reasonable to create a common standard for calendars, regarding which statistics to present. However, in order to do so, we need to evaluate the needs of different user groups, and methods of different research groups and regions. Such comparisons will also address the need to standardize methods for reconstruction and interpolation of historical data series.

ACKNOWLEDGEMENTS The authors are grateful to all the participants of the POSITIVE program, especially to V.G. Fedotova from the Russian Association of Geography, and to all phenological observers in Europe who collected those valuable databases.

REFERENCES CITED Aasa, A., Trends in Eastern- and Central-European Phenological Calendars (in Estonian with English Summary), M.A. thesis, Institute of Geography, University of Tartu, 2002 Ahas, R., Long-term phyto-, ornitho- and ichthyophenological time-series analyses in Estonia, Int. J. Biometeorol., 42(3), 119-123, 1999. Ahas R. (Editor), Estonian phenological calendar, Publicationes Instituti Geographici Universitatis Tartuensis 90, 206 pp., 2001. Ahas, R., and A. Aasa, Impact of landscape features on spring phenological phases of maple and bird cherry in Estonia, Landscape Ecology, 16(5), 437-451, 2001. Ahas, R., J. Jaagus, and A. Aasa, The phenological calendar of Estonia and its correlation with mean air temperature, Int. J. Biometeorol., 44(4), 159-166, 2000. Beruchashvili, N. L., Seasonal dynamics of landscapes of the Earth, Communications of Russian Academy Sciences, 6, 24-40 (In Russian), 1994. Defila, C., Pzlanzenphänologische Kalender ausgewählter Stationen in der Schweiz / Calendriers phytophänologiques en Suisse 1951-1990, Beiheft zu den Annalen der Schweizerischen Meteorologischen Anstalt, 233 pp., 1992. Galahov, N. N., Study on structure of climatic seasons of a year, Izdatel´stvo Akademii Nauk SSSR, Moscow (In Russian), 162 pp., 1959. Gornik, W., Investigations into the Problems of Finding a Mean Value in Phenological Data Series, Arboreta Phaenologica, 47, 11-14, 1994. Hopkins, A. D., Bioclimatics - a science of life and climate relations, U.S. Dept. Agr. Misc. Publ. 280, Washington, 1938.

318

Phenology: An Integrative Environmental Science

Hopkins, A. D., and M. A. Murray, Natural guides to the beginning, length and progress of seasons, Acta Phaenologica, 2, 33-43, 1933. Hydrometeorological Printing House, Natural calendars of North Western USSR, Geographical Society of USSR, Leningrad (In Russian), 71 pp., 1965. Ihne, E., Über phänologische Jahreszeiten, Naturw. Wehchenschr., Berlin, 10 pp., 1895. Jaagus J., and R. Ahas, Space-time variations of climatic seasons and their correlation with the phenological development of nature in Estonia, Climate Res., 15(3), 207-219, 2000. Kalamees, K., Estonian mycophenological seasons, in Estonian phenological calendar, edited by R. Ahas, pp. 119-138, Publicationes Instituti Geographici Universitatis Tartuensis 90, 2001. Klaveness, D., and F. E. Wielgolaski, Plant phenology in Norway – a summary of past and present first flowering dates (FFDs) with emphasis on conditions within three different areas, Phenol. Season., 1, 47-61, 1996. Malosheva, T. C., Methodological manual for composition of phenological maps (in Russian), Leningrad, 59 pp., 1968. Menzel A., R. Ahas, E. Koch, K. Peter, and W. Lipa, Common European phenological database of POSITIVE, in EPN conference "The times they are a-changin'" Climate change, phenological responses and their consequences for biodiversity, agriculture, forestry, and human health 5-7 Dec., 2001, (book of abstracts), edited by A. J. H van Vliet, J. A. den Dulk, and R. S. de Groot, p. 88, Wageningen, 2001. Penuelas J., and I. Filella, Responses to a Warming World, Science, 294, 793-794, 2001. Raik, A., Climatic seasons in Estonia (in Estonian with summary in English), Acta et commentationes Universitatis Tartuensis, 144, p. 33-44, 1963. Schnelle F. Pflanzen-Phänologie, Akademishe Verlagsgesellschaft, Geest und Portig, Leipzig, 299 pp., 1955. Schultz G. E., General phenology (in Russian), Nauka, Leningrad, 186 pp., 1981. Schwartz, M. D., Spring Index Models: An Approach to Connecting Satellite and Surface I edited by H. Lieth and M. D. Schwartz, Phenology, in Phenology of Seasonal Climates I, pp. 23-38, Backhuys, Netherlands, 1997. Schwartz, M. D., B. C. Reed, and M. A. White, Assessing satellite-derived start-of-season measures in the conterminous USA, Int. J. Climatol., 22(14), 1793-1805, 2002. White, M. A., P. E. Thorton, and S. W. Running, A continental phenology model for monitoring vegetation responses to interannual climatic variability, Global Biogeochemical Cycles, 11(2), 217-235, 1997.

Chapter 4.7 PLANT PHENOLOGICAL "FINGERPRINTS" Detection of Climate Change Impacts Annette Menzel Department of Ecology, TU Munich, Freising, Germany

Key words:

1.

Plants, Bio-indicator, Climate change, Environmental links, Temperature

INTRODUCTION

Observation of phenological phases is probably the simplest way to track changes in the ecology of species in response to climate change. Other possible responses, such as altered species distribution, population sizes, and community composition, will be much harder and more expensive to detect (Walther et al. 2002). Thus, during recent years, phenology has received increasing attention as a bio-indicator for Global Change. “Snow drops as bearer of bad tidings” was, for example, the title of an article about climate change in a German newspaper in January 2002. Phenology seems to be an ideal climate indicator at regional, national, and international levels, because it is easily understood by the general public, allows the study of changes at a smaller scale, raises awareness of climate change issues, engages the public in the climate change debate, and reconnects people with their natural world (Sparks and Smithers 2002). However, the scientific community also welcomes phenology as a tool for global change research. Among others, the length of the growing season and phenological phases are proposed by the European Environment Agency as global change indicators. Recent reviews about observed phenological changes, e.g. Menzel and Estrella (2001), Walther et al. (2002), Menzel (2002), Sparks and Menzel (2002), and Root et al. (2003) summarize various indications of shifts in Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 319-329 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

320

Phenology: An Integrative Environmental Science

plant and animal phenology that have been reported for the boreal and temperate zones of the northern hemisphere. This review, however, will be confined to plant phenological changes, as responses of animal life are described in the chapters of Part 6.

2.

OBSERVED PLANT - FINGERPRINTS

The picture of observed changes is consistent: in the last three to five decades, spring phases, such as leaf unfolding and flowering, have advanced by 0.12 to 0.31 days per year in Europe, and 0.08 to 0.38 days per year in North America. Fewer data on autumnn phases exist (mainly in Europe), and autumn is delayed by 0.03 to 0.26 days per year in Europe. In order to compare recent reported results correctly we should distinguish between studies analyzing one phenophase at one site (not included here), several phenophases per season at one site (small signs in Plate 1), one phenophase in an observational network (medium signs) and several phenophases per season in a network (large signs). All data in Plate 1 are given as phenophase / site / network average trend in days / year. Mean advances of spring in the network of the International Phenological Gardens in Europe are -0.20 days/year (1959-1993, Menzel and Fabian 1999), -0.21 d/y (1959-1996, Menzel 2000) and -0.27 d/y (1969-1998, Chmielewski and Rötzer 2001). An advance of -0.19 d/y (1951-1998) is given by Defila and Clot (2001) for the Swiss phenological network, -0.12 d/y (1951-1996) and -0.16 d/y (1951-2000) by Menzel et al. (2001) and Menzel (2003) for the German phenological network. Flowering of lilac and honeysuckle in the western U.S. (1957/1968-1994, Cayan et al. 2001) has advanced by -0.15 and -0.35 d/y. Multi-site network data allow assessment of the spatial representativeness of changes. These data show evidence of spatial variability with differences among sites either on maps (Walkovszky 1998; Schwartz and Reiter 2000; Menzel and Estrella 2001; Menzel et al. 2001) or by statistical descriptions. In general, positive trends indicating a delayed onset of spring also occur, however with clearly smaller numbers. Only around one third of stations have statistically significant trends. Thus, analyses of network data clearly reveal phenological responses, although these are sometimes inhomogeneous due to local microclimate conditions, natural variation, genetic differences (not for the IPG, see Chapter 2.3) or other non-climatic factors. On the continental scale geographical differences are evident with delayed rather than earlier onset of spring phases prevailing in southeastern Europe (Menzel and Fabian 1999), advances in spring in Western and

Chapter 4.7: Plant Phenological "Fingerprints"

321

Central Europe compared to delays in Eastern Europe (Ahas et al. 2002), and notable differences between states in the USA (Plate 1, Schwartz and Reiter 2000). Variability within countries is not a simple regression on latitude or longitude (Germany, Menzel et al. 2001), and observed differences between regions in Switzerland may be also due to clear altitudinal gradients of changes (Defila and Clot 2001). Many studies analyzing multi-phased phenological data at a particular site also reveal clear advances of spring events in the recent decades, though with differences between species. Out of the flowering time series of 36 plant species in southern Wisconsin (1936-1998 incl. a 29-year gap, Bradley et al. 1999), 10 exhibit significant advances, 10 non-significant trends (eight earlier, two later) and 16 no trends, with all phenological times series (including bird arrival and first song) observed on average -0.12 d/y earlier. In the Washington, D.C. area, the trend of average first flowering times per year shows a significant advance of -0.08 d/y, from 100 angiosperms (89 have negative, 76 significant negative trends, 1970-1999, Abu-Asab et al. 2001). Flowering of 385 species in southern-central England (1954-2000, Fitter and Fitter 2002) is on average 4.5 days earlier in the 1990s than in the previous four decades, with 16% of the species significant earlier and only 3% significant later. Similar statistics are also reported by Peñuelas et al. (2002) for NE Spain, 1952-2000, with an average advance of flowering of -0.12 d/y and leaf unfolding of -0.33 d/y. / Other single site studies with fewer phases (Ahas 1999; Beaubien and Freeland 2000; Jaagus and Ahas 2000; Sparks and Manning 2000) also report advancing spring phases. Further, there are study sites where none of the time series show significant trends (budburst, floraison, and veraison of grapevine in the Bordeaux area (Jones and Davis 2000); leaf unfolding of Mountain birch and flowering of cowberry and cloudberry on the Kola Peninsula, Russia (Kozlov and Berlina 2002). Several studies report time of season differences with highest advances often in early spring, and notable advances of succeeding phenophases (e.g., Bradley et al. 1999; Defila and Clot 2001; Menzel et al. 2001). Sparks and Smithers (2002) suggest higher temperature changes in early season as the reason for this seasonal differentiation. However, Peñuelas et al. (2002) did not find seasonal differences at their study site in NE Spain (the Mediterranean region). Despite this seasonal variation, large inter-specific plant reactions are reported for example by Menzel and Fabian (1999), Abu-Asab et al. (2001), Defila and Clot (2001), Menzel et al. (2001), Fitter and Fitter (2002), Kozlov and Berlina (2002), and Peñuelas et al. (2002). Fitter and Fitter (2002) note that annuals are more likely to flower earlier that congeneric perennials, and insect-pollinated species more than wind-pollinated ones. Peñuelas et al.

322

Phenology: An Integrative Environmental Science

(2002) observed no differences among Raunkiaer life forms, or among plants of different origin. This generally distinct response of different species may affect the structure of plant communities by altered competition. Relatively few studies have examined end-of-season changes such as leaf color and leaf fall in plants, but at this time of year shifts seem to be less pronounced and show a more heterogeneous pattern. Trends of autumn phenophases of deciduous trees at neighboring stations often show contradictory signals. On average the end of season phases have been delayed by 0.03 to 0.16 days / year (Menzel and Fabian 1999; Menzel 2000; Chmielewski and Rötzer 2001; Defila and Clot 2001; Menzel et al. 2001; Menzel 2003). An extreme delay is detected by Peñuelas et al. (2002) at one site in northeastern Spain, where leaf fall was 13 days later in 2000 compared to 1952, resulting in a mean change of 0.26 days / year. Kozlov and Berlina’s study (2002) represents a counter example of a strong advance of birch first leaf fall (by 22 days) compared to the 1930s. These findings appear to contrast with the typical observation of a longer growing season in Europe and North America, also due to a later autumn. Although the authors report no evidence of direct pollution impact on phenological autumn phases, it remains doubtful whether (solely or predominantly) increasing severity of environmental pollution on the Kola Peninsula is reflected in this extreme advance of birch leaf fall. In this case, phenology may serve more as a bio-indicator of global change (including air pollution) than of “pure” climate change. Fruit ripening as an early autumn phase behaves differently. There is a clear temperature response, with warmer spring and summer advancing this phase. Several examples of advanced fruit ripening are significantly earlier harvest dates of grapevine in the Bordeaux area (Jones and Davis 2000), 9 days earlier fruiting in 2000 than 1974 in NE Spain (-0.35 days / year, Peñuelas et al. 2002), earlier fruit ripening of Sambucus nigra and Aesculus hippocastanum in Germany (-0.20 and -0.05 days / year in the last five decades, Menzel 2003). Again, the observed pattern of changes is different on the Kola Peninsula with 10 days later fruit ripening of cowberry than in the 1930s and no changes in cloudberry ripening (Kozlov and Berlina 2002). In total, the length of the growing season has increased by up to 3.6 days per decade over the last 50 years in Europe (Menzel and Fabian 1999; Menzel 2000; Chmielewski and Rötzer 2001; Defila and Clot 2001; Menzel et al. 2001; Walther et al. 2002; Menzel 2003). Almost twice the typical lengthening is reported by Peñuelas et al. (2002) with +0.67 days / year for the period 1952-2000.

Chapter 4.7: Plant Phenological "Fingerprints"

3.

323

INCONSISTENCIES

The use of phenology as a bio-indicator of climate change has the great advantage that these observations are extremely suitable to illustrate and communicate climate change impacts. However, there are also some inconsistencies that have to be taken into account when comparing reported results. Hughes (2000) mentioned that the nature of scientific publishing is such that papers showing advancing spring due to global warming are more likely to be submitted and published in scientific journals than others. In addition, the observed species and events themselves are subject of selection. Sparks and Smithers (2002) propose that species with strong temperature response and common widespread distribution, widely recognized and well loved species, which are found in rural and urban situations and have been recorded in the past, have been preferred for observations. In general, phenological observations are not an absolute measurement; they always depend on the observational efforts and skills of the observers, and thus are always subject to a certain degree of subjective inaccuracy (Menzel 2002). In addition, phenological records are often full of gaps, due to changes in observers (Defila and Clot 2001). When these gaps or observational errors occur at the beginning or the end of the record, they may involve some problems concerning the temporal representativeness of the observed changes. Shifts of phenological phases are determined by means of linear regression against time and are given as absolute change (days) during the period analyzed or as slope of the regression line (days per year). The main problem, also arising for other meteorological parameters, is that the underlying time period, especially the observations in the starting and ending year, strongly influence the resulting trend. For example, the increase in temperature and the advance of spring phases, respectively, are especially clear for time series ending in the warm years of the 1990s (1989 and later, Rapp 2002; Scheifinger et al. 2002). Figure 1 illustrates this fact by examining the recorded onset of flowering of P. avium at Geisenheim during the last 100 years. The linear trends for the last five decades, especially for the last 18 years (period compatible with NOAA satellites) are much stronger than for the whole century. Thus, in order to properly compare changes, trends should be given in days per year, together with the time period analyzed. Sagarin (2001) concluded that most off the studies would overestimate the reported advance of spring, because their phenological dates are given as calendar dates (CD) and not relative to the vernal equinox (VE). The magnitude of this bias would depend on the length of and period covered by the record, and the actual scatter of the data, and data sets spanning the 20th century would show the largest bias. However, the reported climate change

324

Phenology: An Integrative Environmental Science

impacts on phenology can be influenced only to a minor degree by this bias. Re-analysis of P. avium flowering data, corrected for the date after the VE, revealed that this bias involves only a change in the third decimal place of the trend and is well within one standard error of the trend coefficient.

Figure 4.7-1. Beginning of flowering of P. avium at Geisenheim (49°59’ N, 7°58’ E, data from the Deutscher Wetterdienst). For three different periods, the results of trend analyses for calendar dates (CD) as well as corrected for the date of vernal spring equinox (VE) are given with slope and standard error of the slope (days / year).

Moreover, the concerns about the impacts of climate change on nature via phenological shifts inevitably refer to the relation of two or more events and changed species interactions, such as altered competition by different leaf unfolding dates. Thus there is no bias due to shifting dates of the spring equinox when the analysis involves comparison of interrelated events. Other disturbing factors may be the aging of clones or trees under long-term observation (Menzel and Fabian 1999), effects of the urban heat island (Menzel and Fabian 1999; Rötzer et al. 2000), or the use of different cultivars and varieties in observation of agricultural crops (Menzel 1998).

4.

ENVIRONMENTAL LINKS

The confirmation of the close relationship between spring phenological phases and temperature is not needed to use phenological data as “proxies”

Chapter 4.7: Plant Phenological "Fingerprints"

325

for air temperature, because only few pre-instrumental phenological records are available. In fact, phenology can only communicate global change impacts to the public, if the relation between temperature change and shift of the timing of phenological events is demonstrated. In contrast to the climatic factors controlling autumn phenology, the climate signal driving spring phenology is fairly well understood. Nearly all phenophases correlate with spring temperatures in the preceding months (e.g. Sparks and Carey 1995; Sparks et al. 2000; Abu-Asab et al. 2001; Chmielewski and Rötzer 2001, 2002; Menzel 2003). Some spring events, such as the start of the plant growing season in northern and central Europe, also correlate with the North Atlantic Oscillation (NAO) index corresponding to winter climatic conditions (Post and Stenseth 1999; Ottersen 2001; Chmielewski and Rötzer 2001; Menzel 2003). Available models for spring phases constitute another intelligent possibility to demonstrate the close relationship between spring phases and temperature (see phenological modeling review in Chapter 4.1). The close relationship between spring phases and air temperature is also the basis of phenological spring indices for flowering and bud burst which are commonly used to describe phenological changes (e.g., Beaubien and Freeland 2000; Schwartz and Reiter 2000; Schwartz and Chen 2002). Autumn leaf coloring is considerably less directly explained by temperature. Two opposing influencing factors are detected, with warm late summers delaying leaf coloring, and higher temperatures in May and June Corresponding to temperature, NAO also promoting leaf coloring. influences the anomalies of phenological phases (Menzel 2003). The conclusion of Sparks and Smithers (2002) for the UK—that so far no sizeable effects of rainfall to phenology have been detected—might be true for the whole of Central Europe. However, in their study of Mediterranean vegetation, Peñuelas et al. (2002) found that phenological events of less drought tolerant species or non-irrigated agricultural plants correlated with precipitation. In the future (additional) photoperiodic control, CO2 effects, irrigation, fertilization, and farming practices cannot be disregarded (Menzel 1998; Peñuelas et al. 2002).

5.

COMPARISON TO OTHER FINGERPRINTS

Observed start of season changes are not restricted to plants. They also include changes in the timing of spring activities, include earlier breeding or first singing of birds, earlier arrival of migrant birds, earlier appearance of butterflies, and earlier choruses and spawning in amphibians. Thus various indications for shifts in plant and animal phenology have already been observed in the boreal and temperate zones of the northern hemisphere

326

Phenology: An Integrative Environmental Science

(Walther et al. 2002, for discussion of animal phenology see section 6). Phenological changes in birds and plants are often similar, as described in some cross-system studies (Ahas 1999; Bradley et al. 1999; Peñuelas et al. 2002). An example in Walther et al. (2002) shows that anomalies of leaf unfolding of trees, and arrival/hatching of birds exhibit parallel trends in Germany, with both having advancing tendencies since the 1990s; air temperature and the spring North-Atlantic Oscillation index (NAO) are negatively correlated to these phenophases. However, the timing of change in different taxonomic groups is not always synchronous, and may have profound ecological consequences. Major worries about the impacts of climate change as expressed by phenology concern this disharmonization of fine-tuned relations within ecosystems, leading to a decoupling of species interactions (e.g., review of examples by Walther et al. 2002). The lengthening of the growing season in recent decades, which is apparent from phenological “ground truth” r in mid- and higher latitudes of the Northern Hemisphere, is also found in other data sets, such as satellite images, CO2 records, and in the temperature records. Associated with an increase in the global mean temperature of about 0.6°C since the start of the 20th century, an asymmetry with stronger warming in daily minimum temperatures than in maximums, is detectable in all seasons and in most of the regions studied (Karl et al. 1993; IPCC 2001). Most of the numerous existing definitions of the climatological growing season use the dates when air temperature exceeds a threshold in spring and falls below in autumn. These meteorological measures also reveal lengthening of the warm season (e.g., Rapp and Schönwiese 1994) or of the ice-free season (Magnuson et al. 2000; Sagarin and Micheli 2001). Several studies have analyzed in detail the lengthening of the frost-free season in several countries of the northern hemisphere (e.g., Robeson 2002; Schwartz and Chen 2002; Scheifinger et al. 2003; Menzel et al., in press). However, geographical differences in climate change and corresponding plant responses are quite common. Kozlov and Berlina (2002) report a decline in the length of the snow-free and ice-free periods due to both delayed spring and advanced autumn/winter on the Kola Peninsula, Russia. Thus, regional studies of plant and animal phenology are extremely important as they can shed light on regional peculiarities. At the same time, these examples show that there are differences between climatological and phenological seasons, as a significant increase in snow precipitation may have caused the significant delay of the first thawed patches, whereas spring phases of the observed plants did not change in timing. However, multitemporal satellite data, such as NOAA AVHRR NDVI time series (see Chapter 5.1), also allow an assessment of the lengthening of the growing season in mid- and higher latitudes of the northern hemisphere (e.g.,

Chapter 4.7: Plant Phenological "Fingerprints"

327

Myneni et al. 1997; Tucker et al. 2001). A recent study by Zhou et al. (2001) found a lengthening of the growing season, both due to an earlier start and later end, of 18 ± 4 days for Eurasia, and 12 ± 5 days for North America based on normalized difference vegetation index estimates between 1982 and 1999 for 40-70°N. This extension of the growing season by 1.0 and 0.66 d / y respectively, exceeds the results from phenological “ground truth” especially due to larger changes in autumn. Both satellite and ground results are also well in accord with an advance of the seasonal cycle by –0.2 to –0.28 days/year for the last 2 to 3 decades and an increase in amplitude of the annual CO2 cycle since the 1960s derived from long-term measurements of CO2 concentration (Keeling et al. 1996).

REFERENCES CITED Abu-Asab, M.S., P. M. Peterson, S. G. Shelter, and S. S. Orli, Earlier plant flowering in spring as a response to global warming in the Washington, DC, area, Biodiversity and Conservation, 10, 597-612, 2001. Ahas, R., Long-term phyto-, ornitho- and ichthyophenological time-series analyses in Estonia, Int. J. Biometeorol., 42, 119-123, 1999. Ahas, R., A. Aasa, A. Menzel, V. G. Fedotova, H. and Scheifinger, Changes in European spring phenology, Int. J. Climatology, 22, 1727-1738, 2002. Beaubien, E. G., and H. J. Freeland, Spring phenology trends in Alberta, Canada: links to ocean temperature, Int. J. Biometeorol., 44, 53-59, 2000. Bradley, N. L., A. C. Leopold, J. Ross, and W. Huffaker, Phenological changes reflect climate change in Wisconsin, Proc. Natl. Acad. Sci. (USA), 96, 9701-9704, 1999. Cayan, D.R., S. A. Kammerdiener, M. D. Dettinger, J. M. Caprio, and D.H. Peterson, Changes in the onset of spring in the western United States, Bull. Amer. Meteorol. Soc., 82, 399-415, 2001. Chmielewski, F.M., and T. Rötzer, Response of tree phenology to climate changes across Europe, Agricult. Forest Meteorol., 108, 101-112, 2001. Chmielewski, F.M., and T. Rötzer, Annual and spatial variability of the beginning of growing season in Europe in relation to air temperature changes, Clim. Res., 19, 257-264, 2002. Defila, C., and B. Clot, Phytophenological trends in Switzerland, Int. J. Biometeorol., 45, 203-207, 2001. Fitter, A. H., and R. S. R. Fitter, Rapid changes in flowering time in British plants, Science, 296, 1689-1691, 2002. Hughes, L., Biological consequences of global warming: is the signal already apparent? Tree, 15, 56 – 61, 2000. Intergovernmental Panel on Climate Change, Climate Change 2001: Impacts, Adaptation, and Vulnerability, edited by J. J. McCarthy, O. F. Canziani, N. A. Leary, D. J. Dokken, and K. S. White, Cambridge University Press, 1032 pp., 2001. Jaagus, J., and R. Ahas, Space-time variations of climatic seasons and their correlation with the phenological development of nature in Estonia, Clim. Res., 15, 207-219, 2000. Jones, G. V., and R. E. Davis, Climate influences on grapevine phenology, grape composition, and wine production and quality for Bordeaux, France, Amer. J. Enology and Viticulture, 51, 249-261, 2000.

328

Phenology: An Integrative Environmental Science

Karl, T. R., P. D. Jones, R. W. Knight, G. Kukla, N. Plummer, V. Razuvayev, K. P. Gallo, J. Lindseay, R. J. Charlson, and T. C. Peterson, Asymmetric trends of daily maximum and minimum temperature, Bull. Amer. Meteorol. Soc., 74, 1007-1023, 1993. Keeling, C. D., F. J. S. Chin, and T. P. Whorf, Increased activity of northern vegetation inferred from atmospheric CO2 measurements, Nature, 382, 146-149, 1996. Kozlov, M., and N. Berlina, Decline in the length of the summer season on the Kola peninsula, Russia, Climatic Change, 54, 387-398, 2002. Magnuson, J. J., D. M. Robertson, B. J. Benson, R. H. Wynne, D. M. Livingstone, T. Arai, R. A. Assel, R. G. Barry, V. Card, E. Kuusisto, N. G. Granin, T. D. Prowse, K. M. Stewart, and V. S. Vuglinski, Historical trends in lake and river ice cover in the northern hemisphere, Science, 289, 1743-1746, 2000. Menzel, A., Zeitliche Trends ausgesuchter phänologischer Phasen in Deutschland aus dem Zeitraum 1951-1996, Unveröff, Bericht an den Deutschen Wetterdienst, 17 pp., 1998. Menzel, A., Trends in phenological phases in Europe between 1951 and 1996, Int. J. Biometeorol., 44, 76-81, 2000. Menzel, A., Phenology: its importance to the global change community, Climatic Change, 54, 379-385, 2002. Menzel, A., Plant phenological anomalies in Germany and their relation to air temperature and NAO, Climatic Change, 57, 243-263, 2003. Menzel, A., and N. Estrella, Plant phenological changes, in “Fingerprints” of Climate Change – Adapted behaviour and shifting species ranges, edited by G.-R. Walther, C. A. Burga, and P. J. Edwards, pp. 123-137, Kluwer Academic/Plenum Publishers, New York and London, 2001. Menzel, A., and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999. Menzel, A., N. Estrella, and P. Fabian, Spatial and temporal variability of the phenological seasons in Germany from 1951-1996, Global Change Biology, 7, 7 657-666, 2001. Menzel, A., G. Jakobi, R. Ahas, H. Scheifinger, and N. Estrella, Variations of the Climatological Growing Season (1951-2000) in Germany Compared to Other Countries, Int. J. Climatology, (in press), 2003. Myneni, R. B., C. D. Keeling, C. J. Tucker, G. Asrar, and R. R. Nemani, Increased plant growth in the northern high latitudes from 1981 to 1991, Nature, 386, 698-702, 1997. Ottersen, G., B. Planque, A. Belgrano, E. Post, P. C. Reid, and N. C. Stenseth, Ecological effects of the North Atlantic Oscillation, Oecologia, 128, 1–14, 2001. Peñuelas, J., I. Filella, and P. Comas, Changed plant and animal life cycles from 1952 to 2000 in the Mediterranean region, Global Change Biology, 8, 531-544, 2002. Post, E., and N. C. Stenseth, Climatic variability, plant phenology, and northern ungulates, Ecology, 80, 1322–1339, 1999. Rapp, J., and C. D. Schönwiese, "Thermische Jahreszeiten" als anschauliche Charakteristik klimatischer Trends, Meteorologische Zeitschrift, 3, 91-94, 1994. Rapp, J., Konzeption, Problematik and Ergebnisse klimatologischer Trendanalysen für Europa und Deutschland, Berichte des Deutschen Wetterdienstes 212, 2002. Robeson, S.M., Increasing growing-season length in Illinois during the 20thh century, Climatic Change, 52, 219-238, 2002. Root, T. L., J. T. Price, K. R. Hall, S. H. Schneider, C. Rosenzweig, and A. Pounds, Fingerprints of global warming on wild animals and plants, Nature, 421, 57-60, 2003. Rötzer, T., M. Wittenzeller, H. Haeckel, and J. Nekovar, Phenology in central Europe – differences and trends of spring phenophases in urban and rural areas, Int. J. Biometeorol., 44, 60-66, 2000. Sagarin, R., False estimates of the advance of spring, Nature, 414, 600, 2001.

Chapter 4.7: Plant Phenological "Fingerprints"

329

Sagarin, R., and F. Micheli, Climate Change in Nontraditional Data Sets, Science, 294, 811, 2001. Scheifinger, H., A. Menzel, E. Koch, C. Peter, and R. Ahas, Atmospheric mechanisms governing the spatial and temporal variability of phenological observations in central Europe, Int. J. Climatology, 22, 1739-1755, 2002. Scheifinger, H., A. Menzel, E. Koch and C. Peter, Trends of spring time frost events and phenological dates in central Europe, Theoretical and Applied Climatology, 74, 41-51, 2003. Schwartz, M. D., and B. E. Reiter, Changes in North American Spring, Int. J. Climatology, 20 (8), 929-932, 2000. Schwartz, M. D., and X. Chen, Examining the onset of spring in China, Clim. Res., 21, 157164, 2002. Sparks, T. H., and P. D. Carey, The responses of species to climate over two centuries: an analysis of the Marsham phenological record, J. Ecology, 83, 321-329, 1995. Sparks, T. H., and M. Manning, Recent phenological changes in Norfolk, Trans. Norfolk Norwich Nat. Soc, 33, 105-110, 2000. Sparks, T. H., and A. Menzel, Observed changes in seasons: an overview, Int. J. Climatology, 22, 1715-1725, 2002. Sparks, T. H., and R. J. Smithers, Is spring getting earlier?, Weather, 57, 157-166, 2002. Sparks, T. H., Jeffree, E. P., and C. E. Jeffree, An examination of the relationship between flowering times and temperature at the national scale using long-term phenological records from the UK, Int. J. Biometeorol., 44, 82-87, 2000. Tucker, C. J., D. A. Slayback, J. E. Pinzon, S.O. Los, R. B. Myneni, and M. G. Taylor, Higher northern latitude normalized difference vegetation index and growing season trends from 1982-1999, Int. J. Biometeorol., 45, 184-190, 2001. Walkovszky, A., Changes in phenology of the locust tree (Robinia ( pseudoacacia L.) in Hungary, Int. J. Biometeorol., 41, 155-160, 1998. Walther, G. R., E. Post, P. Convey, A. Menzel, C. Parmesan, T. J. C. Beebee, J. M. Frometin, O. Hoegh-Guldberg, and F. Bairlein, Ecological responses to recent climate change, Nature, 416, 389-395, 2002. Zhou, L., C. J. Tucker, R. K. Kaufmann, D. Slayback, N. V. Shabanov, and R. B. Myneni, Variations in northern vegetations activity inferred from satellite data of vegetation index during 1981 to 1999, J. Geophysical Research, 106(D17), 20,069-20,083, 2001.

Chapter 4.8 PHENOCLIMATIC MEASURES Assessment of the Onset of Spring Mark D. Schwartz Department of Geography, University of Wisconsin-Milwaukee, Milwaukee, WI, USA

Key words:

1.

Models, Climatology, Synoptic, Spring, Indices, Lilacs

INTRODUCTION

Interaction between the atmosphere and biosphere is a crucial area of study for improving the accuracy of plant development models at all scales, and increasing knowledge of critical exchanges in the planetary carbon balance. Phenology, the study of biological life cycle events (such as first leaf, first bloom, and senescence events in plants, hatching of insects, or first appearance of birds and small mammals), driven by environmental factors (primarily temperature), can serve as an important framework to facilitate such studies (Schwartz 1994, 1999). The following sections summarize a program that has contributed to this research area through: 1) development and refinement of phenological simulation models, optimized for continental-scale studies, and providing a ready means to process climate data into a form comparable with satellite sensor-derived and conventional phenological data; and 2) evaluation of the effects of springtime plant development on energy exchange, mass exchange, and measured characteristics of the lower atmosphere. Results from this work have: 1) provided spring phenological models and a suite of associated measures (derived from daily maximum-minimum temperature data) that allow a first assessment of the possible impacts of climate change on mid-latitude phenology at the local, regional, and global Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 331-343 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

332

Phenology: An Integrative Environmental Science

scale; 2) increased understanding of the magnitude and impact of spring plant growth initiation in mid-latitude seasonal climates on lower atmospheric temperature, moisture, sensible-latent heat balance, and carbon balance, and 3) produced measures of phenological development over large areas that can be compared to and provide a means to assess the accuracy of satellite sensor-derived phenological products (see Chapter 5.1).

2.

CONTINENTAL-SCALE PHENOLGICAL MODELS

2.1

Conceptual Issues

Traditional simple models of phenological development are constructed based on measuring degree-day accumulations (see Chapters 4.1, 4.9, and 6.2) from some arbitrary or calculated starting date up till the time of the phenological event (stage) being studied. Several assumptions are implicit in this approach: 1) degree-day accumulations get treated like “water filling a bucket,” in that no consideration is usually given to the manner in which they add up, only the time that the total passes a predetermined threshold is regarded as important; and 2) plant-climate relationships are deemed to be highly site-specific, such that regional or continental-scale phenological models are not practical. While several earlier researchers (Wang 1963, Caprio 1974) had previously taken steps toward challenging these assumptions, a more comprehensive reevaluation was begun in the mid1980s (Schwartz 1985; Schwartz and Marotz 1986, 1988). Specifically, the timing and magnitude of periods of warm air advection (driven by synopticscale weather systems) was tested as a method of model variable selection, and models were developed based on data from stations distributed over a large (3000 x 2000 km) region in eastern North America. Another principle that was adhered to in model development was simplicity of input data. The goal was to produce the best models possible using data that would likely be available for continental-scale and ideally global-scale applications. Thus, daily maximum-minimum temperatures were selected as the primary input data.

Chapter 4.8: Phenoclimatic Measures

2.2

Model Development

2.2.1

Database issues

333

Several characteristics of phenological data and common limitations of the information collected by phenological networks are important factors in model feasibility and design. First, natural plants of the same species show genetic variability across their ranges, that affects their phenological responses (Nienstaedt 1974). For example, if red oak (Quercus rubra) trees from across eastern North America were all gathered and grown at a common location, trees from northern regions would achieve spring budburst earlier each year than their southern-origin “cousins,” as they have adapted to growing in a region with less solar-thermal energy available (Schlarbaum and Bagley 1981). Thus, a continental-scale phenological monitoring network using only native species would produce a “mixed signal” of genetic variability and environmental factors. Many continentalscale phenological networks have accounted for this problem by using varietal clones at all locations, thus minimizing genetic effects on phenology (see examples of these types of network designs in Chapters 2.3 and 2.4). Such cloned data are well suited for developing general continental-scale phenological models, but these responses must still be compared back to native species for maximum utility. So for the broadest application, networks should ideally be designed to include both a selection of appropriate native species, and a few wide-ranging cloned species (Schwartz 1994, 1999). A second issue is the typical nature of network station data. For example, cloned lilac and honeysuckle data have been collected from several hundred stations in eastern North America, and several thousand stations in the western USA since the late 1950s (Schwartz 1994). However, only a small fraction (about 10%) of these stations have periods of record extending 20 years or more. Thus, direct comparison of long-term phenological monitoring is only possible at a few sites from the raw data. With a continental-scale phenological model, a first assessment of phenological changes can be made comprehensively (in all areas where daily maximumminimum temperature data needed to drive the model are available), providing a baseline for assessing phenological change on a global scale, and in comparison to available native species data. Another approach to compensate for incomplete network data is using phenological model output and satellite-derived data in concert with native species data to help reconstruct phenological time-series. Zhao and Schwartz (2003) applied this strategy to data collected by the Wisconsin Phenological Society in the state of Wisconsin over the 1965-1998 period

334

Phenology: An Integrative Environmental Science

(see Chapter 2.4 for more information about this network). These data (mostly bloom dates for native wild flowers) were extensive in total, but few individual stations had records longer than ten years. Therefore, satellitederived start of season (SOS) dates were used to produce an initial regional assessment of phenological response across the state (Reed et al. 1994). The coherence of these regions was then tested using a phenological model (Spring Indices, Schwartz 1997). The result was three phenological regions within the state where all phenological observations could subsequently be treated as if from one site. Additional “boot strapping” techniques were used to further collapse and fill remaining data voids. One of the prime results of this study were three to four native species indices (distributed from early to late spring) for each phenological region, which could be examined over the study period for changes. “Early” spring in the southwestern portion of Wisconsin appeared to have the most significant advancement (Zhao and Schwartz 2003). 2.2.2

Model construction, evaluation, and refinement

Schwartz (1985) and Schwartz and Marotz (1986) developed and tested continental-scale phenological models based on cloned lilac (Syringa chinensis ‘Red Rothomagensis’) first leaf event data recorded in eastern North America. A traditional degree-day accumulation model was compared to a “synoptic control model” where the timing of phenological events was deemed to be related to the number and intensity of synopticscale weather systems affecting a location, especially those with well developed warm air advection. In both model forms, the phenological data served as the dependent variable in a regression equation expressed as an arbitrary constant (1000) divided by the actual phenological date (day of the year, January 1st = 1). Independent variables (such as GDH accumulation at the time of the phenological event) were all divided by the phenological date. The equation thus takes the form: (1000 / date) = Regression Constant + (A1*X1/date) + . . . (An*Xn/date) (1) This formulation ensured that different dates for the same phenological event, at different stations and during different years, did not bias the predictive capabilities of the technique. As subsequently elaborated by Schwartz et al. (1997) this transformation allows the multiple regression procedure to optimize the coefficients (A1-n) for the threshold variability of the independent variables, since they are each “standardized” by the date of the phenological event. Now, by simply multiplying through by the date of the phenological event, weather sequence variability can be reintegrated into

Chapter 4.8: Phenoclimatic Measures

335

the equation. Thus, the coefficients and regression constantt determined from (1) can be assimilated into an equation form more efficient for error testing and subsequent operational prediction: 1000 = (Regression Constant*Date) + (A1*X1) + (A2*X2) + . . . (An*Xn) (2) In this form, when an array of daily maximum-minimum temperatures are “processed” with the model, the output grows in size until it reaches value 1000 at the time of the predicted phenological event. Thus, one “best” prediction is provided for every case (Schwartz et al. 1997). In the traditional model form, variables including conventional degreeday accumulation, and hours accumulated in selected temperature ranges, to account for possible non-linear plant responses to temperature were calculated after a January 1st start date (plant chilling requirements were assumed to be satisfied at that time). For the synoptic model form, degreeday accumulation totals over a three-day period were initially calculated from January 1st until the time of first leaf during all station years, and then descriptive statistics were derived from these totals, such that the threshold for values one standard deviation above the mean was calculated for the entire study area. This threshold then served as the proxy for “high energy” synoptic events, which were simply counted from January 1st till the time of first leaf. The results demonstrated that the synoptic model form was a better predictor than the traditional form (7.5 days mean absolute error compared to 8.0 days), and that specific synoptic events (representing less than 20% of the total days in spring) exert as much or more influence on the arrival of spring (first leaf emergence) than annual progressions in energy received brought about by changing earth-sun geometry (Schwartz and Marotz 1986). Schwartz and Marotz (1988) expanded the synoptic model form to include testing for a potential “capstone effect” wherein counting the number of high-energy synoptic events is coupled with looking at preferred times for their occurrence in the week before the first leaf event. Surface synoptic classification showed that warm air advection, associated with the passage of a low pressure system was common about five to seven days before and at the time of first leaf, and this cycle was also supported by Fourier analysis of degree-day accumulations over the spring season. This “synoptic + capstone” model form decreased average prediction mean absolute error by about one day (to 6.6 days). However, the influence of synoptic weather systems on phenological development appeared to change regionally as the season progressed, being critically important in the south early, of moderate influence over most of the study region during the middle, and of minimal

336

Phenology: An Integrative Environmental Science

importance in the far north late in the spring season (Schwartz and Marotz 1988). Schwartz (1997) reexamined model construction, and produced a “second generation” of models that included predictions for the timing of first bloom as well as first leaf phenological events. One of the most noteworthy changes was an explicit accounting for chilling requirements. Earlier model forms had used a January 1st start date, assuming that chill was satisfied by that date, but not evaluating the validity of the assumption in any way. While it is generally understood that deciduous species need exposure to a period of cold temperatures before they can respond effectively to springtime warmth, no definitive research has established what those chill requirements are for the lilac (Syringa chinensis ‘Red Rothomagensis’) and honeysuckle ((Lonicera tatarica ‘Arnold Red’ and L. korolkowii ‘Zabeli’) indicator species, or even a definitive way to measure chill accumulation. Several “boot strapping” approaches were evaluated, and chill hours proved superior to chill units for these species and modeling approach (Richardson et al. 1974). A combination of previous techniques was used to develop working estimates of the chill requirements for each indicator plant. Daily running chill accumulation was plotted against the growing degree hours (GDHs, measuring heat accumulation) that would accumulate for that day until the time of the phenological event. These values were then stratified for all cases in 25 unit “wide” groups. The standard error of the GDH values within each group was subsequently calculated and plotted. The chill requirement was the chill value with the minimum standard error. For the lilac this was 1375 chill hours, and 1250 for both species of honeysuckle (base temperature of 7.2°C). The GDH accumulation (base temperature of –0.6°C) portions of the new models used the same “synoptic capstone” form as in previous versions. Models were evaluated by randomly splitting the database in half, using one part to develop the multiple regression coefficients, and the other half to assess error. Mean absolute errors for the first leaf models were 6.0 days for the Arnold Red honeysuckle, 6.2 days for the lilac, and 6.5 days for the Zabeli honeysuckle. First bloom model used a traditional “bucket filling” approach, and had errors of 4.2 days for the lilac and Zabeli honeysuckle, and 4.4 days for the Arnold Red honeysuckle.

2.3

Spring Indices

A useful component of evaluating change is a standard measurement tool. One of the reasons for using indicator plants in phenological studies is not to represent all species, but to provide a standard reference that can be

Chapter 4.8: Phenoclimatic Measures

337

compared among different sites and years. Ideally, models based on such plants will also be readily connected to sequences of weather events and changes in characteristics of the lower atmosphere. Any such variables proposed as “spring indices” would need to be thoroughly tested before adoption, and could involve a suite of related measures. Schwartz (1990) proposed a “first approximation of a comprehensive spring index” based on an averaging of the modeled dates of first leaf models for the lilac and two species of honeysuckle indicators. The proposed index showed the best ensemble relationship to an array of other measures, including last frost date, and “break-point” changes in the sequence of surface maximum temperatures and lower atmospheric lapse rates. Schwartz (1993) used a first leaf model-derived “spring index” to reconstruct a representation of past phenological responses in eastern North America over the 1908-1987 period. Last –2.2°C frost date was also examined, and a new measure, the “damage index” (spring index first leaf date minus last frost date) was calculated to assess the risk of frost damage to spring plant development. The long-term (1961-1987 average) pattern of first leaf date and its standard deviation were examined spatially, with central Illinois-Indiana and the southeast showing the greatest variability. Also, four distinct patterns: 1) earlier-than-normal over the entire region, 2) early dates in the south/southeast and late dates in the northwest, 3) laterthan normal over the entire region, and 4) late dates in the south and early dates in the northwest and north-central portions of the study area were examined. These patterns appeared to run for periods of three to six years, with no apparent temporal sequence. The damage index showed an interesting pattern of moderately negative values in the west (last frost occurring about 10 days after first leaf) highly negative values in the southeast (last frost occurring 20 to 35 days after first leaf) and the far northeast where last frost occurred before first leaf. No strong long-term trends were present in the indices. When a slightly longer period of record was examined in Schwartz (1998) a short-term change in spring index first leaf (toward earlier dates) was detected over the 1978 to 1990 period, which corresponded to satellite-derived data (Myneni et al. 1997). A “spring indices suite of measures” was introduced by Schwartz and Reiter (2000) to evaluate 1900-1997 springtime changes in North America, and to parallel Menzel and Fabian’s (1999) exploration of Europe’s changing growing season over the 1959-1993 period. This study also incorporated the “second generation” first leaf and first bloom models developed by Schwartz (1997). The suite includes eight measures: 1. First –2.2°C frost date in autumn

338

Phenology: An Integrative Environmental Science

2. Composite chill date—the average date when winter cold requirements among the three indicator plants (lilac and two honeysuckle species) have been satisfied, meaning they are ready to respond to spring warmth 3. Spring indices first leaf—“early spring,” the average date among the three indicator plant models that leaves grow beyond their winter bud tips, related to a general onset of growth in grasses and shrubs 4. Spring indices first bloom—“late spring,” the average date among the three indicator plant models that flowers start to open, related to a general onset of growth in dominant forest vegetation 5. Last –2.2°C frost date in spring 6. Frost period, the number of days from first frost date in autumn to last frost date in spring 7. Damage index value, the difference in days between the first leaf date and last frost date, indicating the relative internal timing of spring and potential for plant damage in a given year 8. Average annual temperature, and all twelve average monthly temperatures for comparative purposes The results showed that spring indices (SI) first leaf date advanced 5.4 days, SI first bloom date 4.2 days, and last frost date 4.2 days toward earlier arrival over the 1959-1993 period, which compared favorably with the 6.0 day advance for the same period reported in Europe (Menzel and Fabian 1999). The strongest regional patterns of change were in northwestern USAsouthwestern Canada, and northeast USA-Canadian Atlantic provinces. The lack of autumn phenology data made direct comparison difficult, but unlike Europe, North America does not seem to be getting warmer in that season. Rather, some areas in the central continent are cooling and enduring earlier frosts. Unlike North America and Europe, China does not seem to show a longterm trend toward earlier onset of spring. Schwartz and Chen (2002) used a methodology similar to Schwartz and Reiter (2000) applied to China. They first tested the SI models, and found them as accurate as in North America, when compared to actual lilac phenology data. Last frost dates in spring were getting earlier (by 6.0 days), and first frost dates in autumn later (by 3.9 days) over the 1959-1993 period. Thus, Schwartz and Chen (2002) reported a lengthening of the frost-free period of about 10 days, with the greatest changes concentrated in many northern regions of China. Based on the positive results of applying the SI models in China, and similar (but unpublished) comparisons in Estonia and Germany, the SI models appear robust enough to be applied to all compatible climate zones in mid-latitudes. At this time a project is underway to calculate the SI suite of measures for the greatest number and longest period of record climate stations in all regions of the mid-latitudes where the indicator lilac and honeysuckle plants

Chapter 4.8: Phenoclimatic Measures

339

can theoretically survive. Changes over the 1961-2000 period will be compared among all regions. The results will provide a valuable assessment of past changes and a baseline for future phenological studies.

3.

SPRING PLANT-CLIMATE INTERACTIONS

3.1

Effects on Conventional Climate Data

When plants break dormancy and foliage reappears (the “green-up”) in spring, a rapid increase in transpiration and changes in surface albedo alter thermodynamic properties of the surface layer. Normally, seasonal variations in tropospheric characteristics and year-to-year variability in the green-up date (which can range up to 1.5 months) mask the impact of these effects on surface maximum temperatures. Schwartz and Karl (1990) showed that these changes could be detected when lilac (Syringa chinensis “Red Rothomagensis” clone) phenology data is used as the indicator of transpiration onset. An important first step was to arrange the data relative to a phenological time scale. The difference between calendar date and lilac first leaf was computed for every daily observation. These difference values (number of days after first leaf) were then used for all further examination and manipulation of the data. Simple plots using average hypsometric layermean temperature and average surface daily maximum temperature were examined for variations in the thickness-maximum temperature relationship relative to first leaf date. For the same thickness value, stations in agricultural inland areas showed at least a 3.5°C reduction in surface daily maximum temperatures over any two-week period subsequent to first leaf compared to a two-week period prior to first leaf. Values were smaller for stations near major water bodies (1.5°C reduction). Schwartz (1992) continued this work and showed that the “green wave” (onset of leafing or “green-up” as indicated by cloned lilac first leaf and modeled first leaf dates) occurs in conjunction with discontinuities (signals) in lower atmospheric lapse rate, surface vapor pressure, relative humidity, visibility, and V (south-north) wind component. These signals are all consistent with the start of plant photosynthetic activity, seasonal shifts in atmospheric circulation patterns, and physical changes in the nature of the surface layer. Based on the results of that study, Schwartz (1992) proposed a plausible springtime sequence of events and the expected impact of vegetation green-up on surface-layer thermodynamic properties. In that scenario, air/soil temperatures first begin to rise rapidly with decreasing snow cover/dropping albedo. Since the rise is large compared to the increase in air moisture, relative humidity starts to decrease rapidly. As

340

Phenology: An Integrative Environmental Science

relative humidity drops, water droplet evaporation in the air increases, and the drops grow smaller, improving visibility. Now, when plants begin to actively transpire (with foliage appearance), albedo and latent heat increase. This leads to less sensible heat (energy to heat the air), and more moisture in the air. In response, surface temperatures slow their rate of increase. Relative humidity, responding to the changes in temperature and moisture, abruptly reverses direction and increases, while visibility stops improving and levels off. Fitzjarrald et al. (2001) developed an effective method for identifying the timing of this abrupt shift in lower atmospheric conditions, and the intensity of change over the spring season, using commonly available daily maximum-minimum temperatures. Spring intensity appears to be related to a variety of factors, including long-term changes in forest cover fraction. The diurnal temperature range (DTR, difference between daily maximum and minimum temperature) shows a discontinuity at the onset of spring characterized by a rapid increase for several weeks, followed by an abrupt leveling off (Figure 1). The trend then remains essentially flat for the remainder of the warm season. Ruschy et al. (1991) concentrated on changes in the DTR due to variations in snow and cloud cover. Karl et al. (1993) did a comprehensive analysis of DTR changes in all seasons, adding evidence that cloud cover ceiling height and percentage coverage may also play a role. Schwartz (1996) performed a detailed examination of the spring DTR discontinuity and found that several factors (snow cover loss, more frequent southerly winds, and increased ceiling heights) were responsible for the initial rapid increase in spring DTR. The subsequent leveling off was connected with increased atmospheric moisture and coincides with the onset of plant transpiration (as measured by first leaf of cloned lilacs). These results further defined the onset of spring in mid-latitudes as a modally abrupt rather than gradual seasonal transition.

3.2

Effects on Energy Balance and Carbon Balance

Changes in lower atmospheric characteristic during spring are ultimately related to variations in the surface energy balance, and have implications for carbon dioxide flux (net ecosystem exchange) and balance. In recent years, direct measurement of these variables at eddy-covariance sites in the Ameriflux (western hemisphere) and Fluxnet (worldwide) networks have made more extensive studies of phenological interactions possible. Wilson and Baldocchi (2000) showed that leaf emergence was one of the important factors that controlled seasonal changes inn latent and sensible heat fluxes at a broadleaved temperate deciduous forest site in North America (near Oak

Chapter 4.8: Phenoclimatic Measures

341

Ridge, Tennessee, USA). Extensions of this work are discussed in Chapter 7.2.

Figure 4.8-1. Average diurnal temperature range variations (data subset where first leaf date followed last snow date by at least 14 days, n = 150, 62% of all available station-years) for twelve locations in eastern North America, 1961-1986, relative to Syringa chinensis “Red Rothomagensis” (cloned lilac) first leaf date. Every seven days, ± 1 standard error bars are shown, and average dates of last –2.2°C freeze and snow-on-the-ground (“+” symbols) are also shown for comparison (adapted from Schwartz 1996).

Schwartz and Crawford (2001) explored phenological relationships to energy and carbon flux at several Ameriflux sites in eastern North America. Ongoing work is expanding that study by adding additional stations and years to the analyses. Results to date show that net ecosystem exchange (NEE, carbon dioxide flux), latent minus sensible heat flux (best reflecting the relative change in these two variables), and net radiation all show distinct changes relative to Spring Indices’ (SI) first bloom date (measure of phenological development, Schwartz and Reiter 2000). Two aspects are most important: 1) that the patterns of change are quite similar, despite different land cover types and geographical locations of the sites, and 2) existing differences between the sites can be logically explained by the

342

Phenology: An Integrative Environmental Science

differences in land cover type. For example, when arranged in phenological time, using SI values, northern deciduous forest and grassland sites show the earliest response, middle forest sites show intermediate response timing, and southern deciduous sites show the latest response. Thus, phenology is exhibiting considerable power in organizing the aggregated carbon and energy flux responses of diverse stations in the spring, and is even able to reflect different rates of Spring NEE (carbon flux) change on an annual basis at the different sites. Additional research is needed to further confirm these relationships.

4.

FUTURE RESEARCH

Many additional questions, such as the precise nature of regional variations in native species phenological responses, and the potential of phenological measurements to assist in refinement of carbon flux assessment, remain to be addressed. Future research into plant-climate interactions at the onset of spring can be greatly enhanced by expanded collection of standardized phenological data worldwide. This will require cooperation and collaboration among existing phenological data networks (such as those in Europe, see Chapters 2.3 and 2.7), establishment (or reactivation) of national phenology networks in North America and Asia, and implementation of the global phenological monitoring (GPM) plan (see Chapter 2.6) and proposed phenology protocols at Fluxnet sites around the world (http://www.uwm.edu/~mds/Pheno_phases_Fluxnet_proposed.pdf).

REFERENCES CITED Caprio, J. M., The Solar Thermal Unit Concept in Problems Related to Plant Development and Potential Evapotranspiration, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 353-364, Springer-Verlag, New York, 1974. Fitzjarrald, D. R., O. C. Acevedo, and K. E. Moore, Climatic consequences of leaf presence in the eastern United States, J. Climate, 14, 598-614, 2001. Karl, T. R., P. D. Jones, R. W. Knight, G. Kukla, N. Plummer, V. Razuvayev, K. P. Gallo, J. Lindseay, R. J. Charlson, and T. C. Peterson, A new perspective on recent global warming: Asymmetric trends of daily maximum and minimum temperature, Bull. Amer. Meteor. Soc., 74, 1007-1023, 1993. Menzel A., and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999. Myneni, R. B., C. D. Keeling, C. J. Tucker, G. Asar, and R. R. Nemani, Increased plant growth in the northern high latitudes from 1981-1991, Nature, 386, 698-702, 1997. Nienstaedt, H., Genetic variation in some phenological characteristics of forest trees, in Phenology and Seasonality Modeling, edited by H. Lieth, pp. 389-400, Springer-Verlag, New York, 1974.

Chapter 4.8: Phenoclimatic Measures

343

Reed, B.C., J.F. Brown, D. VanderZee, T.R. Loveland, J.W. Merchant, and D.O. Ohlen, Measuring phenological variability from satellite imagery, J. Vegetation Science, 5, 703714, 1994. Richardson, E. A., S. D. Seeley, and D. R. Walker, A model for estimating the completion of rest for ‘Redhaven’ and ‘Elberta’ peach trees, Hortscience, 9, 331-332, 1974. Ruschy, D. L., D. G. Baker, and R. H. Skaggs, Seasonal variation in daily temperature ranges, J. Climate, 4, 1211-1216, 1991. Schlarbaum, S. E., and W. T. Bagley, Intraspecific genetic variation of Quercus rubra L., northern red oak, Silvae Genetica, 30, 50-56. Schwartz, M. D., The Advance of Phenological Spring Across Eastern and Central North America, Ph.D. dissertation, Dept. of Geography, University of Kansas, Lawrence, 1985. Schwartz, M. D., Detecting the Onset of Spring: A possible application of phenological models, Clim. Res., 1(1), 23-29, 1990. Schwartz, M. D., Phenology and Springtime Surface Layer Change, Mon. Wea. Review, 120(11), 2570-2578, 1992. Schwartz, M. D., Assessing the Onset of Spring: A Climatological Perspective, Phys. Geography, 14(6), 536-550, 1993. Schwartz, M. D., Monitoring global change with phenology: the case of the spring green wave, Int. J. Biometeorol., 38, 18-22, 1994. Schwartz, M. D., Examining the Spring Discontinuity in Daily Temperature Ranges, J. Climate 9(4), 803-808, 1996. Schwartz, M. D., Spring Index Models: An Approach to Connecting Satellite and Surface Phenology, in Phenology of Seasonal Climates II, edited by H. Lieth and M. D. Schwartz, pp. 23-38, Backhuys, Netherlands, 1997. Schwartz, M. D., Green-wave phenology, Nature, 394(6696), 839-840, 1998. Schwartz, M. D., Advancing to full bloom: planning phenological research for the 21st century, Int. J. Biometeorol., 42, 113-118, 1999. Schwartz, M. D., and X. Chen, Examining the Onset of Spring in China, Clim. Res., 21(2), 157-164, 2002. Schwartz, M. D., and T. M. Crawford, Detecting Energy-Balance Modifications at the Onset of Spring, Phys. Geography, 21(5), 394-409, 2001. Schwartz, M. D., and T. R. Karl, Spring Phenology: Nature's Experiment to Detect the Effect of “Green-up” on Surface Maximum Temperatures, Mon. Wea. Review, 118(4), 883-890, 1990. Schwartz, M. D. and G. A. Marotz, An approach to examining regional atmosphere-plant interaction with phenological data, J. Biogeography, 13, 551-560, 1986. Schwartz, M. D., and G. A. Marotz, Synoptic Events and Spring Phenology, Phys. Geography, 9(2), 151-161, 1988. Schwartz, M. D., and B. E. Reiter, Changes in North American Spring, Int. J. Climatology, 20(8), 929-932, 2000. Schwartz, M. D., G. J. Carbone, G. L. Reighard, and W. R. Okie, Models to Predict Peach Phenology from Meteorological Variables, HortScience, 32(2), 213-216, 1997. Wang, J. Y., Agricultural Meteorology, Pacemaker Press, 693 pp., 1963. Wilson, K. B., and D. D. Baldocchi, Seasonal and interannual variability of energy fluxes over a broadleaved temperate deciduous forest in North America, Agricult. Forest Meteor. 100, 1-18, 2000. Zhao, T., and M. D. Schwartz, Examining the Onset of Spring in Wisconsin, Clim. Res., 24(1), 59-70, 2003.

Chapter 4.9 WEATHER STATION SITING Effects on Phenological Models Richard L. Snyder1, Donatella Spano2, and Pierpaolo Duce3 1

Department of Land, Air, and Water Resources, University of California, Davis, CA, USA; Department of Economics and Woody Plant Ecosystems, University of Sassari, Sassari, Italy; 3Agroecosystem Monitoring Laboratory, Institute of Biometeorology, National Research Council, Sassari, Italy 2

Key words:

1.

Temperature, Measurement, Standard surface, Heat units, Sensors

INTRODUCTION

The collection of accurate temperature data is critically important for phenological model development as well as for predictions when a model is being used. This is true everywhere, but especially in an arid climate where advection and surface wetting (i.e., by rainfall or irrigation) can affect the temperature measurements. In fact, the use of bad temperature data can lead to errors as large as differences resulting from climate change. Therefore, accurate temperature measurement is critical for both the application and development of phenological models. Temperature readings are affected by the local energy balance, so wetting frequency by irrigation or rainfall can affect readings. Therefore standardization and management of the underlying surface is critical to obtain useful models that can be universally applied. In this paper, the history and proper measurement of temperature is discussed including sensors, shielding, weather station siting, and surface management.

Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 345-361 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

346

2.

Phenology: An Integrative Environmental Science

BACKGROUND AND HISTORY OF TEMPERATURE MEASUREMENT

Although everyone intuitively understands the meaning of temperature, it is really not easily defined. One could describe it as a measure of the heat content of the air. However, heat is an even more vague term. According Horstmeyer (2001), who reported the following statistics, heat does not really exist. What does exist are unimaginable numbers of air molecules (about 2.69×1025 molecules/m3 at sea level) that are moving at incredible speeds (increasing from about 1600 to 1700 km/h for dry air as the temperature increases from 0°C to 40°C). The mass of individual air molecules is small and the total volume of the molecules occupies less than 0.1% of the air volume, but there are about 5.95×1014 collisions/s between molecules. When molecules collide, energy is not created or lost, but kinetic energy is transferred between the molecules. In thermodynamics, temperature is considered a measure of the average kinetic energy of the air molecules expressed as: T M ν 2 R' , where ν 2 is the mean of the squared molecular velocity, M is the molar mass, and R'' is the molar gas constant. Therefore, when the speed of the molecules increases, the temperature increases proportionally to the mean of the squared velocity. As air molecules strike our skin, some kinetic energy is absorbed and conducted from molecule to molecule into our bodies. We describe this “sensation” as “heat” and the term “sensible heat” is used to indicate that it is energy that we “sense”. If the temperature is higher, the molecules move faster, more hit our skin, more kinetic energy is received, and the hotter we feel. When a thermometer is placed in the air, its surface is also bombarded with air molecules at near sonic speeds. These collisions transfer kinetic energy to the thermometer and some is conducted inside. When the energy reaches and heats the liquid in the thermometer, it expands and moves up the capillary tube indicating a higher temperature. Historically, temperature measurement was closely linked to the development of thermometers. The earliest records pertaining to the concept of temperature were based on the writings of Aristotle in about 300 A.D. The first evidence of temperature measurement was in the late 1500's in Europe, with temperature scales appearing in the early 1600’s. In the mid 1600’s, the first scales based on the freezing point of distilled water appeared (Christopher Wren and Robert Hooke). By the 1700’s, Fahrenheit’s scale, which defined 32 degrees as the ice point and 96 degrees as the human body temperature was developed. Fahrenheit did not use the boiling point in his definition, but the boiling point (212°F) was used to manufacture thermometers. Celsius used the ice point and the boiling point to define a temperature scale starting with 0 degrees at the boiling point and

Chapter 4.9: Weather Station Siting

347

100 degrees as the ice point. However, shortly after Celsius proposed his temperature scale, others began to use degrees centigrade for the scale defined with 100 gradations between the ice point and boiling point (i.e., the reverse of the original Celsius scale). Because there were several temperature scales in usage during the late 1800’s, Callendar recommended a single practical temperature using the ice and boiling points of water with standard platinum resistance thermometers for interpolation. During the 20th century, several conferences on the international temperature scale were held in 1927, 1948, 1968, and 1990 (i.e., ITS-27, ITS-48, ITS-68, and ITS-90). The ITS-27 adopted the concepts of Callender, with the addition of thermocouples being the standard instrument. The ITS48 changed the name of the scale from centigrade to Celsius to credit Celsius for developing the scale and because the word centigrade represents a unit of angle measure in French. The ITS-48 also defined the absolute temperature scale as having exactly 100 gradations between ice point and the boiling point of water. The ITS-68 fixed the Kelvin temperature scale to absolute zero by setting 1.0 K = 1/273.16 of the triple point (i.e., the equilibrium between the phases of ice, liquid water, and water vapor). The triple point of water was set at exactly 273.16 K and the ice point was given the value of 273.15 K = 0°C. The boiling point was still set at 373.15 K = 100°C. They also eliminated the use of the term degrees for the absolute temperature scale (i.e., the unit is Kelvin rather than degrees Kelvin). Later, the ITS-90 reported the boiling point was really 99.975°C, so the definition of the boiling point was eliminated from the absolute temperature scale. However, at standard temperature and pressure, the Celsius scale still is defined as going from 0°C at the freezing point to 100°C at the boiling point, and the Fahrenheit scale goes from 32°F at the freezing point to 212°F at the boiling point. Today, absolute temperature (K) and temperature in Celsius, where (0°C = 273.15 K), are commonly used in science.

3.

MEASUREMENT THEORY

3.1

Sensors

Primary thermometers measure temperature in the sense that they measure variables, which are directly dependent on temperature, with coefficients that are virtually independent of temperature. For example, a gas thermometer, which measures temperature by the thermodynamic relationship between gaseous pressure, volume and temperature, is a primary thermometer. Secondary thermometers (e.g., thermistors, diodes, transistors, thermocouples, liquid-in-glass, liquid-in-metal, and metal deformation

348

Phenology: An Integrative Environmental Science

thermometers) measure variables that depend on temperature with coefficients that may be highly dependent on temperature. Secondary thermometers are the main types used for meteorological standards and operational meteorological instruments. Because they are the most commonly used in phenological studies, mercury-in-glass thermometers and thermistors will be discussed in this chapter. A mercury-in-glass thermometer is a secondary type thermometer because the liquid expansion involves coefficients, which are temperature dependent and not necessarily theoretically predictable. Most meteorological thermometers are full immersion types, so the entire glass thermometer is fully exposed to the air. By making the capillary (the bore in which the liquid rises) small relative to the liquid bulb (reservoir), the sensitivity and resolution of the thermometer increase. However, a limit is reached when the capillary surface tension forces degrade performance.

Figure 4.9-1. Maximum and minimum liquid-in-glass thermometers.

Maximum and minimum thermometers are two main types of liquid-inglass thermometers that operate on different principles (Figure 1). The maximum thermometer has a constriction in the capillary bore at a short

Chapter 4.9: Weather Station Siting

349

distance from the fluid reservoir. The liquid expands past the constriction, and expands to a maximum temperature. On cooling, the liquid above the constriction contracts leaving a vacuum barrier, while the liquid below the constriction contracts into the reservoir. This leaves the upper end of the liquid at the maximum temperature. The minimum thermometer has a small rider, which is pushed by the surface forces of the meniscus of the liquid within the capillary bore. The upper end of the rider is left at a minimum temperature position when the liquid rises again. Thermistors are made of sintered semiconductor materials (e.g., manganese, nickel, copper, iron, cobalt, and uranium oxides) that are pressed into the thermistor form and aged to promote stability. The resistance to electrical current decreases as the temperature rises and simple electronics are used to monitor the resistance. The physics involved in resistance sensor operation is described in Quinn (1983). With a thermistor, a curvilinear relationship between resistance and temperature is needed to calibrate the sensors. Older thermistors were somewhat unstable and, during their lifetime, the resistance would drift requiring frequent calibration. However, now thermistors are aged before calibration to minimize drift. Because the bead thermistors are manufactured with their leads inside the powdered material before they are pressed, whereas disk thermistors are manufactured prior to the leads being sprayed or printed on the material, thermistors made in the shape of beads are more stable than thermistors pressed into disk shapes. A major advantage of thermistors is that they are small sensors. The wire/metal resistors may be made very thin, but they still must have an appreciable length (several millimeters at the minimum) to have a practical level of resistance (tens of ohms) for temperature measurement. Thermistors can be made as small as 0.03 mm.

3.2

Shielding

Although having accurate sensors solves part of the problem with temperature measurement, proper shielding and ventilation are needed to minimize radiation effects on the sensor energy budget. The three main energy forms that affect energy balance on a thermometer are radiant, convective/sensible heat transfer, and thermal conduction. To avoid energy transfer from the mount to the sensor, the contact area should be small and insulated with plastic, ceramic, or cardboard materials that reduce conduction. For electronic sensors, the leads should be small, nonconductive, and maintained in the same shelter environment as the sensor to avoid direct exposure to radiation or other heat sources. Temperature sensors are frequently shielded from short wave radiation by putting them in a Stevenson Screen (Figure 2), which is made of wood

350

Phenology: An Integrative Environmental Science

and painted white to reflect away most solar radiation. For electronic sensors, white plastic shields like the Gill shield (Gill 1983) are often used (Figure 2). Long wave radiation is unavoidable, because it is emitted by all surfaces at terrestrial temperatures.

Figure 4.9-2. A standard Stevenson Screen, image by Jack Kelly Clark, © Regents of the University of California, used with permission (left), and a Gill radiation shield (right).

While the goal is to measure air temperature with a thermometer, the actual temperature recorded is a weighted mean of the air temperature and the housing or shield temperature around the sensor. Assuming that short wave radiation reaching the thermometer is negligible, Monteith and Tt) in terms of the Unsworth (1990) expressed thermometer temperature (T shield temperature (T Ts), air temperature (T), T and resistances to sensible (rH) and radiation (rrR) heat transfer as: Tt =

rH Ts rRT rR + rH

(1)

by assuming the net radiation and convective heat transfer are equal (R ( n = C) C as illustrated in Figure 3. When estimating air temperature, the goal is to have Tt approach T by making rH much smaller than rR and/or by making Ts very close to T T. To make rH much smaller than rR, one can maximize convective heat transfer by making the sensor very small or by ventilating the sensor. Generally, ventilation of a few meters per second is adequate. When ventilating, the air should be pulled over the sensor to avoid heating from the fans. In naturally ventilated Stevenson Screens or Gill shields,

Chapter 4.9: Weather Station Siting

351

errors can result when wind speed is low. Using double walls and roof and painting the shelter white to reflect solar radiation makes Ts closer to T. T

Figure 4.9-3. Drawing of the energy balance between a thermometer and its surrounding radiation shield where Rn is the net radiation, C is convective heat transfer, σ is the StefanBoltzmann constant, Ts, Tt, and T are the shield, thermometer, and air temperatures (K), ρ is the air density, Cp is the specific heat of air at constant pressure, and rR and rH are resistances to radiative and sensible heat transfer.

3.3

Height

When using temperature data in phenological models the height of the temperature measurement can have a big effect on the results. In the USA, the National Weather Service typically reports 10.0 m temperatures, whereas many climate and agricultural weather networks report temperatures for heights varying between 1.5 and 3.0 m. To illustrate the problem, Figure 4 shows the vertical temperature profile changes during the day before and during the night of a spring freeze event in a northern California mountain valley. Clearly, big differences exist between data taken at 0.5 m and 10.0 m heights. Therefore, when developing phenological models, temperature data should be used from a station with a temperature measurement height similar to the sensor heights of stations where it is likely to be used. Similarly, models from the literature should report the temperature measurement height and the model should only be used with data collected at the same height.

4.

HEAT (THERMAL) UNITS

For many applications (e.g., pest management, crop modeling, and irrigation scheduling) it is useful to predict when a crop or pest will develop to a particular phenological stage. When temperatures are higher, organisms

352

Phenology: An Integrative Environmental Science

develop in fewer days because they are exposed to the greater heat accumulation than organisms grown under cooler conditions. The accumulation of heat is called “physiological time” and “heat units” are a measure of physiological time, which typically are better than “calendar

Figure 4.9-4. Temperature profile data from a walnut orchard at 0.5, 1.0, 2.0, 6.0, and 10.0 m height prior to and during a radiation freeze night in a northern California valley.

Figure 4.9-5. Days and cumulative degree-days to mature cotton bollworm from neonate larvae to adult stages for different temperature ranges (after Wilson and Barnett 1983).

Chapter 4.9: Weather Station Siting

353

time” for predicting days between phenological stages. For example, the cotton bollworm development from neonate larvae to adult (Figure 5) is closely related to the number of accumulated heat units regardless of the large differences in number of days between the stages due to temperature differences during growth. The cumulative number of degree-days (oD) to develop is about the same, whereas the number of days to develop is greatly different for the different growth temperatures. Most organisms show little growth or development below a lower threshold temperature (T TL) and often there is little increase in development rate above an upper threshold (T TU), but the number of hours between TL and TU provide a measure of heat units for estimating the days between phenological stages. Each hour of time that the organism is exposed to temperature between the two thresholds is called a “degree-hour”. Dividing the accumulated degree-hours by 24 converts the heat units to degree-days, which are commonly used in phenological models. There are numerous papers on methods to estimate degree-days from daily maximum and minimum temperature data and on how to determine threshold temperatures and the cumulative degree-day requirement for various crops and pests (Baskerville and Emin 1969; Allen 1976; Johnson and Fitzpatrick 1977; Parton and Logan 1981; Kline et al. 1982; Snyder 1985; Wann et al. 1985; Reicosky 1989; de Gaetano and Knapp 1993; Kramer 1994; Yin et al. 1995; Pellizzaro et al. 1996; Roltsch et al. 1999; Snyder et al. 1999, 2001; Cesaraccio et al. 2001). In general, researchers have attempted to model diurnal temperature changes using mathematical models with a range of complexity (Zalom et al. 1983; Snyder et al. 1999). Calculating the number of degree-days between the two phenological stages of interest using selected lower and upper threshold temperatures and all appropriate data sets best identifies threshold temperatures. Then the mean cumulative degree-days (oDm) over all of the data sets are computed. The same thresholds are used to calculate degree-days for each data set until the cumulative degree-days adds up to oDm. Then the mean absolute difference or root mean square error between the predicted and observed number of days over all of the data sets is calculated and recorded. This process is repeated for a range of threshold temperatures until the smallest mean absolute difference or root mean square error is identified. The threshold temperatures with the smallest difference between predicted and observed days between stages and the corresponding oDm provide the best prediction model.

354

5.

Phenology: An Integrative Environmental Science

PLANT VERSUS AIR TEMPERATURE

Whereas phenological development is driven by plant temperature, most phenological models use air temperature as the main input. Ideally, plant rather than air temperature should be used in phenological models. Because the plant temperature depends on the surface energy balance, there is no simple method to estimate plant surface temperature from air temperature. Based on the energy balance over a wet surface, the following equation (after Monteith and Unsworth 1990) estimates the surface temperature (T To): To = T +

es − e r γ *( n − a Δ + γ * ρC p Δ + γ *

)

(2)

where T is the air temperature (°C), es is the saturation vapor pressure (kPa) at T, T e is the vapor pressure (kPa), ra is the aerodynamic resistance (s/m), Rn is the net radiation (W/m2), G is the soil heat flux density (W/m2), ρ is the air density (g/m3), Cp is the specific heat of air at constant pressure (J/g K), Δ is the slope of the saturation vapor pressure at air temperature (kPa/°C), γ is the psychrometric constant (kPa/°C), and γ* = γ (rs/ra), where rs is the surface resistance (s/m). The adiabatic part (i.e., the middle term on the right) is directly proportional to the vapor pressure deficit (es – e), which is approximately an exponential function of the air temperature. The diabatic term (i.e., the right-hand term) depends on the external energy supply (mainly net radiation). Therefore, a wet surface temperature will be warmer than air temperature when the vapor pressure deficit is high enough that the adiabatic term is bigger than the diabatic term. Clearly, Equation 2 is complex, involving parameters that are not readily available for phenological models, and it only applies to relatively smooth, wet surfaces (e.g., for unstressed plant canopies). However, considerable past research with infrared thermometers has shown that the temperature of unstressed canopies is predictable using the air temperature and vapor pressure deficit (Idso et al. 1981; Jackson et al. 1981).

6.

SITING EFFECTS

6.1

Underlying Surface

The main indirect source of energy for heating the air is the sum of direct and diffuse short-wave (solar) radiation minus that which is reflected away

Chapter 4.9: Weather Station Siting

355

from the surface. Although smaller in magnitude, a more direct source is long-wave radiation from the surface upwards and from the sky downwards. By convention, downward radiation that adds energy to the surface is positive, and radiation away from the surface is negative. During a clear, mid-summer day at noon between about 30° and 50° latitude, the downward short-wave radiation is on the order of 1000 W/m2 and, from a grass surface, about 25% of the energy received is reflected away leaving a net short-wave balance of about 750 W/m2. Under clear skies at the same latitude, the longwave radiation is about -450 W/m2 upwards and about 350 W/m2 downwards leaving a net long-wave radiation balance of about -100 W/m2. Combining the short- and long-wave balances, the mid-latitude net radiation near noon on a clear summer day is about Rn = 650 W/m2 over a grass surface. Under clear skies, the net long-wave radiation changes little with time, but, under overcast skies, because the clouds are warmer than clear sky, it decreases to near -10 W/m2. Short-wave radiation changes considerably over the day with the angle of the sun above the horizon and it mostly decreases with cloud cover. Energy from positive radiation does one of the following: 1. heats the soil surface and conducts downward into the soil, 2. vaporizes water that contributes to upward latent heat flux, 3. heats air near the surface, which then convects upwards as sensible heat flux, and contributes to 4. miscellaneous consumption for heating the plants and photosynthesis. However, energy flux also reverses direction and does one of the following: 1. conducts heat upward from the soil to the surface, 2. condenses water vapor and convert the energy to sensible heat, 3. transfers sensible heat to a colder soil surface, and 4. miscellaneous release losses due to cooling plants and respiration. Using the convention that positive Rn at the surface is partitioned into energy that heats the soil (G), vaporizes water (λE), heats the air (H), or contributes to miscellaneous energy consumption (M (M) used for heating the plants and photosynthesis, the energy balance equation is written as: Rn = G + λE + H + M

(3)

Usually, M is relatively small and is ignored for energy flux calculations. If the soil temperature decreases with depth below the surface, then G is positive. Similarly, if the air temperature decreases with height above the surface then H is positive. If the temperature profiles are reversed then G or H are negative and heat is transferred to the surface. If more water vapor is vaporized than condensed on the surface, then the water vapor flux is

356

Phenology: An Integrative Environmental Science

upwards and λE is positive. If more water condenses than vaporizes, then λE is negative. Temperature is a measure of the sensible heat content of air at the height of the sensor. Because the temperature gradient with height is typically much greater than with horizontal distance, vertical fluxes of sensible heat are generally much more important than horizontal fluxes. Therefore, energy balance at the surface, which largely determines the sensible heat content at the sensor level, is extremely important. If the surface is warmer than the air above, a positive H will likely increase the sensible heat content and hence the temperature of air at the sensor level. If the surface is colder than the air above, a negative H will tend to lower the sensible heat content and temperature at the sensor height. Assuming no horizontal advection and the same incoming long-wave and solar irradiance, the outgoing long-wave and, hence, net radiation depend mainly on the surface albedo and temperature, which is affected by G, H, H and λE. Albedo is affected by surface properties and the angle of incidence of the solar radiation. The relative partitioning of absorbed incoming radiation to G, H H, and λE determines the surface temperature. Soil heating is affected by thermal conductivity and soil heat capacity. Latent heat flux is mainly affected by the presence of water to vaporize and secondarily by the water vapor content of the air and turbulence, which transfers sensible and latent heat to and from the surface. Sensible heat flux is the residual energy after the Rn contributions to heating the soil (G) and evaporating water (λE) are removed. Sensible heat content of air near the surface determines the air temperature. For a given incoming flux density of radiation, the presence or absence of water is the main factor affecting changes in sensible heat content of air near the surface and, hence, temperature. For example, recording temperature above a transpiring grass surface will result in lower temperature than measurements above a dry, bare ground surface. And the difference between temperature recordings over the two surfaces is greater as the climate becomes more arid. In general, most climate and weather forecast stations in the USA are located over plots of land with the natural vegetation of the region. Most agricultural weather stations are located over grass or irrigated grass surfaces to standardize the site for comparisons and to allow for model development that can be transferred to other locations. In arid climates, variations in the energy balance due to intermittent rainfall, can greatly affect temperature readings. For example, Snyder et al. (2001) compared annual degree-day calculations, with a 10°C lower threshold, from stations located over irrigated grass and over bare soil in four regions of California and found that the cumulative degree-days were between 3.2 and 10.7% higher for the unirrigated, bare soil. The difference is likely due to differences in energy

Chapter 4.9: Weather Station Siting

357

balance over bare soil that is similar to the grass during the rainy season but can be quite different when the soil dries and the grass continues to transpire. Also, it is likely that differences in vegetation around the weather stations contributed to differences in the degree-day accumulation. The type of underlying surface around a weather station should be selected depending on the purpose of the data. If the purpose is to characterize phenological development of natural vegetation, the stations should be sited in a natural setting that represents the conditions of the vegetation being studied. However, if the data are used to model phenology of irrigated crops, it is best to site the stations over an irrigated grass surface that removes variability due to intermittent rainfall.

6.2

Fetch

Fetch is the upwind distance of the same vegetation underlying a weather station. Inadequate fetch is a problem for some micrometeorological measurements and it can lead to errors in temperature measurements if the purpose is to collect data over a standard surface (e.g., irrigated grass). Under extremely arid conditions, the authors have observed systematic overestimation of temperature by about 4% from a station downwind from a desert when compared to a station over irrigated grass 177 m from the edge of the desert (Figure 6). While this amounts to only about 1.6°C higher temperature at 40°C, the readings were systematic and they led to considerable differences in degree-day accumulations (Snyder et al. 2001).

6.3

Surrounding Environment

Even when temperatures are recorded over the same standard surface, sometimes big differences are recorded due to the surrounding environment. For example, Figure 7 shows the corresponding temperatures recorded at Torrey Pines and Miramar Naval Air Station near San Diego, California. Torrey Pines is on the coast and Miramar is located about 30 km inland. Clearly, although these sites are not far apart, there were big differences in the temperature data with the coastal site having warmer temperatures when values are low and the inland site having warmer temperatures when values are high. If phenological models were to be applied in this region, many stations would be needed to account the temperature effects of the ocean and how they change with distance from the coast.

358

Phenology: An Integrative Environmental Science

Figure 4.9-6. Corresponding hourly temperature data (May-June 2000) measured on the west edge of a grass field and 177 m east of the west edge of a large grass field surrounded by desert near Indio, California. The prevailing wind was from the west.

Figure 4.9-7. Corresponding hourly temperature data (June-November 2000) from Torrey Pines on the coast and Miramar Naval Air Station, which is 30 km inland.

In areas that are far from the ocean, there are often big differences in temperature over similar underlying surfaces. For example, Figure 8 shows the corresponding temperatures measured at the Indio CIMIS station and the Vintage Country Club, which are only about 30 km apart in the below sea

Chapter 4.9: Weather Station Siting

359

level desert in southern California. In this example, the regression statistics indicate that the temperature at the Vintage Country Club is about 3% less than at the Indio CIMIS station, but that is because the regression was forced through the origin and the regression line is strongly influenced by the high temperatures during midday, which are similar. However, at lower temperatures, during the morning and afternoon, the temperatures at the Vintage Country Club were considerably more variable and lower. The difference is mainly due to considerably less wind and mountains that shade the station in the afternoon at the Vintage Country Club. Because the diurnal temperature curves are different, the same temperature curve model cannot be used to estimate degree-days at both sites. Therefore, the use of temperature driven phenological models is problematic in an area with multiple microclimate zones.

Figure 4.9-8. Corresponding hourly temperature data (June - November 2000) measured at the Vintage Country Club near Indio, California and the Indio CIMIS station. The wind speed at the Vintage Country Club averaged about 8% of the Indio CIMIS site.

7.

CONCLUSIONS

Temperature is the driving factor in most phenological models, and proper measurement is critical for both development and use of the models. In addition to selecting accurate sensors, they should be mounted at an appropriate height and properly shielded from short-wave radiation (double shielding is best). Choosing small sensors that respond rapidly, protecting electronic leads, and ventilation (in areas with little wind) can improve

360

Phenology: An Integrative Environmental Science

accuracy of the temperature measurements. Data should be collected at a height that is typical of other weather stations in the area where the model will be used. Generally, agricultural weather stations collect temperature data at 1.5 to 2.0 m height and weather services tend to measure at 10.0 m height. For phenological models of natural vegetation, it is best to site the weather station in a similar environment without irrigation. However, when the models are used for irrigated crops, the stations should be sited over an irrigated grass surface to avoid temperature fluctuations due to intermittent rainfall at the measurement site. Strong temperature gradients can occur near large water bodies (e.g., the ocean or large lakes) and in hilly or mountainous regions where sunlight is blocked during part of the day. In such regions, more weather stations are needed to better characterize microclimate differences. However, even when the temperature data are accurately determined, inaccuracies in model predictions can occur because it is plant temperature rather than air temperature that truly drives the phenological development. Although little or no literature on the topic exists, perhaps using vapor pressure deficits to estimate plant from air temperature could improve models and make them more universally applicable.

REFERENCES CITED Allen, J. C., A modified sine wave method for calculating degree days, Environ. Ent., 5, 388396, 1976 Baskerville, G. L., and P. Emin, Rapid estimation of heat accumulation from maximum and minimum temperatures, Ecology, 50, 514-517, 1969. Cesaraccio, C., D. Spano, P. Duce, and R. L. Snyder, An improved model for degree-days from temperature data, Int. J. Biometeorol., 45, 161-169, 2001. de Gaetano, A., and W. W. Knapp, Standardization of weekly growing degree day accumulations based on differences in temperature observation and method, Agric. For. Meteorol., 66, 1-19, 1993. Gill, G. C., Comparison testing of selected naturally ventilated solar radiation shields, Final Report Contract # NA-82-OA-A-266, NOAA, St. Louis, 15 pp., 1983. Horstmeyer, S., Building blocks – What goes on in a cubic meter of air?, Weatherwise, 54(5), 20-27, 2001. Idso, S. B., R. D. Jackson, P. J. Pinter, R. J. Reginato, and J. L. Hatfield, Normalizing the stress-degree-day parameter for environmental variability, Agric. Meteorol., 24, 45-55, 1981. Jackson, R. D., S. B. Idso, R. J. Reginato, and P. J. Pinter, Canopy temperature as a crop water stress indicator, Water Resour. Res., 13, 651-656, 1981. Johnson, M. E., and E. A. Fitzpatrick, A comparison of methods of estimating a mean diurnal temperature curve during the daylight hours, Arch. Meteorol. Geophys. Bioklimatol., Series B, 25, 251-263, 1977.

Chapter 4.9: Weather Station Siting

361

Kline, D. E., J. F. Reid, and F. E. Woeste, Computer simulation of hourly dry-bulb temperatures, Virginia Agricultural Experiment Station, No. 82-5, Virginia Politechnical Institute and State University, Blacksburg, 18 pp., 1982 Kramer, K., Selecting a model to predict the onset of growth of Fagus sylvatica, J. Appl. Ecol., 31, 172-181, 1994. Monteith, J. L., and M. H. Unsworth, Principles of Environmental Physics, 2nd ed., Edward Arnold, London, 291 pp., 1990. Parton, W. J., and J. A. Logan, A model for diurnal variation in soil and air temperature, Agric. For. Meteorol., 23, 205-216, 1981. Pellizzaro, G., D. Spano, A. Canu, and C. Cesaraccio, Calcolo dei gradi–giorno per la previsione delle fasi fenologiche nell’actinidia, Italus Hortus, 5, 24-30, 1996. Reicosky, L. J., J. M. Winkelman, J. M. Baker, J. M., and D. G. Baker, Accuracy of hourly air temperatures calculated from daily minima and maxima, Agric. For. Meteorol., 46, 193209, 1989. Roltsch, J.W., F.G. Zalom, A.J. Strawn, J.F. Strand, and M.J. Pitcairn, Evaluation of several degree day estimation methods in California climates, Int. J. Biometeorol., 42, 169-176, 1999. Quinn, T. J., Temperature, Academic Press, New York, 495 pp., 1983. Snyder, R. L., Hand calculating degree-days, Agric. For. Meteorol., 35, 353-358, 1985. Snyder, R. L., D. Spano, C. Cesaraccio, and P. Duce, Determining degree-day threshold from field observations, Int. J. Biometeorol., 42, 177-182, 1999. Snyder, R. L., D. Spano, P. Duce, and C. Cesaraccio, Temperature data for phenological models, Int. J. Biometeorol., 45, 178-183, 2001. Wann, M., D. Yen, and H. J. Gold, Evaluation and calibration of three models for daily cycle of air temperature, Agric. For. Meteorol., 34, 121-128, 1985. Wilson, L. T., and W. W. Barnett, Degree-days: an aid in crop and pest management, Calif. Agric., 37, 4-7, 1983. Yin, X., M. J. Kropff, G. McLaren, and R. M. Visperas, A nonlinear model for crop development as a function of temperature, Agric. For. Meteorol., 77, 1-16, 1995. Zalom, F. G., P. B. Goodell, W. W. Wilson, and W. J. Bentley, Degree-Days: The calculation and the use of heat units in pest management, Leaflet 21373, Division of Agriculture and Natural Resources, University of California, Davis, 10 pp., 1983.

PART 5

REMOTE SENSING PHENOLOGY

Chapter 5.1 REMOTE SENSING PHENOLOGY Bradley C. Reed1, Michael White2, and Jesslyn F. Brown1 1

SAIC, USGS EROS Data Center, Sioux Falls, SD, USA; 2Department of Aquatic, Watershed, and Earth Resources, Utah State University, Logan, UT, USA

Key words:

1.

Remote Sensing, Greenness, Growing season, Data smoothing

INTRODUCTION

The role of remote sensing in phenological studies, while still in development, is increasingly regarded as a key to understanding large area seasonal phenomena. Repeat observations from satellite-borne multispectral sensors provide a mechanism to move from plant-specific to regional scale studies of phenology. In addition, we now have sufficient time-series (since 1982 at 8-km resolution covering the globe and since 1989 at 1-km resolution over the conterminous US) to study seasonal and interannual seasonal characteristics from satellite data. The need for large area phenology measures to document and support global change studies is growing. From 1998-2001 the Earth experienced four of the warmest years during the period of climatic record. How these warmer temperatures affect vegetation characteristics from the timing of spring greenup and subsequent environmental changes such as, among others, CO2 assimilation, surface roughness, and evapotranspiration are the subject of scrutiny in the global change sciences. In this chapter we discuss the basis for remote sensing phenology, approaches to generate phenology measures from satellite data, and issues to consider when developing and interpreting satellite-derived phenology measures. Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 365-381 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

366

2.

Phenology: An Integrative Environmental Science

GREENNESS MEASURES

Most remote sensing phenology studies utilize data collected from the red and near infrared portions of the electromagnetic spectrum that are transformed into a vegetation or greenness index. Vegetation indices (VIs) are designed to take advantage of the spectral reflection/absorption characteristics of plants to describe the “greenness” of each pixel in a satellite observation. Plants absorb red light and reflect energy in the near infrared portion of the electromagnetic spectrum. The normalized difference vegetation index (NDVI) is the most commonly used index for large area studies (Goward et al. 1985; Tucker and Sellers 1986; Malingreau et al. 1989; Loveland et al. 1991; Townshend et al. 1994). However several alternative vegetation indices have been produced to reduce canopy background effects and atmospheric contamination, notably the soil adjusted vegetation index (SAVI, Huete 1988) and the soil and atmospherically resistant vegetation index (SARVI, Kaufman and Tanré 1992). In a study comparing these indices Huete et al. (1997) noted that each has strengths and weaknesses for certain applications. The NDVI is more strongly coupled to red band reflectance, while the other indices are more coupled to near infrared reflectance. The NDVI thus seems to be well suited for studies concerned with the photosynthetic capacity of vegetation cover (e.g., fPAR and fractional green cover), while the SAVI and SARVI are suitable for studies that are concerned with structural canopy parameters (e.g., LAI, biomass) that are more apparent in near infrared reflectance (Huete et al. 1997). In addition, the NDVI is less sensitive in densely vegetated regions. So while a single optimal index does not exist for use in phenology studies, NDVI continues to be the standard, primarily because of its ready availability and strong heritage.

3.

SENSORS AND DATA SMOOTHING

Relatively high resolution satellite sensors, principally the Landsat series, were utilized for phenological studies, especially in the 1970’s and 1980’s (e.g. Wiegand et al. 1979; MacDonald and Hall 1980). These early efforts were directed toward utilizing multi-date Landsat imagery for land classifications or toward establishing a relationship between temporal spectral variability and crop calendars (Crist and Malila 1982; Meltzer et al. 1982). These studies were based on observations of the temporal profile of the greenness component of the crop canopy, rather than of individual plants (Badhwar 1980). The same principle is still applied today with large area satellite phenology, but is taken further. Most researchers are tracking the

Chapter 5.1: Remote Sensing Phenology

367

integrated greenness component of largely heterogeneous 1-km pixels, rather than single plants or dominant plant types. Research efforts using the relatively high resolution sensors, such as Landsat, for measuring phenology over large areas are inherently limited due to the relatively long revisit period (16 days) being insufficient for identifying critical phenological dates and the fine spatial resolution of the sensors requiring prohibitively large computing resources. The satellite sensor that has primarily been used for phenology studies is the National Oceanic and Atmospheric Administration’s (NOAA) advanced very high resolution radiometer (AVHRR). This sensor has a near daily repeat cycle of the Earth and a 1-km spatial resolution. Both the temporal density of the data and the moderate spatial resolution make this sensor well suited for studying large area phenology. AVHRR-NDVI data are readily available in a consistently processed database from 1982-present at an 8-km resampling grid covering the globe and at 1-km resolution from 1989 covering the conterminous United States. Other moderate resolution satellite sensors including SPOT Vegetation (1 km data, launched in 1998) and Envisat MERIS (300m data, launched in 2002) have the proper instrumentation for greenness studies. The moderate resolution imaging spectroradiometer (MODIS) launched in December, 1999 may well become the standard sensor for phenology studies. Improved geometry, radiometry, and overall data quality of MODIS will provide high quality data for time-series analysis as soon as a sufficient population of data has been collected. Vegetation index data, NDVI and enhanced vegetation index or EVI, are being routinely produced at 250-m, 500-m, and 1-km spatial resolution. Vegetation indices are normally composited over a set period of time, typically 10 days or two weeks, to eliminate spuriously low values caused by cloud contamination, haze, and other atmospheric effects (Eidenshink 1992; Los 1994). Even with the composited data, however, there are often lingering effects that tend to reduce the NDVI value and, more importantly, disturb the temporal profile of the vegetation signal. Figure 1 illustrates a typical NDVI time-series collected over a corn/soybean land cover type. There is a clear greenup, maturity, senescence phenology depicted for a 2year period. However, there are also several downward spikes in the NDVI signal that are dramatic, but too short-lived to be a function of a real decline in vegetation condition. To eliminate the spurious data, which can affect algorithms that are searching for increasing or decreasing trends representing real phenological shifts, a temporal smoothing of the data is typically performed. Temporal VI smoothing has been performed for some time, including the Best Index Slope Extraction (BISE, Viovy et al. 1992), compound mean and

368

Phenology: An Integrative Environmental Science

median filters (e.g., VanDijk et al. 1987), and splines (White 1997). Swets et al. (1999) proposed a weighted least-squares approach that they believe eliminates some of the time-shifts caused by over generalization of the signal by giving added weight to locally (in the time domain) high values (Figure 1). Whatever smoothing algorithm is used, the signal should retain its temporal nuances without over generalization, eliminate spurious downward spikes in the VI, and retain sustained temporary declines in VI that are representative of declines in vegetation condition. original smoothed 0.6 0.5 NDVI

0.4 0.3 0.2 0.1

Year 1

53

49

45

41

37

33

29

25

21

17

13

9

5

1

0 Year 2

Figure 5.1-1. Time-series NDVI data over a corn/soybean land cover type. Clouds, haze, and other atmospheric perturbations may cause downward spikes in the temporal NDVI signal that may be corrected using a temporal smoothing technique.

But what, exactly, do satellite sensors measure with these vegetation indices? Some approaches toward satellite phenology are designed to detect the first sustained positive change in the vegetation index during the springtime. Others seem to be measuring leaf expansion of the dominant overstory species. When utilizing satellite phenology derivations, one should be careful about interpreting the phenomena that are being estimated by the various approaches. Just as different approaches for estimating growing season length (e.g., days with fPAR > 0.5; frost-free days; carbon uptake period) may have different results (see Chapter 7.1 in this volume), so may alternative approaches using satellite data.

Chapter 5.1: Remote Sensing Phenology

4.

369

ESTIMATING THE START OF GROWING SEASON

The key phenological variables that may be estimated from satellite remote sensing are the time of the start of the growing season (SOS), end of growing season (EOS) and the growing season length (GSL). When deriving phenological characteristics from remote sensors, the object is not species, population or even community specific phenology, but rather the characteristics of individual pixels. If the pixel size is on the order of hundreds of meters or a kilometer or greater, the objective usually becomes characterizing the phenology of a mosaic of several vegetation types. Because of the more general target, the specificity of phenology that can be determined is also reduced. The contribution of the individual phenology of the various components of a pixel are mixed and homogenized. The integrated phenology of the pixel is usually addressed in terms of a more general statement such as the start or the end of the growing season, rather than more specific measures such as first-leaf or bud-burst. The various approaches that have been used for satellite phenology can be grouped into three types; threshold-based, inflection points, and trend derivatives.

4.1

Threshold-based Phenology

Threshold-based measures use either a pre-defined or a relative reference value for defining the start of season. For example, in one of the earliest efforts at setting a definitive date for SOS, Lloyd (1990) used a NDVI=0.099 as the threshold value in a study using AVHRR NDVI data. When the NDVI reached this threshold value, SOS was reached. Similarly, EOS was defined as the time at which the NDVI reached this threshold. Such a reference value can be very effective at defining local SOS values, but there is difficulty in using such values over environments with different soil background characteristics or land cover types. Reed et al. (1994) showed that over the conterminous US, the NDVI value at SOS might vary from 0.08 to as high as 0.40. Another threshold-based approach, seasonal midpoint NDVI (SMN) uses the midpoint between minimum and maximum NDVI to define the SOS and EOS (White et al. 1997, 1999; Schwartz et al. 2002). This permits the threshold to be tied to the seasonal amplitude of the individual pixels and thus, the dynamic characteristics of each pixel. Similarly, Jönsson and Eklundh (2002) use the 10% distance between minimum and maximum NDVI as representing the SOS.

370

4.2

Phenology: An Integrative Environmental Science

Inflection Point Phenology

Badhwar (1984) created a model to describe greenness-based phenology from Landsat data. The goal of this study was to separate small grain crops from other vegetation types. Six seasonal parameters were extracted from the temporal profile based on Crist (1982) and adapted to Landsat data and included two inflection points, the width of the profile and a peak time. The goal of this study was not necessarily to document the phenology of small grains, but was rather to aid the discrimination of small grains from other land cover types that have similar spectral characteristics, such as hay and pasture. Moulin et al. (1997) used 1º latitude x 1º longitude grid cell size AVHRR VI data over the globe to extract the beginning, maximum, and end of the seasonal phenological cycle. They assumed that each pixel displayed either a single seasonal cycle each year or no cycle at all and that a growth cycle includes growth, senescence, and dormancy. The beginning of the cycle (growth) is defined as where the (left) time derivative transitions from 0 to a positive number. Similarly, the end of the seasonal cycle is defined as where the (right) time derivative transitions from negative to zero. They note that the SOS is relatively homogeneous compared to the EOS, but that reasonable land cover interpretations may be made from the resulting image maps. Problems in using the temporal NDVI signal for global assessment included atmospheric contamination (which may be reduced by smoothing) and difficulty in interpreting evergreen, snow effects, and the slow rate of senescence in some biomes. Zhang et al. (2003) identified four phases of vegetation phenology: greenup, maturity, senescence, and dormancy. Their algorithm to identify these phases is based on finding when the curvature of the NDVI time-series changes from one linear stage to another, i.e. from decreasing to increasing or increasing to decreasing sections. The points with maximal curvatures are where the phenology changes phases. They applied this algorithm to 1 km AVHRR/NDVI data from 1995-1996 and over New England using MODIS NDVI derived from nadir, BRDF-adjusted reflectance (Zhang et al. 2001). The advantage of this approach is that it permits the description of multiple growing seasons and discriminates an additional phase of phenology (maturity) compared to most other methods.

4.3

Curve Derivative Phenology

Similar to the previously discussed approaches, phenology estimates derived from curve characteristics seek to identify where the VI data exhibit a rapid, sustained increase. In a study using phenology along with climate

Chapter 5.1: Remote Sensing Phenology

371

and soil variables for carbon cycle models, Kaduk and Heimann (1996) defined budburst as the date of the largest increase in NDVI value after the monthly mean temperature had reached 5ºC. The inclusion of the satellitebased bud-burst resulted in an improved NPP model that was closer to observed CO2 at several stations in the northern hemisphere. Reed et al. (1994) determined SOS by employing a backward-looking or delayed moving average (DMA). NDVI values are compared to the average of the previous n observations to identify departures from the established trend. The DMA values can be thought of as a predicted value based on previous observations. A trend change is defined as when the NDVI values become larger than those predicted by the DMA. They are then able to define twelve seasonal metrics for each year.

5.

ESTIMATING END OF SEASON AND DURATION OF GROWING SEASON

The end of the growing season is nearly always calculated with a similar approach as the start of the growing season. For example, in the thresholdbased approaches, the same threshold as the SOS is used for EOS, except it is defined during a sustained decrease, rather than increase in VI (e.g., Lloyd, 1990). Inflection points are also utilized for identifying the EOS (Moulin 1997; Zhang 2001) and Reed et al. (1994) use a forward-looking moving average for EOS, rather than the DMA, which is employed for the SOS. The duration of the growing season is then simply calculated as the difference between the EOS and SOS.

6.

WHAT IS SATELLITE PHENOLOGY MEASURING?

While a variety of approaches are being utilized to derive phenology from satellite sensors, no ideal technique has yet been developed. A critical concern, given all of the constraints that are inherent in satellite-derived phenology (such as pixel size, temporal resolution, vegetation index limitations, atmospheric contaminants, etc.), is what key features can we detect and measure? As Menzel (2002) notes, even ground-based observations of phenology are inherently subject to a degree of inaccuracy due to the reliance on the observational skills of observers. So, while satellite phenology is confounded by a number of factors beyond the control

372

Phenology: An Integrative Environmental Science

of the analyst, it remains an objective measurement of changing environmental dynamics. Non-vegetation environmental conditions that affect vegetation indices during the time of the onset of the growing season include snowmelt, soil moisture, preparing the soil for planting, litter, and the aforementioned atmospheric variables. Snowmelt and soil moisture changes may be directly related to phenological stages that are of importance to scientists. In some environments, herbaceous vegetation is green beneath the snow cover and is able to begin photosynthesis as soon as it is exposed to sunlight. In other environments, a soil moisture increase may be the key to starting vegetation growth. Identifying SOS is important because that is when the land surface/atmosphere boundary begins changes that are associated with energy, mass, and momentum exchange.

6.1

Evaluation of Methods

Despite the growing number of efforts toward characterizing satellite phenology, there are relatively few that provide a thorough evaluation of the results. As mentioned earlier, it is difficult to obtain an objective measure of phenology over a single plant type, so the difficulty in observing the phenology over a 1-km pixel, is compounded. Schwartz et al. (2002) conducted an assessment of two SOS measures (delayed moving average or DMA and seasonal midpoint NDVI or SMN). They found that the DMA was relatively earlier than the SMN because each is measuring fundamentally different processes. They believed the DMA was detecting the first sustained flush of greenness, while the SMN was designed to predict initial leaf expansion of the dominant overstory species. The DMA appeared to have a systematic bias toward earlier SOS at higher latitudes, which is probably due in part to the ability of northern plant species to initiate growth with less energy (Schwartz and Crawford 2001). The SMN appeared to be more related to land cover type than the DMA, but with greater error in the individual years of study. A simple example pixel (Figure 2) shows that, in general, those methods employing inflection points will show the earliest SOS, followed in order by the DMA, the time of greatest increase, and the SMN showing the latest time for SOS. Inflection points depict the time at which the VI first begins to increase, while the DMA is a more conservative measure of the increase designed to eliminate false SOS. The time of greatest increase is when the early season growth is at its peak (probably what the lay-person thinks of as the beginning of spring), while SMN occurs when the vegetation has reached an established stage.

Chapter 5.1: Remote Sensing Phenology

373 Pixel value DMA Inflection pt.

0.6

Max. Increase

NDVI

HMN

0.4

0.2

25

22

19

16

13

10

7

4

1

0 Time

Figure 5.1-2. Time-series NDVI with SOS plotted using delayed moving average (DMA), inflection point, time of maximum increase, and seasonal midpoint NDVI (SMN).

Moreover, Duchemin et al. (1999) found a good correlation between the satellite-derived budburst and the budburst timing predicted from air temperature using the thermal time model in nearly monospecific forests of France. Chen et al. (2001) analyzed the temporal relationships between growing season parameters of multi-species community phenology in the field observation areas, and the corresponding threshold NDVI values obtained by surface-satellite analyses at the pixels overlaying the areas in northern China from 1982 to 1993. They found that the threshold NDVI values have good implications for detecting the start of phenological growing season, but not for detecting the end (Chen et al. 2001).

6.2

Confounding Issues

Regardless of what approach is being utilized to estimate phenology from satellite data, there are a number of confounding issues that must be addressed, including: 1) regions with no clear growing season (Figure 3a), 2) evergreen types (Figure 3b and 3c), and 3) multiple growing seasons (Figure 3d). One of the difficulties with remote sensing phenology is creating an algorithm to handle the wide variety of real time-series curves, rather than a model curve. For example, in some areas, particularly desert shrublands (Figure 3a) and some evergreen biomes, there is not a distinct seasonal signal. The algorithm should be able to identify these regions and assign them a value such as “no growing season” or “evergreen seasonality”.

374

Phenology: An Integrative Environmental Science

a

b

c

d

Figure 5.1-3. Issues in satellite-derived phenology include a) regions with no clear start of season; b) evergreen needleleaf trees with false seasonality, c) evergreen needleleaf trees with little seasonality; and d) multiple seasons.

Chapter 5.1: Remote Sensing Phenology

375

However, some evergreen biomes do have a distinct seasonal NDVI signal. Figure 3b is a pixel from the spruce-fir-pine coniferous forest of the Rocky Mountains in Colorado and Figure 3c is a pixel from the fir-hemlockpine forest of the Cascades in Oregon. These two evergreen pixels have a distinctly different seasonality as measured by the NDVI time series. The Rocky Mountain pixel has a misleading strong seasonal signal that is similar to that of croplands, which go from bare soil to a very strong vegetation signal. The low winter NDVI values are likely the result of snow cover. The evergreen pixel depicted in Figure 3c is likely the result of a denser canopy cover with less influence from snow cover. Perhaps the use of snow cover data, coupled with the NDVI data could help reduce the effects of snow on defining the seasonality of evergreen regions. Figure 3d illustrates a pixel from a region with a bimodal growing season from the Imperial Valley of California. Multiple growing seasons are rare in the conterminous U.S., but are common in subtropical regions and in agricultural regions of Asia. The challenge in satellite phenology is addressing how to depict both growing seasons. Depending on the ultimate application of the phenology database, the end-user may be interested in characterizing only the primary growing season, each growing season, or the two growing seasons together. It is possible to describe multiple growing seasons, but the manner in which the information is presented merits research.

7.

APPLICATIONS OF SATELLITE PHENOLOGY

Scientists have applied remote sensing of phenology for many purposes including monitoring biospheric activity, developing prognostic phenology models, and deriving land cover maps. They have also utilized satellite based phenology measures studies to study trends in growing season length and vegetation production. Myneni et al. (1997a) presented evidence that photosynthetic activity of terrestrial vegetation at high latitudes increased from 1981 to 1991 due to an increase in the length of growing season. This increase corresponded to an increase in the amplitude of the seasonal cycle of atmospheric CO2 since the early 1970s and an advance in the time of the drawdown of CO2 in spring of up to seven days (Keeling et al.1996). Tucker et al. (2001) showed similar results, but over an expanded timeframe (19821999). Their study showed an average May to September NDVI increase of 8-9% (excluding 1991 and 1992 due to the Mt. Pinatubo eruption) over the period of study at high latitudes (45ºN to 75ºN). The NDVI increases were associated with earlier starts of the growing season and with increases in the surface temperature at these latitudes.

376

Phenology: An Integrative Environmental Science

For applications such as biospheric monitoring and prognostic model development, a particular set of phenological metrics such as SOS, EOS, greenness amplitude, rate of greenup, etc. are often required. We term these active applications. In a passive phenological application, specific stages of phenological development are not extracted or predicted; rather, the seasonal progression of phenological development, as observed in variables such as leaf area index (LAI) or the fraction of photosynthetically active radiation absorbed by plant canopies (FPAR) is used as an input to a model, much like meteorological information. In ecological models, the concept of light use efficiency (LUE) is often used. Here, plant productivity, either gross primary production (GPP) or net primary productivity is calculated as follows: n

productivity = ε × f (environment )∑ APAR

(1)

0

where İ is the dry matter conversion efficiency (g/J), f(environment) represents a series of scalars that are used to down-regulate the potential value of İ, and APAR is the photosynthetically active radiation absorbed by plant canopies. Usually the environmental scalars include precipitation and temperature. APAR is calculated from observations or modeled predictions of photosynthetically active radiation and a remotely-sensed prediction of FPAR, usually as a function of NDVI (Asrar et al. 1984; Myneni et al. 1997b). In other words, APAR = FPAR*PAR. The LUE approach is highly flexible: GPP can be modeled by using high İ values or NPP through lower İ values. This tuning of İ to produce either GPP or NPP is probably successful due to the relatively conservative nature of autotrophic respiration (Waring et al. 1998). Similar production estimates also can be obtained by applying high İ and stringent environmental controls or low İ and loose or nonexistent environmental controls; the precise method must be guided by the characteristics of the input remote sensing dataset and the region under study. The LUE approach has been applied in many modeling approaches; we highlight two here. The Carnegie-Ames-Stanford Approach (CASA) is a well known LUE model (Potter et al. 1993) and employs the generalized LUE concepts to predict NPP. In various forms, CASA has been used to model agricultural productivity (Lobell et al. 2002), trace gas emissions (Potter et al. 2001), derive plant allocation patterns (Friedlingstein et al. 1999), and investigate the terrestrial carbon sink (Potter and Klooster 1999). The 3-PGS model (Physiological Processes Predicting Growth, satellite

Chapter 5.1: Remote Sensing Phenology

377

forcing, Coops et al. 1998) uses concepts from the silvicultural 3-PG model (Landsberg and Waring 1997) coupled to a LUE concept and simplified biogeochemical concepts (Running and Coughlan 1988) to model productivity and allocation. 3-PGS is designed to function for more localized forest stand development applications and has been used extensively to estimate forest productivity (Coops and Waring 2001a; Coops and Waring, 2001b; Coops et al. 2001). Both approaches, one designed for regional to global applications and one for local forest application, rely on the implicit use of vegetation phenology through the timing and magnitude of remotelysensed FPAR, usually at a biweekly or monthly time step. To demonstrate the concept of phenological variability in LUE models, we present 1982-1999 monthly FPAR for a single 8km pixel in the northcentral US (Figure 4). Clearly, in spite of a regular seasonal progression characteristic of a northern grassland site, there is considerable year-to-year phenological variation. The date at which FPAR fell below 0.5 ranged from August to October (extensive variability existed in winter due to variability in snow cover); and the amount and timing of maximum FPAR varied considerably. A LUE approach, although it would not use explicit phenological stages, would represent the full range of phenological variability shown in Figure 4.

Figure 5.1-4. 1982-1999 monthly FPAR for a north-central US grassland pixel. Solid line shows the average values. Shaded area shows ± one standard deviation.

378

8.

Phenology: An Integrative Environmental Science

CONCLUSIONS

Remote sensing phenology is able to consistently generate estimates of the start, peak, duration, and end of the growing season over large areas. The elements of phenology that can be estimated from remote sensing are necessarily more coarse than direct observations of individual plant phenology, such as bud burst or first leaf, but are rather summaries of the constituents of pixels and do not normally represent any one vegetation type. A variety of approaches are being used to derive phenology from remote sensing, each of which may be measuring fundamentally different phenomena of seasonality. For example, some approaches for estimating the start of the growing season may be measuring environmental conditions directly preceding the growing season, while others are measuring progressive stages of the spring greenup. The study for which the resulting estimates are being used should carefully consider which approach is most appropriate for that particular application. It is apparent that improvements in estimating phenology from remote sensing are needed. Principally, we need to understand what factors, both vegetative and non-vegetative, are influencing the vegetation index values and resulting phenology estimates. We also need focused ground surveys of phenology in a variety of environments that are specifically designed to validate remote sensing phenology. This would involve designing an approach for making large area estimates of phenology over heterogeneous study areas, rather than plant specific measures. Such an integrated effort would be logistically difficult and expensive, but would make a solid contribution toward not only documenting remote sensing phenology, but would also contribute to answering a number of global change questions.

ACKNOWLEDGEMENTS Michael White was supported by NASA grant NAG5-11282.

REFERENCES CITED Asrar, G., M. Fuchs, E. T. Kanemasu, and J. L. Hatfield, Estimating absorbed photosynthetically active radiation and leaf area index from spectral reflectance in wheat, Agronomy Journal, 76, 300-306, 1984. Badhwar, G. D., Crop emergence date determination from spectral data, Photogrammetric Engineering and Remote Sensing, 46, 369-377, 1980. Badhwar, G. D., Use of Landsat-derived profile features for spring small-grains classification, Int. J. Remote Sensing, 5(5), 783-797, 1984

Chapter 5.1: Remote Sensing Phenology

379

Chen, X. Q., C. X. Xu, and Z. J. Tan, An analysis of relationships among plant community phenology and seasonal metrics of Normalized Difference Vegetation Index in the northern part of the monsoon region of China. Int. J. Biometeorol., 45, 170-177, 2001 Coops, N. C., and R. H. Waring, Estimating forest productivity in the eastern Siskiyou Mountains of southwestern Oregon using a satellite driven process model, 3-PGS, Can. J. Forest Research (Revue Canadienne De Recherceh Forestiere), 31, 143-154, 2001a. Coops, N. C., R. H. Waring, The use of multiscale remote sensing imagery to derive regional estimates of forest growth capacity using 3-PGS, Remote Sensing Environ., 75, 324-334, 2001b Coops, N. C., R. H. Waring, and J. J. Landsberg, Assessing forest productivity in Australia and New Zealand using a physiologically-based model driven with averaged monthly weather data and satellite derived estimates of canopy photosynthetic capacity, Forest Ecology and Management, 104, 113-127, 1998. Coops, N. C., R. H. Waring, and J. J. Landsberg, Estimation of potential forest productivity across the Oregon transect using satellite data and monthly weather records, Int. J. Remote Sensing, 22, 3797-3812, 2001. Crist, E. P., Cultural and environmental effects on the spectral development patterns of corn and soybeans – field data analysis, NASA Report SR-E2-04224, Environmental Research Institute of Michigan, Ann Arbor, Michigan, 67 pp., 1982. Crist, E. P., and W. A. Malila, Development and evaluation of an automatic labeling technique for spring small grains, NASA Report SR-EL-04065, AgRISTARS Report NAS9-15476, Environmental Research Institute of Michigan, Ann Arbor, Michigan, 67 pp., 1982. Duchemin B., J. Goubier, and G. Courrier, Monitoring phenological key stages and cycle duration of temperate deciduous forest ecosystems with NOAA/AVHRR data. Remote Sensing of Environ., 67, 68-82, 1999 Eidenshink, J., The 1990 Conterminous U.S. AVHRR data set, Photogrammetric Engineering and Remote Sensing, 58, 809-813, 1992. Friedlingstein, P., G. Joel, C. B. Field, and I. Y. Fung, Toward an allocation scheme for global terrestrial carbon models, Global Change Biology, 5, 755-770, 1999. Goward, S. N., C. J. Tucker, and D. G. Dye, North American vegetation patterns observed with the NOAA-7 Advanced Very High Resolution Radiometer, Vegetatio, 64, 3-14, 1985. Huete, A. R., A soil adjusted vegetation index (SAVI), Remote Sensing Environ., 25, 295309, 1988. Huete, A. R., H. Q. Liu, K. Batchily, and W. vanLeeuwen, A comparison of Vegetation indices over a global set of TM images for EOS-MODIS, Remote Sensing Environ., 59, 440-451, 1997. Jönsson, P., and L. Eklundh. Seasonality extraction by function fitting to time-series of satellite sensor data, IEEE Transactions on Geoscience and Remote Sensing, 40, 18241832, 2002. Kaduk, J., and M. Heimann, A prognostic phenology scheme for global terrestrial carbon cycle models, Clim. Res., 6, 1-19, 1996. Kaufman, Y. J., and D. Tanré, Atmospherically resistant vegetation index (ARVI) for EOSMODIS, IEEE Transactions Geosciences and Remote Sensing, 30, 2-27, 1992. Keeling, C. D., J. F. S. Chin, and T. P. Whorf, Increased activity of northern vegetation inferred from atmospheric CO2 measurements, Nature, 382, 146-149, 1996.

380

Phenology: An Integrative Environmental Science

Landsberg, J. J., and R. H. Waring, A generalized model of forest productivity using simplified concepts of radiation use efficiency, carbon balance and partitioning, Forest Ecology and Management, 95, 209-228, 1997. Lloyd, D., A phenological classification of terrestrial vegetation cover using shortwave vegetation index imagery, International Journal of Remote Sensing, 11, 2269-2279, 1990. Lobell, D. B., J. A. Hicke, G. P. Asner, C. B. Field, C. J. Tucker, and S. O. Los, Satellite estimates of productivity and light use efficiency in United States agriculture, 1982-98, Global Change Biology, 8, 722-735, 2002. Loveland, T. R., J. W. Merchant, D. O. Ohlen, and J. F. Brown, Development of a land-cover characteristics database for the conterminous U.S., Photogrammetric Engineering and Remote Sensing, 57, 1453-1463, 1991. Los, S. O., C. O. Justice, and C. J. Tucker, A global 1º by 1º NDVI data set for climate studies derived from the GIMMS continental NDVI data, Int. J. Remote Sensing, 15, 3493-3518, 1994. MacDonald, R. B., and F. G. Hall, Global crop forecasting, Science, 208, 670-679, 1980. Malingreau, J. P., C. J. Tucker, and N. Laporte, AVHRR for monitoring global tropical deforestation, Int. J. Remote Sensing, 10, 855-867, 1989. Meltzer, M. D., R. C. Cicone, and K. I. Johnson, The evaluation of a semi-automated procedure for classifying corn and soybeans without ground data, Proceeding of the Eighth Symposium on Machine Processing of Remotely Sensed Data, 7-9 July, West Lafayette Indiana, 108-115, 1982. Menzel, A., Phenology: Its importance to the global change community, Climatic Change, 54, 379-385, 2002. Moulin, S., L. Kergoat, N. Viovy, and G. Dedieu, Global-scale assessment of vegetation phenology using NOAA/AVHRR satellite measurements, J. Climate, 10, 1154-1170, 1997. Myneni, R. B., C. D. Keeling, C. J. Tucker, G. Asrar, and R. R. Nemani, Increased plant growth in the northern high latitudes from 1981 to 1991, Nature, 186, 695-702, 1997a. Myneni, R. B., R. R. Nemani, and S. W. Running, Estimation of global leaf area index and absorbed PAR using radiative transfer models, IEEE Transactions on Geoscience and Remote Sensing, 35, 1380-1393, 1997b. Potter, C. S., S. E. Alexander, J. C. Coughlan, and S. A. Klooster, Modeling biogenic emissions of isoprene: exploration of model drivers, climate control algorithms, and use of global satellite observations, Atmospheric Environment, 35, 6151-6165, 2001. Potter, C. S., and S. A. Klooster, Detecting a terrestrial biosphere sink for carbon dioxide: Interannual ecosystem modeling for the mid-1980s, Climatic Change, 42, 489-503, 1999. Potter, C. S., J. T. Randerson, C. B. Field, P. A. Matson, P. M. Vitousek, H. A. Mooney, and S. A. Klooster, Terrestrial ecosystem production: a process model based on global satellite and surface data, Global Biogeochemical Cycles, 7, 811-841, 1993. Reed, B. C., J. F. Brown, D. VanderZee, T. R. Loveland, J. W. Merchant, and D. O. Ohlen, Measuring phenological variability from satellite imagery, J. Veg. Science, 5, 703-714, 1994. Running, S. W., and J. C. Coughlan, A general model of forest ecosystem processes for regional application. I. Hydrological balance, canopy gas exchange and primary production processes, Ecol. Modelling, 42, 125-154, 1988. Schwartz, M. D., and T. M. Crawford, Detecting energy balance modifications at the onset of spring, Phys. Geography, 5, 394-409, 2001. Schwartz, M. D., B. C. Reed, and M. A. White, Assessing satellite-derived start-of-season measures in the conterminous USA, Int. J. Climatology, 22, 1793-1805, 2002.

Chapter 5.1: Remote Sensing Phenology

381

Swets, D. L., B. C. Reed, J. D. Rowland, and S. E. Marko, A weighted least-squares approach to temporal NDVI smoothing, Proceedings of the 1999 ASPRS Annual Conference, From Image to Information, Portland, Oregon, May 17-21, 19999 [CD-ROM], Bethesda, Maryland, American Society for Photogrammetry and Remote Sensing, 2000. Townshend, J. R. G., C. O. Justice, and D. Skole, The 1 km resolution global data set: needs of the International Geosphere Biosphere Programme, Int. J. Remote Sensing, 15, 34173441, 1994. Tucker, C. J., and P. J. Sellers, Satellite remote sensing of primary productivity, Int. J. Remote Sensing, 7, 1395-1416, 1986. Tucker, C. J., D. A. Slayback, J. E. Pinzon, S. O. Los, R. B. Myneni, and M. G. Taylor, Higher northern latitude normalized difference vegetation index and growing season trends from 1982 to 1999, Int. J. Biometeorol., 45, 184-190, 2001. VanDijk, A., S. L. Callis, C. M. Sakamoto, and W. L. Decker, Smoothing vegetation index profiles: an alternative method for reducing radiometric disturbance in NOAA/AVHRR data, Photogrammetric Engineering and Remote Sensing, 53, 1059-1067, 1987. Viovy, N., O. Arino, and A. S. Belward, The best index slope extraction: A method for reducing noise in NDVI time series, Int. J. Remote Sensing, 13, 1585-1590, 1992. Waring, R. H. J. J. Landsberg, and M. Williams, Net primary production of forests: a constant fraction of gross primary production?, Tree Phiology, 18, 129-134, 1998. White, M. A., M. D. Schwartz, and S. W. Running, Young students, satellites aid understanding of climate-biosphere link, EOS Transactions, 81, 1,5, 1999. White, M. A., P. E. Thornton, and S. W. Running, A continental phenology model for monitoring vegetation responses to interannual climatic variability, Global Biogeochemical Cycles, 11, 217-234, 1997. Wiegand, C. L., A. J. Richardson, and E.T. Kanemasu, Leaf area index estimates for wheat from Landsat and their implications for evapotranspiration and crop modeling, Agronomy Journal, 71, 336, 1979. Zhang, X., J. C. F. Hodges, C. B. Schaaf, M. A. Friedl, A. H. Strahler, and F. Gao, Global vegetation phenology from AVHRR and MODIS data. Proceedings of the IEEE 2001 International Geoscience and Remote Sensing Symposium, Sydney, Australia, [CD-ROM], 2001. Zhang, X., M. A. Friedl, C. B. Schaaf, A. H. Strahler, J. C. F. Hodges, F. Gao, and B. C. Reed, Monitoring vegetation phenology using remotely sensed data from MODIS. Remote Sensing of Environ., 84, 471-475, 2003.

PART 6

PHENOLOGY OF SELECTED LIFEFORMS

Chapter 6.1 AQUATIC PLANTS AND ANIMALS Wulf Greve German Center for Marine Biodiversity Research (Senckenberg Research Institute), Hamburg, Germany

Key words:

1.

Aquatic, Marine, Fresh water, Hydrosphere, Ecology

THE HYDROSPHERE

Oceans cover 70.8% of the earth to a mean depth of 3,729 m and thereby provide a volume of 1,350 millions km3 of inhabitable biosphere (Gerlach, 1994). Childress (1983) calculated that this volume forms 99.5% of the earth's biosphere (Figure 1). This large volume of the biosphere contributes only about 40% of global primary production, however, as productivity is limited by light and by nutrients. Light penetrates just the upper 100 m of the water, and the majority of ocean zones lack nutrients, except in regions with upwelling, currents from less productive areas, or run-off from the coasts. These processes depend on wind speed, wind direction, and the rainfall pattern in the catchment areas of rivers (which in turn vary seasonally). Light, wind, temperature, and production conditions are more constant in the depths of lakes and oceans. The deep-water basins are often linked to surface biology by decaying organic sediments (Lampitt 1985). These nutrients reach the deep-sea benthos in days to weeks, transmitting a seasonal production signal to the abyssal organisms. In some lake and sea basin sediments, such materials contain records of millennia of upper-layer biotic processes, as documented in micropaleontology (Schmiedl et al. 2003) and sedimentology (Alheit and Hagen 1997) literature. Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 385-403 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

386

Phenology: An Integrative Environmental Science

RIVER FLOODPLAIN LAKE EPILIMNION HYPOLIMNION

MANGROVE FOREST SALT MARSH ESTUARY SUPRALITTORAL TTO TIDAL RANGE EULITTORAL SUBLITTORAL

SHELF SEA

DEEP SEA PELAGIAL BENTHAL ABYSSAL

Figure 6.1-1. Schematic of the hydrosphere, indicating areas addressed in this chapter.

2.

ORGANISMS OF THE HYDROSPHERE

Life first developed in the sea. Organisms of the hydrosphere live permanently within water, or temporarily depend on it for nutrition or propagation. Almost all phyla of the plant and animal kingdom live in the world oceans, while several taxonomic units are not represented in freshwater or terrestrial ecosystems (Table 1). The marine taxonomic variation spans from diatoms and dinoflagellates, supporting the bulk of primary production in the world oceans, to kelp, seagrass, and mangroves, adapted to saline water. Many other organisms depending on, but living partially outside the sea (like sea birds and the inhabitants of the salt marshes and mangrove forests), are generally treated as marine organisms. Limnetic systems like the freshwater tidal zone and areas occasionally or seasonally flooded, become part-time members of the hydrosphere. The marine littoral zone is influenced by the lunar driven tides with ranges up to 15 meters, exposing organisms to constant changes in saline water, sunshine, rain, and temperature. The following are some examples of biota where an intensive interaction of hydrological and terrestrial organisms takes place. The upper surface of oceans and lakes is inhabited by pleuston exposed to the air, next comes the neuston, organisms settling or temporarily living below the water surface. The pelagos are organisms of the water, consisting of phytoplankton and zooplankton, drifting organisms, and nekton, actively swimming organisms. The bottom of lakes and seas is inhabited by benthos, the organisms living on or within the sediment (phytobenthos, plants, and zoobenthos, animals, are distinct sub-groups). Additional details are available in the literature (e.g., Parsons et al. 1984).

Chapter 6.1: Aquatic Plants and Animals

387

Table 6.1-1. Distribution of plant and animal phyla in the aquatic realm (marine and fresh water, or limnetic) with indication of frequent („) or seldom (+) occurrence. Taxonomic units Marine Fresh Taxonomic units Marine Fresh Algae Lamellibranchia „ „ Cyanophyceae „ „ Cephalopoda „ Rhodophyceae „ + Articulata Chrysophyceae + „ Annelida „ „ Xanthophyceae + „ Pentastomida „ „ „ „ Tardigrada „ „ Bacillariophyceae „ + Arthropoda Phaeophyceae Coccolithophorida „ + Chelicerata Dinophyceae „ „ Merostomata „ Euglenophyceae + „ Arachnida + „ Chlorophyceae „ „ Pantopoda „ Prasinophyceae „ + Mandibulata Charophyceae „ Crustacea „ „ „ „ Antennata Angiosperma Gymnosperma + „ Insecta + „ Protozoa „ „ Hemichordata „ Porifera „ + Echinodermata „ Cnidaria „ + Pogonophora „ Ctenophora „ Chordata „ Tunicata „ Tentaculata Scolecida Appendicularia „ „ „ Acrania „ Plathelmintes Nemertini „ + Vertebrata Nemathelmintes „ „ Pisces „ „ Kamptozoa „ + Amphibia „ Mollusca Reptilia „ „ „ „ Aves „ „ Gastropoda Mammalia „ „

3.

PHYSICAL FORCING OF THE HYDROSPHERE

3.1

Light

Solar radiation is the prime light source for any organism. Solar radiation varies in energy intensity, wavelength, and irradiation, at hourly, daily, and seasonal time-scales, and also spatially, depending on earth-sun geometry, cloud cover, turbidity and water depth. These light conditions directly or indirectly cause physiological responses. Lunar light may also render a physiologically relevant signal for reproductive synchronization (Franke 1990).

388

3.2

Phenology: An Integrative Environmental Science

Temperature

Water is a transparent and fluid substance, with a great solubility for ions. It can be transported through the atmosphere in a gaseous state from which it is released as rain or snow under low temperatures. Snow may aggregate to layers of ice in glaciers. Since water at 4°C has the greatest density/weight, abyssal temperatures are always close to this value. Thermal stratification of water bodies extends from warmer surface waters to colder bottom waters, forming a thermocline from the mixed surface waters to the deeper layers. As water below 4°C gets lighter again, freezing begins at the water surface and surface temperatures around 4°C favor the mixing of the water column. The temperature range of the world oceans is narrower than that of terrestrial systems (-2°C to 43°C versus –68°C to 65°C) (Kinne 1963). The annual as well as the daily variation is greatest in the temperate zone, decreasing equatorward and poleward. In limnetic biota, this general order is also affected by altitude as environmental temperature decreases with altitude. Temperature changes with weather, climate, advection, and turbulent diffusion. The heat content of a water body is of decisive importance for the organisms living within it (Pohlmann 1996). Extreme changes in seasonal temperatures may be especially useful for understanding the physiological responses in any organism (Orton 1920).

3.3

Moon

According to regional topography, the depth of ocean water is altered by the gravity of the moon and, to a lesser extent, the sun. If both forces supplement each other at “full moon” and at “new moon,” twice monthly “spring tides” with higher high tides and lower low tides occur. The tidal cycle includes one high tide and one low tide and has a length of ~12.25 hours. The tidal range varies from a few centimeters in enclosed seas or the open ocean, to as much as ten or more meters in some shelf seas. Besides being the driving force of the tides, the moon also provides especially bright and especially dark nights at “full moon” and at “new moon.” These events also provide cues to animal behavior making use of the extremely diverse spring tides (Caspers 1951; Neumann, 1967).

3.4

Currents

The motion of water bodies due to: 1) changes in altitude in rivers; 2) varying precipitation or thawing of ice and snow; 3) wind-driven or tidal currents; or 4) the global and regional ocean transport system, dislocates planktonic organisms within the water body and supplies sessile benthos and

Chapter 6.1: Aquatic Plants and Animals

389

nekton with possible food sources. Besides environmental forcing within the water body, the ATTZ (aquatic terrestrial transition zone) is greatly modified by current dependent sea-level changes. In the marine environment, tides are the primary forces that form the littoral zone (through standard and spring tides). Extreme, wind driven high tides can cover the supra-littoral zone with salt marshes. Similar longer-lasting current-based events, are the flood pulses forming temporary river flood plains found in the tropics (Junk 1989). Such temporary aquatic habitats are members of both aquatic and terrestrial systems (Lake 1995).

4.

PHYSIOLOGICAL RESPONSES OF MARINE AND LIMNETIC ORGANISMS

Any population of organisms is exposed to a complex pattern of environmental forcing. Which factors determine phenological responses depends on the species-specific physiological reception mode of the population. According to the ecological niche concept (Elton 1927), any environmental parameter may be important to the physiological responses of the population. If the actual value lies beyond the range of the physiological tolerance of the species, it excludes it from the system. If it lies within the tolerance range, the species-specific physiological response profile leads to a phenological response (Figure 2). Many investigations are concerned with the determination of the tolerance zone of marine and limnetic organisms (Sewell and Young 1999). Single environmental forcing may be observed as well as a combination of environmental cues that influence population responses. Temperature, salinity, and oxygen (Alderdice and Forrester 1968) simultaneously determine the decisive ecological niche of fish. Equally, light and temperature determine the reproduction of brown algae (Henry 1988). Threshold transitions (Werner 1962), physiological forcing by parameter aggregations (Lange and Greve 1997), or combinations of forcing (such as temperature and day-length on the reproductive physiology of the viviparous sea-perch, Cymatogaster aggregata) are decisive cues (Harrington 1959). Temperature controls physiological processes in poikilotherms (nearly all marine organisms according to the Q10 rule, see (Beleheradek 1935). This general functional relationship underlies the species- and process-specific thermal profile, including the limits of temperatures tolerated, and the physiological preference in which differences become especially evident in populations of trophologically sympatric species (Heyen et al. 1998). This element of functional biodiversity deserves increased analysis of its functional and genetic basis (Figure 3, Uhlig and Sahling 1995).

390

Phenology: An Integrative Environmental Science

Figure 6.1-2. Example of a species-specific physiological response to an environmental parameter.

Figure 6.1-3. The thermal profile of reproduction in Noctiluca scintillans, data source Uhlig and Sahling (1995).

5.

SEASONALITY IN THE HYDROSPHERE

Seasonality varies mainly with latitude, altitude, or depth, and large scale climatic forcing such as ENSO (El Nino Southern Oscillation) or NAO (North Atlantic Oscillation). Limnetic systems are closely related to the surrounding terrestrial systems, which may permit the correlation of

Chapter 6.1: Aquatic Plants and Animals

391

terrestrial (e.g., time of cherry blossom) and hydrological signals (e.g., water temperature) with pike-perch year-class success (estimated from abundance of copepod nauplii, Tesch 1962). The scientific treatment of hydrological biota varies more with respect to the pelagial or benthal than to the marine or limnetic community. In plankton research, the food chain paradigm, and a resulting concentration on the succession of trophic levels or guilds, has suppressed investigation of the specific response profiles to environmental forcing of species populations (such as is seen in benthic and fisheries research). The seasonal signal resulting from solar radiation and temperature is integrated into the production cycle of the polar and temperate zone. In general, the description of annual succession is based on trophic relationships (Lenz 2000). However, the succession of populations of a comparable ecological guild can also be described by forcing parameters other than trophology (Greve and Reiners 1988). In contrast, benthology, (especially phytobenthos research) has investigated the physiological responses of single species, mainly in the laboratory (Lüning and tom Dieck 1989) and in studies of geographic distribution (Molenaar et al. 1997). Seagrass and mangroves (as other gymnosperms and angiosperms) can live either continuously or temporarily within the hydrosphere. These organisms have a much longer generation time than planktonic populations and can thus follow the traditional seasonal pattern of budding, blossoming, fruiting, and leaf-fall. Benthic and nektonic animals respond to both the seasonal timing of the benthic production and population succession in open water, where many species live during their larval periods, making use of the pulsed system of suitable food production. The ontogeny of many species includes migration between marine biota and fresh water, sediment and water, or the surrounding terrestrial system and water. Floods or droughts increase variation of the hydrosphere, and organisms have evolved strategies to utilize the new biota to survive unfavorable dry periods. Marine bentho-pelagic coupling includes diurnal migrations from the sediment into the pelagial zone and back, ontogenetic seasonal transfer of sperm (see Greve 1974), eggs, and larvae into the pelagial, and the return of recruits to the benthic or nektonic populations. Further, resting stages such as dormant eggs (Dahms 1995) or phytoplankton cysts add to this bentho-pelagic exchange, especially in limnetic systems and shelf seas. The alternation of generations between polyp and medusiod stages in hydrozoa is another example of bentho-pelagic coupling. All these processes are seasonal and periodic, related either to the length of the day, lunar driven tides, or seasons of the year. They thus offer options to phenological research, though so far the hierarchy of external or internal functional relationships determining the phenology of the

392

Phenology: An Integrative Environmental Science

populations is not generally known. It will be a major effort to understand the primary functional relationships determining the phenology of populations. Two research strategies need to be developed separately: traditional plankton research that discriminates between functional groups (Lenz 2000), and population research that discriminates between speciespopulations with various physiological, trophodynamic, and temporal preferences (Giese 1959; Lindley and Batten 2002).

6.

METHODS OF AQUATIC PHENOLOGY

The classical phenological criteria of plants can applied to organisms with an annual ontogenetic cycle or inter-annually lasting biomass, such as those in the aquatic terrestrial transition zone (ATTZ), which includes the benthic environment as well as birds, seals, and tortoises that feed on marine production. The majority of the hydrosphere is inhabited by plankton with generation times of days, weeks, or months, that results in extreme population dynamics (changes of several orders of magnitude within weeks in the phytoplankton and zooplankton). Some of these changes are due to bentho-pelagic coupling through the cysts and dormant eggs of pelagic species (Dahms 1995) or through larvae of benthic organisms or fish (meroplankton). With such high dynamics, the abundance of the population under investigation itself can be a phenological criterion. The passage of absolute or relative threshold values permits the temporal definition of phenological events (Greve et al. 2001) although it requires collection of a complete year’s data before starting the analysis. Sufficiently frequent measurements over multiple years are required for investigation of the variance in phenology and its functional relationship to physical forcing (Colebrook 1960; Heyen et al. 1998). Measurements on the population dynamics of limnetic systems often have been made on a weekly (Monday-Friday) basis. Measurements in marine systems generally require a more intensive effort from sea-going platforms. Light vessels, weather-ships (Oestvedt 1955) and offshore islands (Greve and Reiners 1988) provide alternatives. There is neither a marine biometeorological observation system (as in terrestrial agrometeorological observation), nor an equivalent to terrestrial phenological gardens (Menzel, 2000). The continuous plankton recorder (CPR) program is the most extensive observation series for marine plankton (Colebrook 1978). Measurements cover the North Atlantic and the North Sea, with monthly sections repeated for a periods of 40 or more years. Besides these measurements, local time-series from the British Channel (near Plymouth or at the station Helgoland Roads and other locations, ICES report 2002) have

Chapter 6.1: Aquatic Plants and Animals

393

been used for the analysis of long-term population and benthos dynamics (Kröncke et al. 1998; Reid et al. 1998; Greve et al. 2001) and for analysis of climatic forcing within marine populations, including phenological findings (Southward et al. 1995).

7.

THE STATE OF PHENOLOGY IN MARINE AND LIMNETIC SYSTEMS

Organisms of the hydrosphere respond similarly to seasonal and climatic forcing as those of terrestrial systems (Walther et al. 2002). The degree to which these responses are documented varies substantially. This includes a broad collection of population data on the annual timing of life histories with special attention to reproductive synchronization. So far, these studies are only partially related to phenology.

7.1

Thallus Algae

Often in marine botany, physiological responses of thallus algae (studied in detail from laboratory and field observations) are used to confirm the validity of derived functional relationships (Lüning and tom Dieck 1989). With this approach, global latitudinal comparisons of phenological responses could be studied in seaweeds (Wiencke et al. 1994; Molenaar et al. 1997). Examples of botanical phenological observations exist for Chlorophyta (Clifton and Clifton 1999), Rhodophyta (Arasaki 1981; Chamberlain, 1985; Breeman et al. 1988), and Phaeophyta (Henry 1988; Deshmukhe and Tatewaki 2001). In these studies reproductive behavior was studied in the field, where tetrasporophytes and gametophytes appeared to be strictly regulated by temperature/day-length responses (Breeman et al. 1988).

7.2

Marine Higher Plants

Growth, reproduction, and leaf fall were documented and timed for sea grass (Reyes et al. 1995; Oliveira et al. 1997), and mangroves (Gwada et al. 2000).

7.3

Flood Plain Forests

The phenology of tropical flood plain forests has been extensively studied, along with the accompanying rich fauna (Junk et al. 1989). The phase transition of a terrestrial ecosystem being converted into an aquatic

394

Phenology: An Integrative Environmental Science

habitat for months with fish immigrating into the canopy of trees and back again into a forest system is pulsed by seasonal floods. Phenology is applied here to the terrestrial as to the hydrological system.

7.4

Marine Benthic Communities

In zoology, the timing of reproduction was the subject of phenological studies on echinodermata (Stanwell-Smith and Peck 1998; Garrido and Barber 2001), polychaeta spawning synchrony (Hardege and Bentley 1997), gonad maturity and spawning in bivalve mollusks. In the sepiolid squid Rossia pacifica, the hatching of eggs appeared to be keyed to the new moon, and extended over more than two months. Low seawater temperatures may affect this process (Summers 1985).

7.5

Marine Vertebrates

The gonadal maturation of fish has been extensively studied. The silverside Atherinomorus lacunosus is able to adjust its phenology to environmental conditions (Conand 1993). In several species of gulls, molt has a species-specific phenology (Howell et al. 1999).

7.6

Limnetic Animals

Fresh water systems are rich in populations living part-time in the water such as insects, amphibians, and some crustaceans. Besides hatching information, phenological data are used for life history strategy analysis of Chironomidae (Mihuc and Toetz 1996). In twenty-nine species of aquatic dance flies (Empididae ( , Clinocerinae and Hemerodromiinae) it was possible to discriminate the variance of temporal orientation for closely related species in the same area (Wagner and Gathmann 1996). Even the autumnal phenology of odonata reveals a species-specific phenology (Joedicke 1998). In the amphibian Rana yavapaiensis, the timing of reproductive periods corresponds to ecological advantages (Sartorius and Rosen 2000).

7.7

Limnetic Plankton

In the epilimnion of lakes, the succession of populations is well, though rarely, documented with phenology (Adrian 1997). Investigation of the iceduration period provides information on the phenology of a forcing parameter. The clear water phase (standing for the greatest filtration efficiency of zooplankton) also indicates a phenological timing that responds

Chapter 6.1: Aquatic Plants and Animals

395

to spring temperature and dominant forcing of the North Atlantic Oscillation (NAO, Adrian et al. 1999). Consequences of climatic changes were also studied in lakes (Straile and Adrian 2000; Gerten 2001). Experimental global warming has been the subject of investigations in rivers (Hogg et al. 1995).

7.8

Marine Dinoflagellates

The NAO is also correlated with population dynamics of the marine Further, the dinoflagellate Noctiluca scintillans (Heyen et al. 1998). seasonality of N. scintillans shows a correlation with seasonal ambient temperatures (Greve et al. 2001).

7.9

Marine Zooplankton

In the life history of highly dynamic populations, triggers or indicators of improved nutritional conditions could be very helpful. In the copepod Calanus finmarchicus, Miller et al. (1991) suspected that photoperiod ended the resting period in February-March and induced maturation. An internal long-range timer is also suspected of ending the resting period. However, as the availability of food and the number of predators can hardly be tested in advance, and the production of juveniles is energy- and time-costly, reproduction is an optimization strategy. The match-mismatch hypothesis (Cushing 1990) describes this dilemma, especially for meroplanktonic populations that generally have no parental care. Animal migration provides a further example of marine phenological optimization, such as in the timing of the first appearance of immigrating Crangon crangon and Pleuronectes platessa into the spring Wadden Sea (Van der Veer and Bergmann 1987), the first fish-feeding in competing populations (Juanes et al. 1994), and the mutual predation of Pleurobrachia pileus and Calanus helgolandicus representing a dichotomy in the system equilibria (Greve 1995). Variability in the phenology of marine populations could help explain and predict the influences of global warming and other forces on ecological equilibrium changes, which in turn decide the fate of the complete biocoenosis (Greve et al. 2001). This scenario can be observed in North Sea plankton, where high frequency zooplankton time-series are available. Since 1975, every Monday, Wednesday, and Friday two plankton samples (150 and 500) are collected at the Helgoland Roads station (54°´11´18”N 7°54´E). The analysis of the population phenology resulted in calculation of an annual 15% cumulative abundance value, which corresponds with the mean SST (surface salinity temperature) over the April to June period for juvenile ctenophores (Figures 4 and 5).

Phenology: An Integrative Environmental Science 80

26

60

22

40

18

20

week

396

14

temperature-sum I-III Pleurobrachia pileus juv. 1994

1993

1992

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

10 1975

0

year

Figure 6.1-4. Winter temperature (°C) and start of season of P. pileus at Helgoland Roads.

Figure 6.1-5. Correlation of a phenological criterion (start of season of juvenile Pleurobrachia pileus) with ambient mean water temperature.

The heat content of the North Sea changed at the end of the 1980s (Pohlmann 1996). SST in the winter months showed this change in correspondence with the NAO. The response of the P. pileus population was highly non-linear. Until 1987, the ctenophore showed a distinct spring bloom in May/June, with abundance increases of more than four orders of magnitude. Since then (for the period analyzed), winter abundances increased and summer abundances decreased. Abundance changes also decreased in speed and dimension. Since P. pileus is a key species of the German Bight ecosystem (it controls the copepod populations during the

Chapter 6.1: Aquatic Plants and Animals

397

times of maximum abundance, see Greve and Reiners 1988) the change in phenology has important consequences for the ecosystem equilibrium. This change is one of the few documented examples for non-linear climate-related ecosystem responses. Similar changes have been analyzed in competing populations, such as clupeid fishes (Alheit 1997). Global warming consequences, such as the lateral displacement of populations (Southward et al. 1975) have been documented, but are not yet a topic in global zooplankton research (Marine Zooplankton Colloquium 2001), even though complete biocoenoses in the temperate zone change their stability regimes (Figure 6). A cluster analysis of 153 taxa of zooplankton timeseries at Helgoland Roads (Greve and Reiners 1995) reveals three main biocoenoses for the spring, summer, and fall-winter seasons. During the 1975-1994 period, a phase of approximately 10 years showed a relatively stable temporal distribution, and then the start of the summer biocoenoses began to advance from week 23 to week 17. The end of summer also shifted from week 32 to week 30. Further, the start of spring biocoenoses shifted from week 11 to week 4. This alteration corresponds with and supports the species-specific phenological results observed and calculated for Helgoland Roads zooplankton (Figure 7, Greve et al. 2001).

8.

PHENOLOGICAL OPTIONS IN APPLIED ECOLOGY

The phenological analysis of marine and limnetic populations has demonstrated that the timing of life history events also responds to changes in physical forcing (Walther et al. 2002). The expected continuation of global temperature increases will alter life in the aquatic biocoenoses, for which biometeorological measurements provide a valuable source of information. These observations will have to follow a variety of research strategies, as benthic, planktonic, nektonic, and submerse terrestrial populations will continue to require diverse measurements, in-situ observations of sedentary and plankton organisms, and physiological laboratory studies. Ecosystem management will have to include phenologically determined functional relationships, secondary match/mismatch consequences, lateral displacement of populations, and community changes in the administrative framework (Avila et al. 1996).

398

Phenology: An Integrative Environmental Science

1994 1992 1990

Ind. * m-3

1988

10000

year

1986 1000 1984

100

1982

10

1980

1

1978 1976 1 2 3 4 5 6 7 8 9 10 11 12

month

Figure 6.1-6. A phase transition of the ecosystem equilibrium in 1988/1989, indicated by a shift in the annual abundance cycle from periodic blooms, to permanent mean abundances of the local key–species P. pileus in the German Bight.

Chapter 6.1: Aquatic Plants and Animals

399

Figure 6.1-7. Forward shift of zooplankton biocoenoses, derived from cluster-analyses of 153 taxa at the station Helgoland Roads.

Observation strategies in the hydrosphere are based on diverse population-specific techniques, which could be improved by inclusion and documentation of phenological criteria. The resulting scientific analyses and standardization of techniques may lead to a more universal approach to hydrological phenology, and systems for collecting those parameters which are of special administrative importance and observational feasibility. Such monitoring systems can be the basis of early warning, operative administrative modeling, and sustainable management of organic resources in a changing world, and could be imbedded into Global Observing Systems (e.g., GOOS). Mankind’s attitude towards terrestrial natural systems has been shaped by the perspective of farmers and gardeners, who are actively involved in the care of plants and animals. Attitudes about the sea, where humans have traditionally been only fishers and hunters, are now slowly changing due to increased awareness of aquaculture. Marine phenological research will profit from this change.

ACKNOWLEDGEMENTS Rita Adrian and Inka Bartsch, who provided information on their fields of research, supported this study. The Helgoland Roads time-series analysis was undertaken with the support of grants DFG 282/3-1,2 and BMBF 03F181A.

400

Phenology: An Integrative Environmental Science

REFERENCES CITED Adrian, R., Calanoid-cyclopoid interactions: evidence from an 11-year field study in a eutrophic lake, Freshwater Biology, 38, 315-325, 1997. Adrian, R., N. Walz, T. Hintze, S. Hoeg and R. Rusche, 1999, Effects of ice duration on plankton succession during spring in a shallow polymictic lake, Freshwater Biology, 41, 621-632. Alderdice, D. F. and C. R. Forrester, Some effects of salinity and temperature on early development and survival of English sole ((Parophyrus vetulus), J. Fish. Res. Bd. Can., 25, 495-521, 1968. Alheit, J., and E. Hagen, Long-term climate forcing of European herring and sardine populations. Fish. Oceanogr., 6, 130-139, 1997. Arasaki, S., A comparison of the phenology of intertidal Porphyra on the coasts of Japan and western North America, Proceedings of the International Seaweed Symposium, 8, 273277, 1981. Avila, M., R. Otaiza, R. Norambuena and M. Nunez, Biological basis for the management of 'luga negra' (Sarcothalia crispata Gigartinales, Rhodophyta) in southern Chile, Hydrobiologia, 326-327, 245-252, 1996. Beleheradek, J., Temperature and living matter, in Protoplasma Monogr., Borntraeger, Berlin, 1935. Breeman, A. M., E. J. S. Meulenhoff and M. D. Guiry, Life history regulation and phenology of the red alga Bonnemaisonia hamifera, Helgol. Meeresunters., 42(3-4), 535-551, 1988. Caspers, H., Rhythmische Erscheinungen in der Fortpflanzung von Clunio marinus (dipt. Chiron.) und das Problem der lunaren Periodizität bei Organismen, Arch. Hydrobiol. (Suppl. Bd), 18, 415-594, 1951. Chamberlain, Y. M., Trichocyte occurrence and phenology in four species of Pneophyllum (Rhodophyta, Corallinaceae) from the British Isles, Br. Phycol. J., 20, 375-379, 1985. Childress, J., Oceanic Biology: Lost in Space?, in Oceanography, the present and the future, edited by G. B. Peters, pp. 127-135, Springer-Verlag, New York, 1983. Clifton, K. E. and L. M. Clifton, The phenology of sexual reproduction by green algae (Bryopsidales) on Caribbean coral reefs, J. Phycology, 35, 24-34, 1999. Colebrook, J. M., Continuous Plankton Records: Methods of analysis, 1950-1959, Bull. Mar. Ecol., 5, 51 – 64, 1960. Colebrook, J. M., Continuous Plankton Records: Zooplankton and Environment, North-East Atlantic and North Sea, 1948 – 1975, Oceanol. Acta, 1(1), 9 – 23, 1978. Conand, F., Life history of the silverside Atherinomorus lacunosus (Atherinidae) in New Caledonia, J. Fish Biology, 42, 851-863, 1993. Cushing, D. H., Recent studies on long term changes in the sea, Freshwater Biol., 23, 71-84, 1990. Dahms, H. U., Dormancy in the Copepoda -- an overview, Hydrobiologia, 306, 199-211, 1995. Deshmukhe, G. V. and M. Tatewaki, Phenology of brown alga Coilodesme japonica (Phaeophyta, Dictyosiphonales) with respect to the host-specificity along Muroran coast, North Pacific Ocean, Japan, Indian Journal of Marine Sciences, 30, 161-165, 2001. Elton, C., Animal Ecology, Methuen, London, 207 pp., 1927. Franke, H.-D., Photopollution: Coastal artificial light affects reproductive synchronisation in a litoral polychaete, Verhandlungen der Deutschen Zoologischen Gesellschaft, 83, 481, 1990.

Chapter 6.1: Aquatic Plants and Animals

401

Garrido, C. L., and B. J. Barber, Effects of temperature and food ration on gonads and oogenesis of the green sea urchin, Strongylocentrotus droebachiensis, Marine Biology, 138, 447-456, 2001. Gerlach, S. A., Marine Systeme, Springer, Berlin, 226 pp., 1994. Gerten, D. and R. Adrian, Differences in the persistency of the North Atlantic Oscillation signal among lakes, Limnol. Oceanogr., Vol. Suppl., 448-455, 2001 Giese, A. C., Comparative physiology: annual reproductive cycles of marine invertebrates, A. Rev. Physiol., 21, 547-576, 1959. Greve, W., Planktonic Spermatophores found in a culture device with spionid Polychaetes, Helgol. Wiss. Meeresunters, 26, 370-374, 1974. Greve, W., Mutual predation causes bifurcations in pelagic ecosystems: the simulation model PLITCH (PLanktonic swITCH), experimental tests, and theory, ICES Journal of marine Science, 52, 505 – 510, 1995. Greve, W., U. Lange, F. Reiners and J. Nast, Predicting the Seasonality of North Sea Zooplankton, in Burning issues of North Sea ecology, Proceedings of the 14th international Senckenberg Conference North Sea 2000, edited by I. Kröncke, M. Türkay and J. Sündermann, pp. 263-268, Senckenbergiana marit., Frankfurt am Main, 2001. Greve, W., and F. Reiners, Plankton time - space dynamics in German Bight - a systems approach, Oecologia, 77, 487-496, 1988. Greve, W. and F. Reiners, Biocoenotic process patterns in German Bight, in Biology and Ecology of Shallow Coastal Waters, edited by A. Eleftheriou, A. Ansell and C. J. Smith, pp. 67-71, Olsen and Olsen, Fredensborg, Denmark, 1995. Gwada, P., T. Makoto and Y. Uezu, Leaf phenological traits in the mangrove Kandelia candel (L.) Druce, Aquatic Botany, 68, 1-14, 2000. Hardege, J. D., and M. G. Bentley, Spawning synchrony in Arenicola marina: evidence for sex pheromonal control, Proc. Royal Soc. London, Series B. Biol.Sci. 264, 1941-1047, 1997. Harrington, R. W., Effects of four combinations of temperature and daylength on the ovogenetic cycle of a low-latitude fish, Fundulus confluentus GOODE and BEAN, Zoologica, N. Y., 44, 149-168, 1959. Henry, E. C., Regulation of reproduction in brown algae by light and temperature, Botanica Marina, 31, 353-357, 1988. Heyen, H., H. Fock, and W. Greve, Detecting relationships between the interannual variability in ecological time series and climate using a multivariate statistical approach - a case study on Helgoland Roads zooplankton, Clim. Res., 10, 179-191, 1998. Hogg, I. D., D. D. Williams, J. M. Eadie and S. A. Butt, The consequences of global warming for stream invertebrates: A field simulation, J. Thermal Biol., 20, 199-206, 1995. Howell, S. N. G., J. R. King and C. Corben, First prebasic molt in herring-, Thayer's-, and glaucous-winged gulls, J. Field Ornithology, 70, 543-554, 1999. International Council for the Exploration of the Sea (ICES), Report of the Working Group on Zooplankton Ecology, ICES CM, C:07, 1-57, 2002. Joedicke, R., Autumnal phenology of central European Odonata. 2. Observations in the y Opuscula Zoologica Fluminensia (Opusc. Zool. Flumin.), Lower Rhine Region, Germany, 159, 1-20, 1998. Juanes, F., J. A. Buckel and D. O. Conover, Accelerating the onset of piscivory: intersection of predator and prey phonologies, J. Fish Biology, 45, 41-54, 1994. Junk, W. J., P. B. Bayley and R. E. Sparks, The Flood Pulse Concept in River-Floodplain Systems, in Proceedings of the International Large River Symposium, edited by D. P. Dodge, pp. 110-127, 1989.

402

Phenology: An Integrative Environmental Science

Kinne, O., The effects of temperature and salinity on marine and brackish water animals, I. Temperature. Oceanogr. Mar. Biol. A. Rev., 1, 301-340, 1963. Kröncke, I., J. W. Dippner, H. Heyen and B. Zeiss, Long-term changes in macrofaunal communities off Norderney (East Frisia, Germany) in relation to climate variability, Marine Ecology Progress Series, 167, 25-36, 1998. Lake, P. S., Of Floods and Droughts: River and Stream Ecosystems of Australia, in River and Stream Ecosystems, edited by C. E. Cushing, K. W. Cummins, and G. W. Minshall, pp. 659-694, Elsevier, Amsterdam, 1995. Lampitt, R. S., Evidence for the seasonal deposition of detritus to the deep-sea floor and its subsequent resuspension, Deep-Sea Research, 32, 885-897, 1985. Lange, U. and W. Greve, Does temperature determine the spawning time, recruitment and distribution of flatfish via its influence on the rate of gonadal maturation?, Deutsche Hydrographishe Zeitschrift, 49(2/3), 251-263, 1997. Lenz, J., Introduction, in Zooplankton Methodology Manual, edited by P. H. W. R. P. Haris, J. Lenz, H. R. Skjoldal, and M. Huntley, pp. 1-32, Academic Press, San Diego, 2000. Lindley, J. A., and S. D. Batten, Long-term variability in the diversity of North Sea zooplankton, J. Mar. Biol. Ass U.K., 82(3), 1-40, 2002. Lüning, K., and I. tom Dieck, Environmental Triggers in Algal Seasonality, Botanica Marina, 32, 389-397, 1989. Marine Zooplankton Colloquium, Future marine zooplankton research- a perspective, Mar. Ecol. Prog. Ser., 222, 297-308, 2001. Menzel, A., Trends in phenological phases in Europe between 1951 and 1996, Int. J. Biometeorol., 44, 76-81, 2000. Mihuc, T. B. and D. W. Toetz, Phenology of aquatic macroinvertebrates in an alpine wetland, Hydrobiologia, 330, 131-136, 1996. Miller, C. B., T. J. Cowles, P. H. Wiebe, N. J. Copley, H. Grigg, Phenology in Calanus finmarchicus; hypotheses about control mechanisms, Mar. Ecol. Prog. Ser., 72, 79-91, 1991. Molenaar, F. J., A. M. Breemann, and L. A. H. Venekamp, Latitudinal Trends in the Growth and Reproductive Seasonality of Delesseria sanguinea, Membranoptera alata, and Phycodrys rubens (Rhodophyta), J. Phycol, 33, 330-343, 1997. Neumann, D., Genetic adaption in emergence time of Clunio populations to different tidal conditions, Helgoländer wiss. Meeresunters, 15, 163-171, 1967. Oestvedt, O.-J., Zooplankton investigation from Weather Ship M in the Norwegian Sea, 1948 – 1949, Hvalradets Skrifter, Scientific Results of Marine Biological Research, 40, 1-93, 1955. Oliveira, E. C., T. N. Corbisier, V. R. De Eston and O. Ambrosio, Jr., Phenology of a seagrass ((Halodule wrightii) bed on the southeast coast of Brazil, Aquatic Botany, 56, 25-33, 1997. Orton, J. H., Sea-temperature, breeding and distribution of marine animals, J. Mar. biol. Ass. U.K., 12, 330-366, 1920. Parsons, T. R., M. Takahashi and B. Hargrave, Biological Oceanographic Processes 3rd. edition, Pergamon Press, Oxford, New York, 330 pp., 1984. Pohlmann, T., Simulating the heat storage in the North Sea with a three-dimensional circulation model, Cont. Shelf Res., 16, 195-213, 1996. Reid, P. C., B. Planque, and M. Edwards, Is observed variability in the long-term results of the Continuous Plankton Recorder survey a response to climate change?, Fisheries Oceanography, 7 (3/4), 282-288, 1998.

Chapter 6.1: Aquatic Plants and Animals

403

Reyes, J., M. Sanson and J. Afonso-Carrillo, Distribution and reproductive phenology of the seagrass Cymodocea nodosa (Ucria) Ascherson in the Canary Islands, Aquatic Botany 50, 171-180, 1995. Sartorius, S. S., and P. C. Rosen, Breeding phenology of the lowland leopard frog ((Rana yavapaiensis): Implications for conservation and ecology, Southwestern Naturalist, 45, 267-273, 2000. Sewell, M. A., and C. M. Young, Temperature limits to fertilization and early development in the tropical sea urchin Exhinometra lucunter, J. Exp. Mar. Biol. Ecol., 47, 291-305, 1999. Schmiedl G., A. Mitschele, S. Beck, K. Emeis, C. Helleben, H. Schulz, M. Sperling, Benthic foraminiferal record of ecosystem variability in the eastern Mediterranean Sea during times of saprobel S5 and S6 deposition, Palaeogeography, Palaeoclimatology, Palaeoecology, 190, 139-164, 2003. Southward, A. J., E. I. Butler and P. Pennycuick, Recent cyclic changes in climate and in abundance of marine life, Nature, 253, 714-717, 1975. Stanwell-Smith, D., and L. S. Peck, Temperature and embryonic development in relation to spawning and field occurrence of larvae of three Antarctic echinoderms, Biol. Bull. Mar. Biol. Lab. Woods Hole, 194, 44-52, 1998. Straile, D., and R. Adrian, The North Atlantic Oscillation and plankton dynamics in two European lakes - two variations on a general theme, Global Change Biology, 6, 663-670, 2000 Summers, W. C., Ecological implications of life stage timing determined from the cultivation of Rossia pacifica (Mollusca: Cephalopoda), Vie Milieu., 35(3/4), 249-254, 1985. Tesch, F. W., Witterungsabhängigkeit der Brutentwicklung und Nachwuchsförderung bei Lucioperca lucioperca L., Kurze Mitteilungen aus dem Institut für Fischereibiologie der Universität Hamburg 12, 37-44, 1962. Uhlig, G., and G. Sahling, Noctiluca scintillans: zeitliche Verteilung bei Helgoland und räumliche Verbreitung in der Deutschen Bucht (Langzeitreihen 1970 -1993), Ber. Biol. Anst. Helgoland, 9, 1-127, 1995. Van der Veer, H.W., and M. J. N. Bergmann, Predation by crustaceans on a newly settled Ogroup plaice Pleuronectes platessa in the Western wadden sea, Mar. Ecol. Prog. Ser., 35, 203-215, 1987. Wagner, R. and O. Gathmann, Long-term studies on aquatic dance flies (Diptera, Empididae) 1883-1993: Distribution and size patterns along the stream, abundance changes between years and the influence of environmental factors of the community, Archiv fuer Hydrobiologie, 137, 385-410, 1996. Walther, G.-R., Eric Post, Peter Convey, Annette Menzel, Camille Parmesan, Trevor J. C. Beebee, Jean-Marc Fromentin, Ove Hoeg-Guldberg and Franz Bairlein, Ecological responses to recent climate change, Nature, 416, 389-395, 2002. Werner, B., Verbreitung und jahreszeitliches Auftreten Rathkea octopunctata (M. Sars) und Bougainvillia superciliaris (L. Agassiz), (Athecata-Anthomedusae), Ein Beitrag zur kausalen marinen Tiergeographie, e Kieler Meeresforsch, 18, 55-66, 1962. Wiencke, C., I. Bartsch, B. Bischoff, A. F. Peters and A. M. Breeman, Temperature Requirements and Biogeography of Antarctic, Arctic and Amphiequatorial Seaweeds. Botanica Marina, 37, 247-259, 1994.

Chapter 6.2 INSECTS Karen Delahaut Department of Horticulture, University of Wisconsin-Madison, Madison, WI, USA

Key words:

1.

Insects, Agriculture, Horticulture, Pest management, Degree-days

INTRODUCTION

For centuries, farmers, gardeners, and scientists have been using phenology to assist in determining the presence of specific life stages of insects. As far back as the mid-18th century, temperature and time have been used to predict insect development. Because insects are poikiloterms their growth is temperature dependent. However, temperature isn’t the sole environmental factor responsible for the rate of insect development; photoperiod and the availability of food resources may also play a role. This chapter explores the use of insect phenology using specific examples of agronomic pests to illustrate the concepts discussed. Although all plants and animals exhibit variability within a species, because insects are generally small, short-lived, abundant, and have relatively short life cycles, individual activity can vary significantly between seasons in a given year or in different years. Insect phenology models therefore vary greatly with respect to their precision and accuracy, and may have lower confidence limits as a result of the greater variation within the species than other invertebrates (Lieth 1974). Seasonality in insects is often a reflection of temperature, humidity and/or photoperiod, with temperature being critical in determining the rate of development or reproductive maturation. Photoperiod is important in determining the onset and end of diapause (suspension of development) in Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 405-419 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

406

Phenology: An Integrative Environmental Science

many species. Seasonal patterns of insect abundance include the arrival of migrants, and dates of appearance and duration of successive life stages. Seasonality can also determine the sharpness of population peaks. Univoltine insects typically have a single abundance peak per year, while bivoltine insects have two, and multivoltine species have three or more. As latitude decreases, the number of generations may also increase. In temperate latitudes, seasonal migration may play a role in the timing of insect seasonality. This migration may be between climatic regions such as to a warmer climate to overwinter, or simply to a suitable hibernation site within a single region (Wolda 1988). Knowledge of insect phenology is primarily used in economic entomology for the timing of pest management activities to target the most susceptible life stage of an insect pest. Because many factors such as latitude, altitude, and maritime effects affect local climatic conditions, the solar calendar is not usually reliable in timing pest management activities.

2.

TEMPERATURE AND DEVELOPMENTAL THRESHOLDS

Temperature is important in activating many insects from dormancy or aestivation and determining their rate of development. Rate-limiting, biochemical reactions determine how fast an insect develops. It is often assumed that development rate is a linear function of temperature above a lower threshold. Once the optimum developmental temperature is reached, insect growth slows, and then declines rapidly. Plotting the developmental rate against temperature will thus produce a bell-shaped curve. When insects are exposed to fluctuating temperatures, overall development often accelerates when compared with the same species exposed to a constant temperature of the same average. This is a wellunderstood phenomenon that is related to the strong non-linearity of developmental response to temperature, the Kaufman effect (Worner 1992). Developmental threshold temperatures are specific. Most insects become active within a range of temperatures between 39°F and 52°F (4°C and 11°C). For many insects, this threshold is near 50°F (10°C). Lower thresholds are often difficult to determine because of the ability of insects to survive at low temperatures without any noticeable development. Therefore, they are often determined by the extrapolation of the linear section of the developmental curve to where it crosses the temperature axis (Arnold 1959). Thus, lower thresholds are statistical quantities and may or may not be actual physiological cutpoints.

Chapter 6.2: Insects

407

There is also an upper threshold for development above which there is no appreciable increase in the rate of development. This temperature is often around 86°F (30°C). Within the bounds of upper and lower thresholds, the rate of an insect’s metabolism is often assumed to be linearly related to temperature, although in reality this relationship is non-linear as a result of complex enzyme reactions involved in the metabolic process.

3.

PHENOLOGICAL EVENTS AND INDICATOR PLANTS

Entomologists, horticulturists, and agronomists often use phenological indicators to time pest control activities with insect development to improve effectiveness, for example in application of short-residual pesticides or biological control agents. Since plant-feeding insects have evolved closely with their host plants, life events of the host are often used to predict insect life events. To be most useful, a phenological indicator must be a clear cut, definitive event that lasts for only a brief period. Plants that are ubiquitous, easy to recognize, and have showy flowers are best suited for phenological correlations. It would not be practical to provide an exhaustive list here, but the more common indicator plants used in pest prediction and management programs include: the chicory (Chicorium intybus), Canada thistle (Cirsium arvense), wild carrot (Daucus ( carota), Tartarian honeysuckle (Lonicera ( tatarica), wild black cherry ((Prunus serotina), Canada goldenrod (Solidago canadensis), bridalwreath spiraea (Spiraea x Vanhouttei), and the common lilac (Syringa vulgaris). The latter is the cornerstone of many phenology monitoring programs (e.g., emergence of adult onion maggots, Delia antiqua). These phenological events can then be compared with a calendar to approximate the time of year when critical life events occur. In one study, emergence data for several apple insects observed over a period of 10 years was grouped more closely together and formed a normal distribution curve with higher peaks and shorter bases when expressed relative to indicator phenology compared with the same data plotted by calendar dates. The use of phenological events to convey insect life history data is particularly useful when making comparisons between different locations. Phenological events are also a better basis for data analysis when evaluating several years of data from a single location. However, phenological indicators are not without problems. A plant’s response to its environment may vary across geographic regions. Flowering events may be affected by the seed origin of the plant as plants from different geographical origins adapt differently to photoperiod and

408

Phenology: An Integrative Environmental Science

temperature. Also, early blooming species are susceptible to late frost damage in early spring. Furthermore, differences in the microclimate in which the indicator plant is grown can affect the accuracy of their correlation with insect development. Snow cover and soil moisture are two such conditions that will affect insect survival and subsequent emergence from dormancy in the spring (Mussey and Potter 1997).

4.

DEGREE-DAYS

Because in some years or locations insect and plant development are not well synchronized, and the inexactness of phenology indicators, degree-day calculations are often needed. The direct comparison of the impact of weather conditions on insect development is more precise and better adapted to prediction. In some cases, insect development can be monitored through both the use of models and indicator plants. Degree-days (also known as day degrees, heat units, and thermal units) use both the upper and lower developmental threshold temperatures along with temperature records to calculate the amount of heat that accumulates while temperature is within the range between the developmental thresholds in a given 24 hour period. To determine when an insect should reach a particular developmental stage, a cumulative total of degree-days is calculated. For example gypsy moth (Lymantria ( dispar) egg hatch correlates with 100-200 degree-days using a lower threshold temperature of 50°F. Pesticide applications are most effective when targeted toward the early instar (stage) larvae, which appear at 200-350 DD50. To use degreeday accumulations for pest prediction, thermal constants must be established. This is the number of days required for a given event, such as instar larval development, pupation, or adult emergence, to occur.

4.1

Limitations of Degree-Days

There are limitations to the use of degree-day models in predicting insect growth. One such limitation is knowing when to begin counting in the absence of a biofix (phenological or other such event). If there are no observable events such as the appearance of a specific insect in a monitoring trap or through visual observation, when should degree-day accumulation begin? In addition, an insect’s microclimate will also impart an effect on development. For example, insects that spend part of their life in the soil will be affected by soil temperature rather than air temperature during this life stage. Also, soil moisture may affect the rate of development for soil-borne

Chapter 6.2: Insects

409

insects. The first adults of the apple maggot (Rhagoletis pomonella) typically appear when Canada thistle (Cirsium arvense) is in bloom, or when 900 degree-days have accumulated using a base temperature of 50°F. However, if soil moisture is limiting, fly emergence is delayed despite the accumulation of the required number of degree-days or Canada thistle bloom. Finally, indicator phenology and degree-days prediction can best be applied to insects that have distinct life stages and generations. The phenology of aphids, which characteristically have overlapping generations, is not easily predicted by simple degree-day methods. Similarly, migratory insects such as the monarch butterfly ((Danaus plexipppus) cannot have their development predicted accurately by degree-day summation in a given location because of their exposure to a wide range of environmental conditions during their migratory journey. Degree-day methods may be difficult to use when the insect pest has multiple biotypes. For example, the European corn borer (Ostrinia nubilalis) has two distinct generations in most areas but are univoltine in some regions.

4.2

Calculating Degree-Days

There are numerous methods and equations for calculating degree-days, and three of the more common are outlined below. It should be noted that the degree-day calculation method used to predict development should always be the same that was originally used to estimate the degree-day requirements. 4.2.1

Standard method

The Standard method of degree-day calculation is based on the difference between the arithmetic mean temperature and the insect’s lower threshold. It is also referred to as the simple, rectangle, or mean-minus-base method. This method is used primarily for its practicality and not for its accuracy. The number of degree-days accumulated during a given 24-hour period is calculated by subtracting the lower threshold from the average daily temperature. DD accumulation = [(max. + min.) ÷2] – lower threshold

(1)

Most agronomic applications use only one of two lower thresholds when calculating degree-days using the standard method. Cool season crops such as alfalfa and peas and their associated insects use a lower threshold of 40°F (4.4°C) while warm season crops use 50°F (10°C) as the lower threshold.

410

Phenology: An Integrative Environmental Science

4.2.2

Modified degree-days

This modification of the standard method includes both an upper and lower threshold temperature. Modified degree-days are often called the “weather bureau” or “86/50” method as the upper threshold used in the calculation is 86°F (30°C) and the lower threshold 50°F (10°C). Other than the inclusion of an upper threshold, this calculation method is similar to the standard method. The following rules are used to calculate modified degreedays: If daily maximum and minimum temperatures are above the lower threshold but not the upper threshold, or if daily maximum temperature exceeds the lower threshold but not the upper threshold and the daily minimum temperature is less than the lower threshold, the standard method (equation 1) is used. If daily maximum temperature exceeds the lower threshold but not the upper threshold, and the daily minimum temperature is less than the lower threshold: DD accumulation = [(max. + lower threshold) ÷2] – lower threshold (2) If daily maximum temperature exceeds the upper threshold and the daily minimum temperature exceeds the lower threshold: DD accumulation = [(upper threshold + min.) ÷2] – lower threshold (3) Finally, if the daily maximum temperature is less than the lower threshold, no degree-days are accumulated. The advantage of using the modified degree-day method over the standard method is that it recognizes that there is a reduction of development in both plants and insects under hot temperatures and that no appreciable development occurs once the upper threshold has been exceeded. 4.2.3

Sine wave

Degree-day accumulations used in predictive models are commonly estimated by measuring the area beneath a sine curve interpolated between daily minimum and maximum temperature readings. The sine wave method of degree-day calculation is the most accurate of the three methods described here particularly when the temperatures approach the upper or lower thresholds. When temperatures fall within the linear range, all three methods provide identical estimates of degree-day accumulation.

Chapter 6.2: Insects

411

The sine wave method is most useful to predict the development of early season insects in temperate climates. The major disadvantage is that calculation is difficult without the use of a computer, although tables have been developed (Wedberg et. al. 1977).

5.

ECONOMICALLY-IMPORTANT PESTS OF AGRONOMIC AND HORTICULTURAL CROPS IN THE UNITED STATES.

Phenological indicator plants and degree-day models provide pest managers with a practical method of predicting when vulnerable life stages of insect pests are present and when to time pest management actions. This approach is most useful for pests with discrete generations that are present for a brief period of time during the growing season. Suppression of insects with multiple generations early in the season can prevent populations from building rapidly and reaching levels that would cause economic damage. Similarly, insects that are provided protection by their host plant such as the squash vine borer ((Melittia cucurbitae) require controls to be applied before the insect enters the plant. Phenological calendars, such as the one developed by Orton and Green (1989) for woody landscape plants allow nurserymen and arborists to properly time the application of short-residual insecticides to the most vulnerable life stage of the insect. A similar calendar exists on the University of California Statewide Integrated Pest Management Project website for fruit, vegetable, and field crop insects: (http://www.ipm.ucdavis.edu/phenology). For phenology to be used effectively, pest managers must understand the pest complexes that occur on the species of plants they manage. Each insect’s life cycle, including the most vulnerable stages, must be understood as well as when the vulnerable stages occur within the season. With this information, pest managers can consult references such as those stated above for degree-day and phenological indicators that correspond with the vulnerable life stages. These references indicate when degree-day calculation should begin, the model that best approximates insect development, and the lower and upper threshold temperatures. Furthermore, knowledge of the available pesticides and biological control agents available to control a specific insect must be fully understood. Before using any published phenological data or degree-day calculations, it is important to validate those data for a particular area. Formally testing the model to determine the degree-day accumulations for specific life stages of insect in a location and comparing these data to the predicted degree-days

412

Phenology: An Integrative Environmental Science

can accomplish this. Formal validation is most often done by University specialists and involves sampling the insect and correlating data with weather data obtained from a reliable weather station for a minimum of three years. Formal validation is often not necessary. For pest management purposes, an informal method of validation—pest sampling during the time when the model predicts that a certain r life stage will be present—will suffice. Over time, this information can be fine-tuned as more data are accumulated. To illustrate how an insect’s development is strongly influenced by the life events of its host plant and how phenological indicator plants can be useful in predicting emergence and timing pest management activities, the Eastern tent caterpillar ((Malacosoma americanum) will be discussed. The range of this insect includes most of the eastern half of the United States. Common host plants include plants in the genus Malus (apple and crabapple), and Prunus species (such as cherry and wild plum), although any deciduous forest species can be attacked. Moths of the Eastern tent caterpillar lay eggs in groups of 150-350 on the twigs of suitable host plants in midsummer. The eggs overwinter and hatch over a three-week period in spring, concurrent with budbreak of the host plant. Mussey and Potter (1997) observed that several indicator plants unrelated to the insect’s host plants may be used to predict egg hatch. In their phenological sequence of events, egg hatch of the Eastern tent caterpillar occurred between the 95% bloom phase of the cornelian cherry dogwood (Cornus mas) and 50% bloom of forsythia (Forsythia x intermedia). This was consistent over the three years of their study. Orton and Green (1989) suggest that pesticide applications be made at the time when the saucer magnolia (Magnolia x soulangiana) is in pink bud which is also reported in the sequence described by Mussey and Potter (1997). As previously discussed, degree-day calculation is difficult in the absence of a biofix. One must keep in mind that specific insects have their own specific biofix. For example, degree-day accumulation for the Colorado potato beetle ((Leptinotarsa decemlineata) begins when the first eggs are observed. The modified degree-day formula is used in calculation with an upper threshold of 86°F (30°C) and lower threshold of 52°F (11°C). From this reference point, the first instar larvae appear once 185 degree-days have accumulated AFTER the date of egg-laying. Subsequently, the second instar larvae appear at 240 DD52, third instar at 300 DD52, fourth instar at 400 DD52, and pupation at 675 DD52. If a pest manager mistakenly begins degree-day accumulation on a set calendar date such as January 1 or when the crop is planted, their calculations will be incorrect. Another serious insect pest in the corn belt of the Midwestern United States is the European corn borer (Ostrinia nubilalis). This nocturnal

Chapter 6.2: Insects

413

lepidopteran causes economic damage to sweet corn, grain and seed corn; snap and lima beans; and peppers. The adult moths lay eggs on the foliage and after egg hatch foliar feeding by early instar larvae begins. Once the larvae have reached the third instar, they begin to bore into the stalks or fruit of the host plant on which the eggs were laid. The modified degree-day model is used with the same upper and lower thresholds as the Colorado potato beetle. Unlike the Colorado potato beetle, degree-day accumulation does not begin when egg masses are observed in the field but rather at the beginning of the calendar year. The geographical location in which the crop is grown will determine exactly when appreciable degree-days have accumulated, with the southernmost locations beginning accumulation during the months of January through March while northern locations may not begin accumulating degree-days until April. Using this method, the first spring moths appear at 375 DD50, first eggs laid at 450 DD50 and peak moth flight at 631 DD50. Phenological indicator plants have been identified for these events as well. Bridalwreath spiraea (Spiraea x Vanhouttei) is in full bloom when the first spring moths appear. The pagoda dogwood (Cornus alternifolia) is in late bloom stage when eggs are laid and the black locust ((Robinia pseudoacacia) is in full bloom when the spring moth generation peaks. The best treatment period to control the first generation of the European corn borer is at the larval stage before they enter the plant at between 800-1000 DD50. The San Jose scale (Quadraspidiotus perniciosus) was introduced into the United States in the late 1800s from Asia and has become a serious pest that can attack over 700 species of deciduous and fruit trees including apple, aspen, cotoneaster, cottonwood, maple, peach, pear, poplar, walnut, and willow, to name a few. It is ubiquitous in temperate regions of North America. The scales infest the fruit and bark of the host plants and can kill the tree if populations are high enough. Management of the San Jose scale is targeted toward the crawlers, the most susceptible life stage. The scale overwinters as a partially-grown nymph protected by the hard scaly covering that makes it difficult to kill. Winged males emerge in the spring and mate. Females give birth to live, mobile crawlers that are most vulnerable to pesticides during the first 24 hours. To predict when the vulnerable crawler stage is be present, the sine method of degree-day calculation is used with a lower threshold of 51°F (10.5°C) and an upper threshold of 90°F (32.2°C). Degree-day accumulation begins once the first crawlers are observed and the optimum treatment time is between 200-300 DD51 thereafter. As there may be several generations of San Jose scale per year depending on the location, it is important to follow-up control practices on successive generations. The generation time from adult to adult is 1050 DD51.

414

Phenology: An Integrative Environmental Science

Development of soil-dwelling insects, such as the cabbage maggot ((Delia radicum), onion maggot ((Delia antiqua), and seedcorn maggot (Delia ( platura), is more difficult to predict using indicator phenology and degreedays. These maggots are the economically-important larval stages of flies. The insects overwinter as pupae in the soil and the adults emerge in late spring to early summer. Damage results when the maggots bore into or feed upon seeds or roots of susceptible host plants. Because all life stages except the adults inhabit the soil, soil temperature and moisture play a key role in their development. Therefore, a model that uses exclusively air temperature may be misleading or inaccurate. The air temperature, rate of evaporation, amount of solar radiation, soil color, and shading vegetation from the crop canopy all influence soil temperature and moisture levels. Two models for calculating the development of the cabbage maggot differ in the use of air or soil temperatures (Eckenrode et. al. 1971,1972). Both use a base threshold of 43°F (6.1°C) but the number of degree-days required for the emergence of the spring adults is 300 DD43 when air temperature is used and 233 DD43 when soil temperatures are used in the calculation. In heat accumulation studies conducted in controlled growth chambers, it took 243 DD for adult emergence. This suggests that using soil temperatures may be more accurate than using air temperatures for predicting adult emergence. There are many other insects for which indicator phenology and degreeday models are used in prediction and pest management. Table 1 lists the more common insect species in the United States for which these methods are useful. Table 6.2-1. Insects monitored with phenology or degree-days.

Common Name Alfalfa weevil American plum borer Apple maggot Armyworm Artichoke plume moth Ash plant bug Asparagus beetle Bagworm Barberry caterpillar Beet armyworm Beet leafhopper Birch leafminer Black cutworm Black vine weevil Blackberry leafhopper

Scientific Name Hypera postica Euzophera semifuneralis Rhagoletis pomonella Pseudaletia unipuncta Platyptilia carduidactyla Tropidostepes amoenus Crioceris asparagi Thyridopteryx ephemeraeformis Coryphista meadii Spodoptera exigua Circulifer tenellus Fenusa pusilla Agrotis ipsilon Otiorhynchus sulcatus Dikrella californica

Chapter 6.2: Insects Common Name Blue alfalfa aphid Boxwood leafminer Boxwood psyllid Bronze birch borer Cabbage aphid Cabbage looper Cabbage maggot Calico scale California red scale Carrot weevil Celery looper Cereal leaf beetle Citricola scale Citrus thrips Citrus red mite Codling moth Colorado potato beetle Corn earworm Corn leaf aphid Cotton aphid Cotton bollworm Cottony maple scale Crucifer flea beetle Cuban laurel thrips Diamondback moth Dogwood borer Eastern spruce gall adelgid Eastern tent caterpillar Egyptian alfalfa weevil Elm leaf beetle English grain aphid European alder leafminer European corn borer European elm scale European pine sawfly European pine shoot moth European red mite Euonymus scale Fall cankerworm Fall webworm Flatheaded appletree borer

415 Scientific Name Acyrthosiphon kondoi Monarthropalpus flavus Cacopsylla buxi Agrilus anxius Brevicoryne brassicae Trichoplusia ni Delia radicum Eulecanium cerasorum Aonidiella aurantii Listronotus oregonensis Anagraphafalcifera Oulema melanopus Coccus pseudomagnoliarum Scirtothrips citri Panonychus citri Cydia pomonella Leptinotarsa decemlineata Helicoverpa zea Rhopalosiphum maidis Aphis gossypii Helicoverpa zea Pulvinaria innumerabilis Phyllotreta cruciferae Gynaikothrips ficorum Plutella xylostella Synanthedon scitula Adelgis abietis Malacosoma americanum Hypera brunnipennis Pyrrhalta luteola Sitobion avenae Fenusa dohrnii Ostrinia nubilalis Gossyparia spuria Neodiprion sertifer Rhyacionia buoliana Panonychus ulmi Unaspis euonymi Alsophila pometaria Hyphantria cunea Chrysobothris femorata

416

Phenology: An Integrative Environmental Science

Common Name Fletcher scale Fruittree leafroller Fuller rose beetle Green peach aphid Greenhouse whitefly Gypsy moth Hawthorn lace bug Hawthorn mealybug Hemlock eriophyid mite Holly leafminer Honeylocust borer Honeylocust plant bug Honeysuckle aphid Hop vine borer Imported cabbageworm Indian meal moth Inkberry leafminer Japanese beetle Juniper scale Juniper tip midge Juniper webworm Lecanium scale Leaf crumpler Lesser peachtree borer Lilac borer Locust borer Locust mite Lygus bug Maple bladder gall mite Meadow spittlebug Mediterranean fruit fly Melon aphid Melon fly Mexican bean beetle Mimosa webworm Nantucket pine tip moth Navel orangeworm Northern corn rootworm Obscure scale Olive scale Omnivorous leafroller

Scientific Name Parthenolecanium fletcheri Archips argyrospilus Asynonychus godmani Myzus persicae Trialeurodes vaporariorum Lymantria dispar Corythucha cydoniae Phenacoccus dearnessi Nalepella tsugifoliae Phytomyza ilicis Agrilus difficilis Diaphnocoris chlorionis Hyadaphus tataricae Hydraecia immanis Pieris rapae Plodia interpunctella Phytomyza glabricola Popillia japonica Carulaspis juniperi Oligotrophus betheli Dichromeris marginella Lecanium corni Acrobasis indigenella Synanthedon pictipes Podosesia syringae Megacyllene robiniae Eotetranychus multidigii Lygus hesperus Vasates quadripedes Philaenus spumarius Ceratitis capitata Aphis gossypii Bactrosera cucurbitae Epilachna varivestis Homadaula anisocentra Rhyacionia frustrana Amyelois transitella Diabrotica barberi Melanapsis obscura Parlatoria oleae Platynota stultana

Chapter 6.2: Insects Common Name Onion maggot Onion thrips Orange tortrix Oriental fruit fly Oriental fruit moth Oystershell scale Pacific spider mite Pea aphid Peach twig borer Peachtree borer Pear psylla Pear rust mite Pine bark aphid Pine needle scale Pink bollworm Plum fruit moth Potato leafhopper Potato tuberworm Roundheaded appletree borer Russian wheat aphid San Jose scale Scurfy scale Seedcorn maggot Serpentine fruit fly Sod webworm Spotted tentiform leafminer Spring cankerworm Spruce budworm Spruce bud scale Spruce eriophyid mite Spruce needleminer Spruce spider mite Squash bug Squash vine borer Strawberry spider mite Sunflower beetle Sunflower moth Sunflower stem weevil Sweet potato whitefly Tobacco budworm Tomato fruitworm

417 Scientific Name Delia antiqua Thrips tabaci Argyrotaenia citrana Bactrocera dorsalis Grapholita molesta Lepidosaphes ulmi Tetranychus pacificus Acyrthosiphon pisum Anarsia lineatella Synanthedon exitiosa Cacopsylla pyricola Epitrimerus pyri Pineus strobi Chionaspis pinifoliae Pectinophora gossypiella Grapholitha funebrana Empoasca fabae Phthorimaea operculella Saperda candida Diuraphis noxia Quadraspidiotus perniciosus Chionaspis furfura Delia platura Anastrepha serpentina Pediasia trisecta Phyllonorycter blancardella Paleacrita vernata Choristoneura fumiferana Physokermes piceae Nalepella halourga Endothenia albolineana Oligonychus ununguis Anasa tristis Melittia cucurbitae Tetranychus turkestani Zygogramma exclamationis Homoeosoma electellum Cylindrocopturus adspersus Bemisia tabaci Heliothis virescens Helicoverpa zea

418

Phenology: An Integrative Environmental Science

Common Name Tomato pinworm Twolined chestnut borer Twospotted spider mite Variegated cutworm Vegetable leafminer Viburnum crown borer Walnut scale Western cherry fruit fly Western grape leafhopper Western pine shoot borer Whitemarked tussock moth Willow aphid Wooly apple aphid Yellownecked caterpillar Yellow poplar weevil Zimmerman pine moth

Scientific Name Keiferia lycopersicella Agrilus bilineatus Tetranychus urticae Peridroma saucia Liriomyza sativae Synanthedon fatifera and S. viburni Quadraspidiotus juglansregiae Rhagoletis indifferens Erythroneura elegantula Eucosoma sonomana Ostyia leucostigma Pterocomma smithae Eriosoma lanigerum Datana ministra Odontopus calceatus Dioryctria zimmermani

The use of indicator phenology and degree-days lend themselves well to predicting the development of insects. Whether this interest is for purely recreational purposes such as observing insect development at a local nature center, or as part of the work of an agronomist, horticulturist, or pest manager, phenology will bring you more in tune with nature.

REFERENCES CITED Arnold, C. Y., The determination and significance of the base temperature in a linear heat unit system, Proc. Amer. Soc. Hort., 74, 430-445, 1959. Eckenrode, C. J., and R. K. Chapman, Effect of various temperatures upon rate of development of the cabbage maggot under artificial conditions, Ann. Entomol. Soc. Amer., 64, 1079-1083, 1971. Eckenrode, C. J., and R. K. Chapman, Seasonal adult cabbage maggot populations in the field in relation to thermal-unit accumulations, Ann. Entomol. Soc. Amer., 65, 151-156, 1972. Lieth, H. (Editor), Phenology and Seasonality Modeling, Springer-Verlag, New York, 444 pp., 1974. Mussey, G. J., and D. A. Potter, Phenological correlations between flowering plants and activity of urban landscape pests in Kentucky, J. Economic Entomology, 90, 1615-1627, 1997. Orton, D. A., and T. L. Green, Coincide: the Orton system of pest management, Plantsmen’s Publications, Illinois, 190 pp., 1989. Wedberg, J. L., E. J. Ruesink, E. J. Armbrust and D. P. Bartell, Alfalfa weevil pest management program, Univ. Ill. Ext. Circ., 1136, 8 pp., 1977. Wolda, H., Insect Seasonality: Why?, Ann. Rev. Ecol. Syst., 19, 1-18, 1988.

Chapter 6.2: Insects

419

Worner, S. P. Performance of phenological models under variable temperature regimes: consequences of the Kaufmann or rate summation effect, Environ. Entomology, 21, 689699, 1992.

Chapter 6.3 BIRDS Tim H. Sparks1, Humphrey Q. P. Crick2, Peter O. Dunn3, and Leonid V. Sokolov4 1

Centre for Ecology and Hydrology, Monks Wood, UK; 2British Trust for Ornithology, Thetford, UK; 3Department of Biological Sciences, University of Wisconsin-Milwaukee, Milwaukee, WI, USA; 4Russian Academy of Sciences, St. Petersburg, Russia

Key words:

1.

Birds, Climate change, Migration, Temperature, Nesting

INTRODUCTION

The Roman mosaic floor at Lullingstone, Kent, UK depicts the four seasons as human characters and having a barn swallow Hirundo rustica on her shoulder uniquely identifies spring. Thus, for almost 2000 years people have associated spring with the arrival of migratory birds. In the UK the (currently) oldest phenological record dates to 1703 and refers to the call of the cuckoo Cuculus canorus. It is hardly surprising then that the timing of bird activity is a productive area of phenological research. Birds are probably the most popular group of plants or animals and are keenly watched. In temperate zones their behavior is very seasonal and therefore ideal for phenological study. Indeed, the volume of data on the timing of bird activity is considered to outweigh that of any other form of phenological data. Unfortunately, large amounts of these data are collected at different sampling intensities and without any large-scale coordination. Interpretation of such data must be undertaken with care and should combine empirical findings with the known ecology of species. It must be remembered that birds are highly mobile and often secretive, and, thus, phenological observation relies on the intensity of recording and on the Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 421-436 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

422

Phenology: An Integrative Environmental Science

density and visibility of the bird species. The most reliable observational measurements derive from those species that are large and obvious, such as the white stork Ciconia ciconia, or occur in large numbers and are associated with human habitation, such as the barn swallow or house martin Delichion urbica. There are many types of data that range from the casual observation of birds in an individual’s garden, to the intensive daily netting of birds at bird observatories, to the national coordinated schemes on the nest timing of birds. We include examples of several of these types of data in this chapter.

2.

CHANGES IN THE TIMING OF MIGRATION

2.1

Spring Arrival of Summer Visitors

The bulk of avian phenological data relates to the passage in spring of birds that overwinter in warmer environments. These include birds that migrate across continents, as well as species that travel shorter distances. The latter group includes species such as skylark Alauda arvensis in central and northern Europe and American robin Turdus migratorius in North America (Inouye et al. 2000). A third group of birds are partially migratory, in that a proportion of their population migrates. This last group includes species such as starling Sturnus vulgaris. Increasingly we learn of normally migratory birds overwintering, such as chiffchaff Phylloscopus collybita in the UK (Geen 2002), white stork in Germany, skylark in Poland and Canada goose Branta canadensis in North America. So, different bird species are migratory in different regions and migrate over varying distances. It is not surprising then that different cultures associate spring with different species. In the UK the reporting of the first cuckoo has long been of high media profile, whereas elsewhere in Europe white stork or skylark may be considered better indicators of spring. In North America, the arrival of robins is traditionally associated with the arrival of spring, although one of the most famous local examples is the arrival of swallows at San Juan Capistrano in southern California. Thus, there may be a different emphasis on which species are recorded returning to their breeding grounds. This is particularly so of observational studies undertaken by individuals. Schemes that operate from bird observatories based on observation or netting are less subject to bias in species choice. In general, short distance migrants return earlier to breeding areas than do long distance migrants, for example in Poland (Tryjanowski et al. 2002), where short distance migrants have shown a greater trend to earlier arrival in recent years than long distance migrants. Short distance migrants have the flexibility of a short passage time to respond quickly to changing

Chapter 6.3: Birds

423

environmental conditions. In Figure 1 changes in mean migration timing at Rybachy, on the Baltic coast of Russia are shown for chaffinch Fringilla coelebs, a short distance migrant, and willow warbler Phylloscopus trochilus, a long distance migrant. During the 44 years from 1959 to 2002, chaffinch arrived an average of 0.34±0.08 days earlier per year (p<0.001), while willow warblers arrived 0.28±0.04 days per year earlier (p<0.001). This is equivalent, respectively, to 15 and 12 days earlier over the recording period. It is worth emphasizing that these are not first migrating birds but rather mean migration dates of birds caught in nets. Thus they derive from a methodology not subject to observational bias and indicate a shift in the migration distribution pattern of the species. The greater variability in arrival dates of earlier species is apparent in this graph (see also Mason 1995) and is a characteristic of migration phenology data sets in general.

140

Day of year

130

120

110

100 1960

1970

1980

1990

2000

Year Figure 6.3-1. The change in mean arrival date of a short distance migrant, chaffinch (solid symbols), and a long distance migrant, willow warbler (open symbol), at Rybachy, Russia 1959-2002. Smoothed (LOWESS) lines have been superimposed. Data source: Rybachy observatory.

In addition to the data on the whole migration distribution available from bird observatories, there are many sources of data supplying first arrival date only. An example from the Essex (UK) Bird Reports (Sparks and Mason 2001) is given in Figure 2. This shows the trend towards earlier arrival over the last two decades of whimbrel Numenius phaeopus and hobby Falco

424

Phenology: An Integrative Environmental Science

subutteo. Both show a significant trend towards earlier arrival; 2.19±0.44 (p<0.001) and 0.67±0.31 (p=0.04) days earlier per year respectively equivalent to 39 and 12 days earlier over the recording period. These first dates need to be treated with caution as they may reflect changes in population size and thus visibility of the species (Sparks 1999; Tryjanowski and Sparks 2001). Indeed the hobby population is increasing in the UK and some of the trend towards earliness may be an artifact of a larger population. However these first dates do also change in highly visible species with static populations, so they can reflect a change in at least one aspect of the migration distribution, even if they do not necessarily tell us about changes to the whole arrival distribution (Sparks et al. 2001). In many regions including Wisconsin, USA (Bradley et al. 1999), Russia (Sokolov et al. 1998) and the UK (Loxton and Sparks 1999) the majority of the bird events recorded have tended to become earlier.

105

120 100 95

Day of yearr

Day of year

90 85 80

110 0

75 70

100

65 60 1980

1990

Year

2000

1980

1990

2000

Year

Figure 6.3-2. The change in first arrival date of whimbrel (left), and hobby (right) from Essex, UK 1981-1998. Smoothed (LOWESS) lines have been superimposed. Data source: Essex Bird Reports.

Not all data show a trend towards earliness, for example barn swallow in Slovakia (Sparks and Braslavskà 2001) and migrants in NE Scotland (Jenkins and Watson 2000). Inevitably the overall picture can be distorted if

Chapter 6.3: Birds

425

only significant results get published. A proper meta analysis to identify the true situation is long overdue.

2.2

Autumn Departure of Summer Visitors

Determining the date of autumn departures of breeding birds is more problematic for observational recorders. Where records exist they usually concern the last observation date of a species, which is usually more difficult to pin down than the first bird in spring. More reliable data derive from netting schemes where birds are trapped on southwards passage. Figure 3 shows the last recorded capture date for two species, tree pipit Anthus trivialis and chiffchaff, from the observatory at Rybachy, Russia. These demonstrate a trend towards earlier departure in the former species of 0.50±0.13 days per year (p<0.001) and a trend towards later departure of the latter of 0.21±0.06 days per year (p=0.002). These trends equate to 21 days earlier and 9 days later, respectively, over the 42-year record. Changes in the autumn departure of birds at Rybachy give a mixed picture and a similar situation was found within the Russian Arctic Circle with some species

305 5

295

Day of year

285

275

265

255

245 1960

1970

1980

1990

2000

Year Figure 6.3-3. The change in the last departure of tree pipit (open symbol) and chiffchaff (closed symbol) at Rybachy, Russia from 1960-2001. Smoothed (LOWESS) lines have been superimposed. Data source: Rybachy observatory.

426

Phenology: An Integrative Environmental Science

departing significantly earlier and some significantly later in a 60+ year record of bird migration (Gilyazov and Sparks 2002). In Rybachy most birds do not show a significant trend one way or the other. However there is plenty of evidence that in years of earlier breeding autumn migration gets underway earlier (e.g., Ellegren 1990). In an examination of autumn departures it must be remembered that species are likely to have different optimal strategies and must balance any benefits of earlier departure from the breeding grounds with costs of missed additional breeding opportunities (in multibrooded species) and mistimed arrival at migratory staging areas or wintering grounds.

2.3

Winter Visitors

Species from colder environments can overwinter in milder environments to the south. In the northern hemisphere the numbers of such species tend to be fewer than summer visitors and the recording effort is lower. In Britain, emphasis has been placed on two species of thrush, redwing Turdus iliacus and fieldfare Turdus pilaris, that move in from Scandinavia. Fewer records exist for overwintering waders. An unpublished study by Sparks and Mason suggests that greater changes have taken place in the timing of short distance migrants than for long distance migrants, but this conclusion is heavily influenced by substantial timing changes in a small number of raptors. Scarcity of data does not make it easy to make generalizations. An example of changed arrival in Bewick’s swan Cygnus bewickiii is shown in Figure 4. While departure of summer visitors from wintering grounds is considered to be driven by photoperiod (e.g., Kok et al. 1991) the cues for migration of winter visitors have received less attention.

3.

INFLUENCE OF CLIMATE ON MIGRATION TIMING

There is substantial evidence that arrival dates are related to local temperatures. This may be an indirect effect through the supply of insect prey, or it may reflect southerly, and hence warm, tail winds aiding migration. There can be little doubt that en-route temperatures must play their part in migration. Indeed Huin and Sparks (1998, 2000) detected an effect on four UK bound migrants, but this aspect has received only modest attention. Undoubtedly this situation will change as temperature data become easier to obtain and as migration routes become better known. Figure 5 displays the relationship between the arrival date of barn swallow as

Chapter 6.3: Birds

427

recorded at four British observatories and spring temperature. The relationship suggests earlier arrival by 2.8±0.5 days for every 1ºC increase (p<0.001).

340

Day of year

320

300

280

260 1970

1980

1990

2000 000

Year Figure 6.3-4. The change in first autumn arrival of Bewick’s swan in Essex, UK from 19661998 (with gaps). Data source: Essex Bird Reports.

A response of migration timing to temperature has been detected in a wide range of studies including those in Poland (Tryjanowski et al. 2002), Slovakia (Sparks and Braslavskà 2001), Russia (Sokolov and Payevsky 1998; Sokolov et al. 1999; Gilyazov and Sparks 2002), and France (Sueur and Triplet 2001).

4.

CHANGES IN THE NEST TIMING OF BIRDS

Some species of bird form very obvious nests. This is particularly so of species that are large, or build nests associated with human habitation. The white stork builds nests on the tops of trees, chimneys or other vertical structures. The rook Corvus frugilegus and grey heron Ardea cineria both build nests in treetop colonies that are very obvious. In the UK the rook is such an obvious species that Robert Marsham included it in his 18th century

428

Phenology: An Integrative Environmental Science

Indications of Spring g (Sparks and Carey 1995) by noting the dates of nest building and the dates on which young could be detected. Hole nesting

Day of year

105 05

95

85

3

4

5

6

7

8

February-April mean temperature Figure 6.3-5. The relationship between mean first arrival date of barn swallow from four British bird observatories 1959-1999 and mean February-April Central England temperature (ºC). Data source: Dungeness, Portland, Bardsey, and Calf of Man observatories.

species can be encouraged to nest in particular locations by providing nest boxes and this makes recording their breeding activity much simpler (e.g., pied flycatcher Ficedula hypoleuca). For other species it is necessary to search for nests and it is usual practice to follow the eggs during laying and hatching and to record the subsequent growth of nestlings. From these dates it is possible to back-calculate first egg date. In the UK the British Trust for Ornithology (BTO) conduct a nest record scheme that receives approximately 30,000 records each year. From these data, statistics on the timing of nesting can be calculated. Crick et al. (1997) showed that a large number of species from a wide range of guilds (types) are breeding progressively earlier. An example from the BTO scheme of reed warbler Acrocephalus scirpaceus is given in Figure 6. Over the last 20 years there has been a marked trend towards earlier breeding in this species. Regression against time suggests that the commencement of egg laying has advanced by 0.77±0.21 days per year (p=0.002), that is by 15 days in 20 years.

Chapter 6.3: Birds

429

This advance in nest timing has been reported from a wide range of species and locations including red-necked starlings Sturnia phillipenisii in Japan (Koike and Higuchi 2002), pied flycatcher in Russia (Sokolov 2000),

Day of year

150

140

130

1980

1985

1990

1995

2000 0

Year Figure 6.3-6. Change in the commencement of nesting (5thh percentile) of reed warbler in the UK 1981-2000. Smoothed (LOWESS) line has been superimposed. Data source: British Trust for Ornithology.

tits Parus spp. in Germany (Winkel and Hudde 1997), goldeneye Bucephala clangula in Germany (Ludwichowski 1997) and tree swallow Tachycineta bicolor in North America (Dunn and Winkler 1999). At Rybachy in Russia there has been a strong advance in the breeding time of many species. Figure 7 displays the situation for pied flycatcher that displays a trend towards earlier breeding of 0.30±0.10 days per year (p<0.01) equivalent to an 8-day advance over the period of recording.

5.

CLIMATE INFLUENCE ON TIMING OF BREEDING

Bird species time their reproduction to maximize the number of offspring produced within a season. Some species are capable of producing multiple broods per year and will need to balance the success of the first brood with

430

Phenology: An Integrative Environmental Science

the options for subsequent brood(s) (Crick et al. 1993). Other species that are single brooded aim to hatch their young at a time of optimal food supply, usually in the form of invertebrates. It is known that the development times of invertebrates under elevated temperatures can halve, for example from 56 to 23 days in the case of the development of winter mothh Operophtera

Day of year

175

170

165

160

1975

1980

1985

1990

1995

2000

Year Figure 6.3-7. Change in the mean nest timing of pied flycatcher at Rybachy, Russia 19752002. Smoothed (LOWESS) line has been superimposed. Data source: Rybachy observatory.

brumata (Buse et al. 1999) but birds cannot accelerate the incubation period of individual eggs. Thus, warmer spring temperatures can lead to earlier peaks in insect abundance and, consequently, a mismatch between the timing of peak food abundance and maximum food requirements of developing young (Visser et al. 1998). The response of bird species to a warming climate is likely to be influenced by the phenology of their prey items. However, experiments under ad-libitum food conditions that artificially warm or cool nests by 2-3°C suggest that temperature can influence the start of egg laying independent of food availability (Meijer et al. 1999), so it is possible that climate change can also have a direct physiological effect on phenology, but this needs more study. In many species the timing of egg laying is correlated with temperature (Ludwichowski 1997; McCleery and Perrins 1998; Brown et al. 1999; Crick and Sparks 1999; Dunn and Winkler 1999; Slater 1999; Przybylo et al. 2000;

Chapter 6.3: Birds

431

Sokolov 2000; Both and Visser 2001). For example, in reed warblers, birds nest an average of 5.0±0.7 days earlier for every degree warmer in spring (p<0.001, Figure 8). Visually, Figure 8 suggests that a straight-line relationship may not be valid over the entire range of temperatures; excluding the colder years would increase the response to temperature detailed above. On the other hand, some bird species do not appear to be

Day of year

150

140

130

3

4

5

6

7

8

February-April mean temperature Figure 6.3-8. The relationship between the commencement of nesting (5thh percentile) in reed warbler in the UK 1981-2000 and mean February-April Central England temperature (ºC). Data source: British Trust for Ornithology.

changing their breeding phenology in response to changes in temperature. This could be due to differences between species in how temperatures at different times of the breeding season affects food abundance. For example, during 1973 to 1995 there has been an increase in late spring temperatures in the Netherlands, but there has been neither change in early spring temperature, nor any change in the laying date of great tits, and it is the temperature during early spring that is most closely correlated with laying date (Visser et al. 1998). Thus, birds may only respond to temperatures during particular time periods that affect the abundance of their breeding resources. Interestingly, great tits and pied flycatchers on the same study area in the Netherlands respond differently to climate change (Both and Visser 2001). The tits, which are residents, breed about two weeks earlier

432

Phenology: An Integrative Environmental Science

than the flycatchers and show no significant change in laying date, because their critical temperature period has not changed over time. In contrast, the flycatchers, which migrate from wintering grounds in Africa, have advanced their laying date, because their critical temperature period, which is later in spring, has become warmer over time (Both and Visser 2001). Despite the advancement in laying date of flycatchers, selection over the past 20 years has become even stronger for an earlier laying date (i.e., more young are recruited from earlier nests). It appears that flycatchers have not responded to this selection pressure because their timing of arrival is determined mostly by photoperiod. Thus, the cue for initiating migration (photoperiod) has become maladaptive in some respects because it no longer provides a cue for the best time to arrive on the breeding grounds (Both and Visser 2001).

6.

CRITICAL APPRAISAL OF THE STATE OF KNOWLEDGE AND WHAT MORE NEEDS TO BE DONE

There can be little doubt that bird phenology has changed in recent decades. This has a most marked effect on the timing of spring migration of summer visitors and on the timing off breeding. The effects on autumn departure of summer visitors and the migration timing of winter visitors are less consistent, possibly reflecting the greater difficulty in collecting these data and possibly reflecting that individual species have their own strategies. We have little doubt that it is easier to collect phenological information on plants than on birds, which are highly mobile and sometimes secretive. In a short chapter like this it has not been possible to discuss other aspects of bird phenology, such as the timing of molt in autumn and the detection of bird song in resident species in spring. There has been a general trend towards earlier arrival of birds in spring and this has been reported from a wide range of geographic locations. Arrival times in general appear to be related to temperatures with earlier arrival in warmer years. However a comparison of arrival date with temperature at destination has tended to ignore the potential influence of temperatures en-route. The response to destination temperatures is typically 2 days/°C, not enough to explain the sometimes dramatic shifts in the timing of migration. We do need to take into account migration route temperatures, land use change and, possibly, adaptation strategies to fully explain the magnitude of recent changes and to be able to predict the consequences of global warming.

Chapter 6.3: Birds

433

Similarly we have detected some very marked trends towards earlier breeding activity, often associated with warmer weather. Breeding earlier often results in larger clutch sizes or greater survival of offspring (Lack 1968), so we might expect warmer springs to benefit bird populations. Indeed, a 2.5°C increase is predicted to increase the carrying capacity of a Norwegian dipper Cinclus cinclus population by 58% (Saether et al. 2000). Similarly, black-throated blue warblers Dendroica caerulescens in New Hampshire had higher annual fecundity in La Niña years (Sillett et al. 2000), when May temperatures tend to be warmer in New England. In some species, the particular pattern of climate change is likely to be important. For example, in capercaillie Tetrao urogallus the weather most favorable for successful breeding consisted of a quickly rising temperature in April when eggs were laid and a warm, dry period in early June when chicks hatched (Moss et al. 2001). It is also important to keep in mind that even though warmer temperatures may increase reproductive success, this increase could easily be eliminated by increases in mortality on the wintering grounds or at later stages in life. Thus, there may be a number of constraints that limit the responses of bird populations to warmer weather (Winkler et al. 2002). Considering how much we know about bird biology in general, it is somewhat surprising that we know so little about such basic issues as the effects of phenology on reproductive success and population demography. More focused studies of the effects of phenology are needed. Phenological data vary enormously in their quality. Some data are better than no data at all, but we do need to be cautious in interpreting change. The most reliable data come from schemes that follow a standardized protocol, either through observation or capture, but even these methods can be criticized. Capture relies on landfall of migrants, which increases in poorer weather conditions. Observation requires good visibility and may not take into account nocturnal migrants unless radar is being used. Population size may influence first events, but are not expected to influence the mean/median or standard deviation of the data. Despite some reservations over data interpretation there can be little doubt that a response to a changing climate is happening in the phenology of bird populations. This evidence is particularly compelling for the timing of breeding and migration arrival where intensive scientific research has taken place. Rich sources of data provide ample opportunity to examine bird phenology and the complex relationships both between and within species and will be a productive area of research in the coming years.

434

Phenology: An Integrative Environmental Science

REFERENCES CITED Brown, J. L., S.-H. Li, and N. Bhagabati, Long-term trend toward earlier breeding in an American bird: A response to global warming?, Proc. Natl. Acad. Sci., 97, 5565-5569, 1999. Both, C., and M. E. Visser, Adjustment to climate change is constrained by arrival date in a long-distance migrant bird, Nature, 411, 296-298, 2001. Bradley, N. L., A. C. Leopold, J. Ross, and W. Huffaker, Phenological changes reflect climate change in Wisconsin, Proc. Natl. Acad. Sci. (USA), 96, 9701-9704, 1999. Buse, A., S. J. Dury, R. J. W. Woodburn, C. M. Perrins, and J. E. G. Good, Effects of elevated temperature on multi-species interactions: the case of Pedunculate Oak, Winter Moth and Tits, Func. Ecol., 13 (Suppl. 1), pp. 74-82, 1999. Crick, H. Q. P., C. Dudley, D. E. Glue, and D. L. Thomson, UK birds are laying eggs earlier, Nature 388, 526, 1997. Crick, H. Q. P., D. W. Gibbons, and R. D. Magrath, Seasonal variation in clutch size in British Birds, J. Animal Ecology, 62, 263-273, 1993. Crick, H. Q. P., and T. H. Sparks, Climate change related to egg-laying trends, Nature 399, 423-424, 1999. Dunn, P.O., and D. W. Winkler, Climate change has affected the breeding date of tree swallows throughout North America, Proc. Roy. Soc. London B, 266, 2487-2490, 1999. Ellegren, H., Timing of autumn migration in Bluethroats Luscinia s. svecica depends on timing of breeding, Ornis Fennica, 67, 13-17, 1990. Geen, G., Common Chiffchaff (Chiffchaff) Phylloscopus collybita, in The Migration Atlas: movements of the birds of Britain and Ireland, edited by C. V. Wernham, M. P. Toms, J. H. Marchant, J. A. Clark, G. M. Siriwardena, and S. R. Baillie, pp. 568-570, T. and A. D. Poyser, London, 2002. Gilyazov, A., and T. Sparks, Change in the timing of migration of common birds at the Lapland nature reserve (Kola Peninsula, Russia) during 1931-1999, Avian Ecology and Behavior, 8, 35-47, 2002. Huin, N., and T. H. Sparks, Arrival and progression of the Swallow Hirundo rustica through Britain, Bird Study, 45, 361-370, 1998. Huin, N., and T. H. Sparks, Spring arrival patterns of the Cuckoo Cuculus canorus, Nightingale Luscinia megarhynchos and Spotted Flycatcher Musciapa striata in Britain, Bird Study, 47, 22-31, 2000. Inouye, D. W., B. Barr, K. B. Armitahe, and B. D. Inouye, Climate change is affecting altitudinal migrants and hibernating species, Proc. Natl. Acad. Sci. (USA), 97, 1630-1633, 2000. Jenkins, D., and A. Watson, Dates of first arrival and song of birds during 1974-99 in midDeeside, Scotland., Bird Study, 47, pp. 249-251, 2000. Koike, S., and H. Higuchi, Long-term trends in the egg-laying date and clutch size of Redcheeked Stalings Sturnia philippensis, Ibis, 144, 150-152, 2002. Kok, O. B., C. A. Van Ee, and D. G. Nel, Daylength determines departure date of the spotted flycatcher Muscicapa striata from its winter quarters, Ardea, 79, 63-66, 1991. Lack, D., Ecological adaptations for breeding in birds, Methuen, London, 409 pp., 1968. Loxton, R. G., and T. H. Sparks, Arrival of spring migrants at Portland, Skokholm, Bardsey and Calf of Man, Bardsey Observatory Report, 42, 105-143, 1999. Ludwichowski, I., Long-term changes of wing-length, body mass and breeding parameters in first-time breeding females of goldeneyes ((Bucephala clangula clangula) in Northern Germany, Vogelwarte, 39, 103-116, 1997.

Chapter 6.3: Birds

435

Mason, C. F., Long-term trends in the arrival dates of spring migrants, Bird Study, 42, 182189, 1995. McCleery, R. H., and C. M. Perrins, Temperature and egg-laying trends, Nature, 391, 30-31, 1998. Meijer, T., U. Nienaber, U. Langer, and F. Trillmich, Temperature and timing of egg-laying of European starlings, Condor, 101, 124-132, 1999. Moss, R., J. Oswald, and D. Baines, Climate change and breeding success: decline of the capercaillie in Scotland, J. Animal Ecology, 70, pp. 47-61, 2001. Przybylo, R., B. C. Sheldon, and J. Merila, Climatic effects on Breeding and morphology: evidence for phenotypic plasticity, J. Animal Ecology, 69, 395-403, 2000. Saether, B. E., J. Tufto, S. Engen, K. Jerstad, O. W. Røstad, and J. E. Skåtan, Population dynamical consequences of climate change for a small temperate songbird, Science, 287, 854-856, 2000. Sillet, T. S., R. T. Holmes, and T. W. Sherry, Impacts of a global climate cycle on population dynamics of a migratory songbird, Science, 288, 2040-2042, 2000. Slater, F. M., First-egg date fluctuations for the pied flycatcher Ficedula hypoleuca in the woodland of mid-Wales in the twentieth century, Ibis, 141, 489-506, 1999. Sokolov, L. V., Spring ambient temperature as an important factor controlling timing of arrival, breeding, post-fledging dispersal and breeding success of Pied Flycatchers Ficedula hypoleuca in Eastern Baltic, Avian Ecology and Behavior, 5, 79-104, 2000. Sokolov, L. V., M. Yu. Markovets, A. P. Shapoval, and Yu. G. Morozov, Long-term trends in the timing of spring migration of passerines on the Courish spit of the Baltic sea, Avian Ecology and Behavior, 1, 1-21, 1998. Sokolov, L. V., M. Yu. Markovets, and Yu. G. Morozov, Long-term dynamics of the mean date of autumn migration in passerines on the Courish spit of the Baltic sea, Avian Ecology and Behavior, 2, 1-18, 1999. Sokolov, L. V., and V. A. Payevsky, Spring temperatures influence year-to-year variations in the breeding phenology of passerines on the Courish Spit, eastern Baltic, Avian Ecology and Behavior, 1, 22-36, 1998. Sparks, T. H., Phenology and the changing pattern of bird migration in Britain, Int. J. Biometeorol., 42, 134-138, 1999. Sparks, T. H., and O. Braslavská, The effects of temperature, altitude and latitude on the arrival and departure dates of the swallow Hirundo rustica in the Slovak Republic, Int. J. Biometeorol., 45, 212-216, 2001. Sparks, T. H., and P. D. Carey, The responses of species to climate over two centuries: An analysis of the Marsham phenological record, 1736-1947, J. Ecology, 83, 321-329, 1995. Sparks, T. H., and C. F. Mason, Dates of arrivals and departures of spring migrants taken from Essex Bird Reports 1950-1998, Essex Bird Report 1999, 154-164, 2001. Sparks, T. H., D. R. Roberts, and H. Q. P. Crick, What is the value of first arrival dates of spring migrants in phenology?, Avian Ecology and Behavior, 7, 75-85, 2001. Sueur, F., and P. Triplet, Réchauffement climatique: les passereaux arrivent-ils plus tôt au printemps?, Avifaune picardie, 1, 111-120, 2001. Tryjanowski, P., and T. H. Sparks, Is the detection of the first arrival date of migrating birds influenced by population size? A case study of the Red-backed Shrike Lanius collurio, Int. J. Biometeorol., 45, 217-219, 2001. Tryjanowski, P., S. KuĨniak, and T. Sparks, Earlier arrival of some farmland migrants in western Poland, Ibis, 144, 62-68, 2002.

436

Phenology: An Integrative Environmental Science

Visser, M. E., A. J. van Noordwijk, J. M. Tinbergen, and C. M. Lessells, Warmer springs lead to mistimed reproduction in great tits (Parus ( major), Proc. Roy. Soc. London B., 265, 1867-1870, 1998. Winkel, W., and H. Hudde, Long-term trends in reproductive traits of tits (Parus major, P.caeruleus) and Pied Flycatchers Ficedula hypoleuca, J. Avian Biol., 28, 187-190, 1997. Winkler, D.W., P. O. Dunn, and C. E. McCulloch, Predicting the effects of climate change on avian life-history traits, Proc. Nat. Acad. Sci. (USA), 99, 13595-13599, 2002.

Chapter 6.4 TIMING OF REPRODUCTION IN LARGE MAMMALS Climatic and density-dependent influences Eric Post Department of Biology, The Pennsylvania State University, University Park, PA, USA

Key words:

1.

Caribou, Mammals, North Atlantic Oscillation, Red deer, Seasonality

INTRODUCTION

This discussion of the reproductive phenology of mammals will focus on large herbivores, sometimes referred to as ruminants or ungulates. Large herbivores have been a major focus of theoretical and empirical studies of the influences of biotic and abiotic factors on reproductive phenology since the pioneering study by Estes (1976) on breeding synchrony in wildebeest (Connochaetes taurinus). The examples used in this chapter derive mainly from multi-annual studies of caribou (Rangifer ( tarandus) and moose ((Alces alces) in arctic and sub-arctic environments, and red deer (Cervus elaphus) on the north-temperate Isle of Rhum, Scotland. The former two species illustrate the influences of abiotic (i.e., climatic) factors on reproductive phenology, while the latter illustrates with striking clarity the influence of population density on timing of calving. The distinction will be made in this chapter between the timing and synchrony of births, as well as between long-term (i.e., evolutionary) and proximal influences on timing and synchrony of births. The distinction between timing and synchrony is important because different forces may act upon the two; moreover, biological and environmental factors may act upon one to influence the other. The distinction between long-term and proximal Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 437-449 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

438

Phenology: An Integrative Environmental Science

influences is aimed at clarifying the difference between general patterns of reproductive phenology and interannual variation about those patterns.

2.

LONG-TERM INFLUENCES AND EVOLUTIONARY CONSIDERATIONS

In many species of large herbivores, there is an obvious season of births. This is especially evident in seasonal environments, though the length of the birth season is highly variable among species. The season of births is referred to as the timing of parturition, while the length of this season (the number of days over which births occur within a population) is referred to as synchrony of parturition. Many competing, though not necessarily exclusive, hypotheses have been forwarded to explain why births occur seasonally and with a high degree of synchrony in populations of several species of large herbivores, including wildebeest, bighorn sheep (Ovis canadensis, Festa-Bianchet 1988), white-tailed deer (Odocoileus virginianus, McGinnis and Downing 1977), Dall’s sheep (Ovis dalli, Rachlow and Bowyer 1991), and mule deer (Odocoileus hemionus, Bowyer 1991). Chief among these are the “predation hypothesis” and the “seasonality hypothesis.” The predation hypothesis, forwarded originally to explain synchronous breeding in colonially nesting birds (Darling 1938), predicts that synchronous reproduction should result from selection by predation against early- or late-born offspring. Wildebeest in Ngorongoro Crater, Tanzania, for example, experience predation by spotted hyenas (Crocuta crocuta) during the calving season. Wildebeest young born into large groups at the peak of the highly synchronized calving season are much less likely to be killed by hyenas than those born at the beginning or end of the calving season (Estes 1976; Estes and Estes 1979). In part, the advantage of being born during the peak of calving derives from predator swamping—the reduction of risk to the individual of being killed because of the greater numbers of potential prey per predator during the peak—while additional benefit may be derived from being born into groups of vigilant mothers congregating during the peak of calving (Rutberg 1987; Bøving and Post 1997). Rutberg (1987) suggested, however, that synchronous parturition is unlikely to have evolved from asynchronous parturition solely in response to selection acting through predation on newborns. Instead, predation on earlyand late-born neonates might act to increase synchrony of parturition in populations that already display seasonal birth peaks timed to coincide with seasonal peaks in resource availability (Rutberg 1987).

Chapter 6.4: Timing of Reproduction in Large Mammals

439

The seasonality hypothesis predicts that selection acting through intraannual variation in weather and resource availability, results in optimization of the timing of parturition by individuals to coincide with the seasonal peak of food availability (Sadleir 1969; Sekulic 1978). In understanding how seasonality might result in the evolution of synchronous parturition, it is important to consider, as Rutberg (1987) pointed out, that natural selection probably does not act on synchrony of parturition because synchrony is a population-level characteristic. Rather, synchrony may result from selection on the timing of parturition by individuals, and the consequent variation in their own reproductive success and that of their offspring as influenced by seasonal variation in resource availability (Rutberg 1987; Ims 1990a). Hence, seasonality may be an important selective force in the evolution of the annual timing of reproduction, with consequences for synchrony, while predation may reinforce or strengthen synchrony without exerting a clearly discernable influence on the annual timing of parturition (Ims 1990a, b). Indeed, evidence from multi-annual studies of depredated populations of Dall’s sheep (Rachlow and Bowyer 1991) and moose (Bowyer et al. 1998) in Alaska, USA, and of mule deer (Bowyer 1991) in California, USA, indicate that, despite predation on newborns, timing and synchrony of

Figure 6.4-1. Timing of calving by caribou and moose in Denali National Park, Alaska, USA, in relation to the winter index of the Arctic Oscillation (AO). Shown are annual first (ż) and median (Ƒ) dates of calving by caribou (adapted from Adams and Dale 1998), and median dates of calving by moose (Ŷ) (adapted from Bowyer et al. 1998), overlain by the AO index (Thompson and Wallace 1998; heavy black line) of the previous winter.

440

Phenology: An Integrative Environmental Science

parturition in the focal populations appear to relate to long-term climatic patterns and their influence on offspring survival. While seasonal reproduction by large mammals may reflect evolutionary strategies by individuals to time parturition to coincide with seasonal peaks in resource availability, multi-annual studies typically reveal variation among years in the onset and median date of parturition. We may, for instance, refer to the birth season of large herbivores in sub-arctic environments as occurring generally in mid to late May, but it would be absurd to state that Alaskan moose give birth on May 18 each year. The median date of parturition by female moose in Denali National Park, Alaska, for example, varied by 7 days over a 5-year period (Bowyer et al. 1998), while that of female caribou in the same area varied by 8 days over 9 years (Adams and Dale 1998, Figure 1).

3.

PROXIMAL INFLUENCES ON TIMING OF PARTURITION

3.1

Direct Climatic Influences

The distinction between the general timing of the season of births and the exact dates over which parturition occurs in a period of years is one of longterm (i.e., evolutionary) vs. proximal influences on the timing of parturition. In highly seasonal or extreme environments, interannual variation in the timing of parturition may relate directly to variation in weather as it influences condition of reproducing females before conception or during pregnancy. Ostensibly, weather may reduce the physical condition of females to the point where oestrus and, consequently, parturition are delayed. Indeed, the interannual variation in onset and median dates of parturition by Alaskan caribou depicted in Figure 1 was highly correlated with late winter snowfall during the previous year (Adams and Dale 1998). Alternatively, or additionally, weather may influence the condition of reproductive females during pregnancy, with consequences for the timing of parturition. Keech et al. (2000), for example, noted that pregnant female moose with good body condition (thick rump fat) gave birth earlier than pregnant females with poor body condition, and concluded that this difference reflected environmental influences during pregnancy. Bowyer et al. (1998) found no relation between the interannual variability in timing of parturition by moose depicted in Figure 1 and local winter weather, but suggested this might have reflected low sample size (5 years of data). Nonetheless, the high degree of correlation between the

Chapter 6.4: Timing of Reproduction in Large Mammals

441

median date of calving by moose and the median date of calving by caribou (rr = 0.85, P = 0.07), and the median date of calving by moose and the first date of calving by caribou (rr = 0.95, P = 0.01) in Denali park (Figure 1), suggests a common climatic influence on interannual variation in the timing of parturition by both species. The winter index of the Arctic Oscillation (Thompson and Wallace 1998) correlates well with both annual data on cumulative snow depth (Mech et al. 1998) for the period in Denali National Park (rr = 0.62, P = 0.04) and annual dates of first calving (r = 0.78, P = 0.01) and median calving (rr = 0.78, P = 0.01) by caribou (Figure 1). Although this correlation was not significant for moose (rr = 0.42, P = 0.49), the directions of these relationships suggest later parturition following snowy winters in both populations.

3.2

Indirect Climatic Influences: Plant Phenology and Productivity

In addition to acting directly on condition of reproductive females both prior to and during pregnancy, weather can also influence the timing of parturition through its influence on the timing of plant growth. Desert bighorn sheep in California, USA, for example, display a peak in lambing just after winter rains when forage plant productivity peaks (Rubin et al. 2000). Similarly, timing of parturition in southern mule deer in California relates to temperatures and precipitation in the last third of gestation, when forage productivity is highest (Bowyer 1991). Caribou inhabiting sub-arctic and arctic environments, where the season of plant growth is short and forage plants display a distinct peak in nutrient quality (Klein 1990), also appear to time parturition to coincide with patterns of plant phenology. On the Southern Alaska Peninsula, USA, the progression of the caribou calving season tracks closely the progression of plant phenology on calving ranges of two herds (Figure 2a, Post and Klein 1999). In this example, the proportion of calves observed on each calving range increased rapidly with the number of forage species emerging, despite the fact that the dates of onset of calving differed between these populations, whose nutritional regimes differed (Post and Klein 1999). Similarly, a comparison among arctic and sub-arctic herds of caribou and wild reindeer, some of which were depredated and others not, revealed a close association between the onset of calving and the onset of the season of plant growth (Figure 2b, Skogland 1989).

442

3.3

Phenology: An Integrative Environmental Science

Density-Dependent Influences

a

proportion emergent

start t t off calving l i

proportion calves

The long-term, individual-based study of red deer on the Isle of Rhum, Scotland, has assumed an importance in the fields of population, behavioral, and evolutionary ecology unparalleled by any other study of large mammals. Between 1971 and 1983, the female segment of the population increased from 57 to 166 (a three-fold increase), while the median date of calving progressed from June 1st to June 11thh over the same period (Figure 3, Clutton-Brock et al. 1985). As density increased on Rhum, competition for resources increased, reducing the condition of females prior to breeding; consequently, the average date of conception progressed from September 25th to October 9th over the period, resulting in later birth dates (CluttonBrock and Albon 1989).

b

start of growing season

Figure 6.4-2. Timing of calving by caribou in relation to plant phenology. In panel (a), the proportion of calves in congregations of female caribou of the Southern Alaska Peninsula Herd, USA, increases with the proportion of forage plants emergent on separate calving ranges, Caribou River(ż) and Black Hill (Ɣ) (adapted from Post and Klein 1999). In panel (b), a cross-population comparison reveals that the start of the calving season is highly correlated with the start off the plant-growing season on ranges in North America and Norway (adapted from Skogland 1989).

As Clutton-Brock and Albon (1989) pointed out, the influence of population density on timing of calving on Rhum was so strong that, after accounting for density, the median date of parturition varied by only four days between 1971 and 1983. In the latter part of this period, from 1977 onward, there is an apparent negative association between the timing of calving and the wintertime state of the North Atlantic Oscillation (NAO) (Figure 3). Because the NAO correlates positively with late winter temperatures in the north Atlantic region, this would seem to suggest, at

Chapter 6.4: Timing of Reproduction in Large Mammals

443

higher densities, a tendency toward earlier calving in years with warmer winters and, presumably, earlier springs. To investigate these relationships, I tested for combined influences of density and climate (the NAO) in a non-linear generalized additive model (GAM) (Hastie and Tibshirani 1999). The results of the GAM support clearly the reported influence of density on timing of calving, and indicate that this relationship tends to level off at the highest densities (Figure 4a). This seems to suggest that although increasing density results in progressively later dates of calving, there is a constraint on this relationship

Figure 6.4-3. Median dates of calving (ż) by female red deer in the population on the Isle of Rhum, Scotland, in relation to female density (Ɣ) and the winter NAO index (heavy line) of the previous year. Data on calving dates and female density are from Clutton-Brock et al. (1985).

imposed, perhaps, by physiological limits on the timing of oestrus and/or the length of the period of gestation. Moreover, the GAM indicates a tendency toward earlier calving one year after warm winters (Figure 4b). Although it is not possible to determine with these data whether this reflects a direct or indirect influence of weather, the timing of plant growth occurs earlier following warm (positive NAO) winters in many regions (Post and Stenseth 1999; Post 2003). The clear influence of density on timing of calving by red deer on Rhum suggests density dependence may influence timing of reproduction in other

444

Phenology: An Integrative Environmental Science

populations as well. Hence, I used the same approach, together with data on annual estimates of population density for caribou in the Denali Herd

calf date

a

female density b

NAO index

Figure 6.4-4. Results of a generalized additive model (GAM) of timing of calving by red deer on the Isle of Rhum in relation to density (a) and the winter NAO index of the previous year (b). Following the general model of the timing of life history events in relation to climate, density, and resources developed in Post et al. (2001), the GAM also included an autoregressive spline function specifying the influence of the timing of calving in the previous year (not shown). Y-axis values are standardized to the null deviance of the fitted model (solid lines), and dashed lines are 95% confidence bands.

calf date

Chapter 6.4: Timing of Reproduction in Large Mammals

445

Density

AO index Figure 6.4-5. Influences of population density (a) and the winter AO index of the previous year on timing of calving by caribou in the Denali Herd, as for red deer in Figure 4.

(Mech et al. 1998) to test for combined influences of density dependence and the Arctic Oscillation on timing of calving in that population. This analysis reveals an effect of density on timing of calving by caribou in the Denali herd (Figure 5a) that mirrors the effect of density on timing of calving by red deer on Rhum. Moreover, the influence of the Arctic

446

Phenology: An Integrative Environmental Science

Oscillation on timing of calving is still evident after incorporating the influence of density (Figure 5b).

4.

CONSEQUENCES OF VARIATION IN TIMING OF PARTURITION

The timing of birth in relation to resources and maternal condition influences the early development of individuals in many species (Lindström 1999). What population-level consequences might arise, though, from environmentally- and biologically-induced variation in timing of parturition? For red deer on Rhum, survival of calves through their first year is strongly dependent upon their birth weight and birth date (Clutton-Brock et al. 1987). Calf survival through their first summer increases with birth weight, but declines with later birth dates (Clutton-Brock et al. 1987). Calves born late are also less likely to survive their first winter (CluttonBrock et al. 1987). The consequences of variation in birth dates for population dynamics are evident in the fact that calf winter mortality is the key factor regulating population size of red deer on Rhum (Clutton-Brock et al. 1985). Climatic conditions during pregnancy can also result in phenotypic variation among cohorts of individuals that relates to conditions in their year of birth (Post et al. 1997). In Soay sheep (O. aries) in the Outer Hebrides of Scotland, the North Atlantic Oscillation influences birth dates and weights of lambs, with consequences for long-term variation among cohorts (Forchhammer et al. 2002) that contribute to population dynamics (Coulson et al. 2001). In the Tanana Flats of interior Alaska, USA, birth date influences juvenile survival of moose: later-born individuals display a shorter time to death in their first year (Keech et al. 2000). How birth date contributes to the dynamics of this population is not, however, clear. In Denali National Park, timing of parturition has no apparent influence on survivorship of young moose (Bowyer et al. 1998). Timing of parturition influences juvenile survival of caribou in the Denali herd indirectly through vulnerability to predation. Caribou calves born during the peak of parturition (five to eight days after the beginning of the calving season) are significantly less likely to be killed by predators than those born before or after the peak (Adams et al. 1995). The contribution of variation in birth dates of caribou calves to their population dynamics is evident in the observations that 98% of calf mortality in the Denali Herd is due to predation, and that predation during the calving season is the main limiting factor for the Denali Herd (Adams et al. 1995).

Chapter 6.4: Timing of Reproduction in Large Mammals

5.

447

CONCLUSION: IMPLICATIONS OF CLIMATE CHANGE

Given that local weather and large-scale climate exert direct and indirect influences on the timing of parturition in large herbivores, we might expect directional climate change to elicit shifts in the timing of parturition by many species. Such climate-mediated shifts in timing of reproduction may, however, be counteracted by changes in density (Forchhammer et al. 1998). As well, seasonal reproduction in northern ungulates may be cued to photoperiod (Leader-Williams 1988), which may also constrain long-term shifts in parturition, as is apparently the case with egg laying in some birds (Both and Visser 2001). Moreover, although the GAMs used to analyze timing of parturition among red deer on Rhum and caribou in Denali Park suggest linear relations between climate and birth dates (Figures 4 and 5), we might expect constraints on gestation length or timing of onset of oestrus to limit the extent to which timing of parturition can advance in response to climatic warming. Finally, considering the influence that timing of parturition exerts on juvenile survival, and the link between juvenile survival and population dynamics in many species of large herbivores (Gaillard et al. 1998), it is reasonable to speculate that climate change may influence population dynamics of many species through timing of reproduction. If so, climatic influences on timing of parturition may contribute to spatial synchrony among populations, as in the high degree of correlation between timing of parturition by moose and that by caribou in Denali Park (Figure 1). The dynamics of caribou and muskoxen (Ovibos moschatus) on opposite coasts of Greenland, for instance, are highly correlated (Post and Forchhammer 2002), and the dynamics of populations of both species in Greenland relate to the North Atlantic Oscillation at time lags that correspond to time to first reproduction in both species (Forchhammer et al. 2002). Further research into the implications of climate change for timing of reproduction in large herbivores, with consequences for population dynamics and synchrony, is certainly warranted.

REFERENCES CITED Adams, L. G., F. J. Singer, and B. W. Dale, Caribou calf mortality in Denali National Park, Alaska, J. Wildl. Manage., 59, 584-594, 1995. Adams, L. G., and B. W. Dale, Timing and synchrony of parturition in Alaskan caribou, J. Mamm., 79, 287-294, 1998.

448

Phenology: An Integrative Environmental Science

Both, C., and M. E. Visser, Adjustment to climate change is constrained by arrival date in a long-distance migrant bird, Nature, 411, 296-298, 2001. Bøving, P. S., and E. Post, Vigilance and foraging behaviour of female caribou in relation to predation risk, Rangifer, 17, 55-63, 1997. Bowyer, R. T., Timing of parturition and lactation in southern mule deer, J. Mamm., 72, 138145, 1991. Bowyer, R. T., V. Van Ballenberghe, and J. G. Kie, Timing and synchrony of parturition in Alaskan moose: long-term versus proximal effects of climate, J. Mamm., 79, 1332-1344, 1998. Clutton-Brock, T. H., and S. D. Albon, Red deer in the highlands, Oxford University Press, Oxford, 260 pp., 1989. Clutton-Brock, T. H., M. Major, and F. E. Guinness, Population regulation in male and female red deer, J. Anim. Ecol., 54, 831-846, 1985. Clutton-Brock, T. H., M. Major, S. D. Albon, and F. E. Guinness, Early development and population dynamics in red deer, I. Demographic consequences of density-dependent changes in birth weight and date, J. Anim. Ecol., 56, 53-67, 1987. Coulson, T., E. A. Catchpole, S. D. Albon, B. J. T. Morgan, J. M. Pemberton, T. H. CluttonBrock, M. J. Crawley, and B. T. Grenfell, Age, sex, density, winter weather, and population crashes in Soay sheep, Science, 292, 1528-1531, 2001. Darling, F. F., Bird flocks and breeding cycle, Cambridge University Press, Cambridge, 124 pp., 1938. Estes, R. D., The significance of breeding synchrony in the wildebeest, E. African Wildl. J., 14, 135-152, 1976. Estes, R. D., and R. K. Estes, The birth and survival of wildebeest calves, Z. Tierpsych., 50, 45-95, 1979. Festa-Bianchet, M., Birthdate and survival in bighorn lambs (Ovis canadensis), J. Zool., 214, 653-661, 1988. Forchhammer, M. C., E. Post, and N. C. Stenseth, Breeding phenology and climate, Nature, 391, 29-30, 1998. Forchhammer, M. C., T. H. Clutton-Brock, J. Lindström, and S. D. Albon, Climate and population density induce long-term cohort variation in a northern ungulate, J. Anim. Ecol., 70, 721-729, 2001. Forchhammer, M. C., E. Post, N. C. Stenseth, and D. Boertmann, Long-term responses in arctic ungulate dynamics to variation in climate and trophic processes, Population Ecology, 44, 113-120, 2002. Gaillard, J. M., M. Festa-Bianchet, and N. G. Yoccoz, Population dynamics of large herbivores: variable recruitment with constant adult survival, Trends Ecol. Evol., 13, 5863, 1998. Hastie, T. J., and R. J. Tibshirani, Generalized additive models, Chapman and Hall, New York, 335 pp., 1999. Ims, R. A., On the adaptive value of reproductive synchrony as a predator-swamping strategy, Am. Nat., 136, 485-498, 1990a. Ims, R. A., The ecology and evolution of reproductive synchrony, Trends Ecol. Evol., 5, 135140, 1990b. Keech, M. A., R. T. Bowyer, J. M. Ver Hoef, R. D. Boertje, B. W. Dale, and T. R. Stephenson, Life-history consequences of maternal condition in Alaskan moose, J. Wildl. Manage., 64, 450-462, 2000. Klein, D. R., Variation in quality of caribou and reindeer forage plants associated with season, plant part, and phenology, Special Issue, Rangifer, 3, 123-130, 1990.

Chapter 6.4: Timing of Reproduction in Large Mammals

449

Leader-Williams, N., Reindeer on South Georgia, Cambridge University Press, Cambridge, 319 pp., 1988. Lindström, J., Early development and fitness in birds and mammals, Trends Ecol. Evol., 14, 343-348, 1999. McGinnis, B. S., and R. L. Downing, Factors affecting the peak of white-tailed deer fawning in Virginia, J. Wildl. Manage., 41, 715-719, 1977. Mech, L. D., L. G. Adams, T. J. Meier, J. W. Burch, and B. W. Dale, The wolves of Denali, University of Minnesota Press, Minneapolis, 227 pp., 1998. Post, E., Large-scale climate synchronizes the timing of flowering by multiple species, Ecology, 84, 277-281, 2003. Post, E., and M. C. Forchhammer, Synchronization of animal population dynamics by largescale climate, Nature, 420, 168-171, 2002. Post, E., and D. R. Klein, Caribou calf production and seasonal range quality during a population decline, J. Wildl. Manage., 63, 335-345, 1999. Post, E., and N. C. Stenseth, Climatic variability, plant phenology, and northern ungulates, Ecology, 80, 1322-1330, 1999. Post, E., M. C. Forchhammer, N. C. Stenseth, and T. V. Callaghan, The timing of life history events in a changing climate, Proc. R. Soc. London B, 268, 15-23, 2001. Post, E., N. C. Stenseth, R. Langvatn, and J.-M. Fromentin, Global climate change and phenotypic variation among red deer cohorts, Proc. R. Soc. London B, 264, 1317-1324, 1997. Rachlow, J. L., and R. T. Bowyer, Interannual variation in timing and synchrony of parturition in Dall’s sheep, J. Mamm., 72, 487-492, 1991. Rubin, E. S., W. M. Boyce, and V. C. Bleich, Reproductive strategies of desert bighorn sheep, J. Mamm., 81, 769-786, 2000. Rutberg, A. T., Adaptive hypotheses of birth synchrony in Ruminants: an interspecific test, Am. Nat., 130, 692-710, 1987. Sadleir, R. M. F. S., The ecology of reproduction in wild and domesticated mammals, Methuen, London, 321 pp., 1969. Sekulic, R., Seasonality of reproduction in the sable antelope, E. African Wildl. J., 16, 177182, 1978. Skogland, T., Comparative social organization of wild reindeer in relation to food, mates, and predator avoidance, Paul Parey Publishers, Berlin, 74 pp., 1989. Thompson, D. W., and J. M. Wallace, The Arctic Oscillation signature in the wintertime geopotential height and temperature fields, Geophys. Res. Lett., 25, 1297-1300, 1998.

PART 7

APPLICATIONS OF PHENOLOGY

Chapter 7.1 VEGETATION PHENOLOGY IN GLOBAL CHANGE STUDIES Michael A. White1, Nathaniel Brunsell2, and Mark D. Schwartz3 1

Department of Aquatic, Watershed, and Earth Resources, Utah State University, Logan, UT, USA; 2Department of Civil Engineering, Duke University, Research Triangle, NC, USA; 3 Department of Geography, University of Wisconsin-Milwaukee, Milwaukee, WI, USA

Key words:

1.

Global Change, Growing season length, Land-atmosphere interactions, Wavelets, AVHRR

INTRODUCTION

Global change, encompassing natural and anthropogenic changes to the Earth system at sub-annual to geologic time scales, has strong interactions with vegetation phenology. In this chapter we will refer to global change as alterations to the Earth system that are certainly or probably influenced by human activity, primarily since the industrial revolution. This form of global change includes irrefutable anthropogenic alterations to terrestrial land cover and alterations to the global climate that are probably anthropogenically influenced. Within this context we discuss three aspects of vegetation phenology: the influence of vegetation phenology on general circulation models (GCMs); a wavelet analysis of phenological patterns and associated evidence of likely phenological responses to direct human-induced land cover alteration; and third, serious challenges regarding the use of phenological data and concepts in global change research.

Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 453-466 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

454

2.

Phenology: An Integrative Environmental Science

GLOBAL CHANGE

Global change is most controversially associated with climate change. The Intergovernmental Panel on Climate Change (IPCC WG I 2001) found that on average, the global climate warmed by approximately 0.5°C in the 20th century. This overall trend, though, masked localized trends that were often much higher and/or confined to increases in seasonal or nighttime temperatures. Despite large uncertainty relating to cloud physics, aerosols, and the influence of land cover changes, most GCMs predict a continued trend of increasing temperatures throughout the 21st century, even if greenhouse gas emissions are reduced or stabilized (IPCC WG I 2001). Due to the lack of an experimental Earth, the precise influence of anthropogenic activity on climate cannot be predicted with absolute accuracy, but the likelihood that humans are at least partially responsible for current climatic changes is extremely high. Regardless of origin, climate change has occurred within the 20th century, the pace of the change is increasing, and it is very likely that the climate will continue to warm. Other forms of global change are not subject to debate. In particular, humans have extensively modified the land use and land cover (LULC) of the terrestrial biosphere. Through conversion of natural habitats to agricultural or urban use, fire suppression, military activity, logging, mining, power generation, accidental or intentional transport of invasive species, etc., the anthropogenic fingerprint is visible and pervasive across the Earth’s surface. Human use of global landscapes for agricultural use is already extensive (DeFries et al. 2002). Of the remainder, much has experienced prior anthropogenic LULC change, the impacts of which can be evident long after the initial disturbance (Foster et al. 1998). While humans have modified other elements of the Earth system such as nutrient cycles (Vitousek 1994) and soils (Imhoff et al. 1997), climate and LULC change are most relevant for phenological issues. Three major areas of phenological responses by both plants and animals to variation in climatic signals are discussed in other chapters: evidence from field datasets showing changes in phenology, phenological responses to urbanization, and implications of phenological variation for the terrestrial carbon cycle. See Menzel (2002) for a general discussion of phenology and global change. In the following sections, we present a focused analysis in two areas in which vegetation phenology may respond to or influence global change. First, we illustrate the importance of vegetation phenology in GCM applications. Second, we present a wavelet analysis showing the influence of edaphic conditions, climate, and LULC on the spatial patterns of phenological variability and discuss likely global change impacts on existing

Chapter 7.1: Vegetation Phenology in Global Change Studies

455

patterns. Finally, by showing four valid definitions of growing season length, we highlight the potential for confusion in phenological research.

3.

PHENOLOGY AND CLIMATE MODELS

In the early 1990s much GCM research focused on coupling global land, ocean, and atmospheric models to simulate the complete Earth system. Around that time, many models of the land surface (Dorman and Sellers 1989; Bonan et al. 2002a) began to incorporate the concept of phenology into coarse resolution modeling. However, for most of these efforts, the global distribution of a plant functional type (PFT) would have identical seasonality profiles. In other words, annual leaf area index (LAI) profiles of grasslands in the central United States and Mongolia would be identical. In the late 20th and early 21st century multiple decades of consistently processed satellite-derived LAI became available, allowing for a consideration of the impacts of actual phenology in a variety of climate modeling activities. Two complementary articles, both using a land surface model (LSM, Bonan 1996) coupled with the Community Climate Model (Kiehl et al. 1998), illustrate the growing acceptance of phenological concepts in modeling activities. In the first, Buermann et al. (2001) used 0.25° LAI data to construct 1981-1991 average, minimum, and maximum LAI profiles for 13 LSM PFTs. Similar to the prescribed LSM condition, each PFT had the same LAI profile; the profile in this case was based on satellite-measured LAI rather than estimated values. The authors compared climatic simulations with the realistic range of satellite-measured LAI and the prescribed LAI from the LSM and found that in general, satellite LAIs were lower than the prescribed LSM LAIs. Consequently, a known cold bias in CCM was partially corrected through an increase in sensible heat fluxes, more than compensating for decreased radiation absorption caused by higher albedo. Second, using 1992-1993 global 1km LAI data, Bonan et al. (2002a) simulated the impact of spatially variable LAI profiles in comparison to the prescribed LSM seasonality. They found that use of the realistic LAI strongly affected model simulations of ground temperature and evaporation, albedo, and soil moisture. Radiation partitioning into sensible and latent fluxes was also impacted. The studies are complementary: one shows the impact of interannual variability (Buermann et al. 2001) while the other shows the impact of spatial variability (Bonan et al. 2002a). In combination, the studies show that failure to incorporate phenological variability in time and space is likely to cause serious errors in climate simulations. Regional simulations demonstrate similar impacts of vegetation phenology on climate. Using the RAMS model, Lu and Shuttleworth (2002) showed that depending

456

Phenology: An Integrative Environmental Science

on the phenological parameterization used, regional climate in the US Great Plains and Rocky Mountain region could be colder and wetter or hotter and drier than in a “default” simulation and argued for the realistic depiction of the timing and magnitude of vegetation activity in climate studies. Vegetation phenology is also an important factor for an emergent research field: coupling climate models with dynamic global vegetation models (DGVMs). In a DGVM, plant biogeography is coupled with biogeochemical concepts to simultaneously simulate climate and ecosystem carbon and water cycles in a world of shifting PFTs. Phenology is already included in most DGVMs run off-line (uncoupled) from GCMs (Cramer et al. 2001); late 20th and early 21st century work focused on executing DGVMs coupled to GCMs (Levis et al. 1999; Foley et al. 2000; Bonan et al. 2002b). Within coupled DGVM-GCMs, phenology affects the following processes: (1) the absolute value and seasonal timing of land surface albedo; (2) the partitioning of net radiation to sensible and latent fluxes; (3) the timing of photosynthesis; (4) the timing and amount of litterfall. Such studies have demonstrated numerous climate-vegetation interactions including an apparent stability of boreal forest area under climatic perturbation (Levis et al. 1999) and shifts in western Africa from desert to grassland (Claussen and Gayler 1997). In cases where coupled LSM-GCMs or DGVM-GCMs are run for the near past, remotely sensed estimates of LAI may be used. For future or far past climates, a prognostic phenology scheme must be used. Most DGVMs and coarse resolution LSMs predict phenological variation based on broad rules relating climatic variation to seasonality (Bonan et al. 2002b) or a productivity index (see models in Cramer et al. 2001). Most, though, use basic and untested climatic triggers to regulate the timing of growth. Compared to other model components, relatively little effort has been devoted to prognostic phenology modeling. Given the importance of phenology within coupled climate models, expansion, implementation, and testing of earlier prognostic schemes (White et al. 1997: Botta et al. 2000) is crucial. Further, an understanding of the scale at which phenological variation occurs and the likely global change impacts on such patterns is important for modeling efforts; we explore these topics in the next section.

4.

WAVELET ANALYSIS

We used the biweekly composite normalized difference vegetation index (NDVI) values from the Earth Resources Observation Systems Data Center (EDC) 1990-1999 conterminous United States (CONUS) 1km Advanced Very High Resolution Radiometer (AVHRR) dataset to calculate the start of

Chapter 7.1: Vegetation Phenology in Global Change Studies

457

season (SOS) and end of season (EOS). NDVI, while potentially confounded by viewing geometry and atmospheric contamination, is related to the fraction of photosynthetically active radiation absorbed by vegetation canopies (Myneni and Williams 1994). NDVI is commonly used in remotely-sensed plant productivity models (Coops 1999; Potter et al. 2001), and has a long history in phenological research (Lloyd 1990; Reed et al. 1994; Moulin et al. 1997; Asner et al. 2000). Our method uses a pixelspecific seasonal midpoint NDVI threshold (SMN) based on the following steps: (1) the annual minimum and maximum NDVI is extracted from cloudand snow-screened data; (2) the midpoint between the minimum and maximum is calculated; (3) the procedure is repeated for each year except 1994 (satellite failure); and (4) SMN is assigned as the average of the nine values. The SMN is thus sensitive to each pixel’s seasonal NDVI cycle. A discussion of the methodology and comparisons with observation is available elsewhere (White et al. 1997; White et al. 1999; Schwartz et al. 2002; White et al. 2002). Using a daily NDVI time series created from cloud-screened and spline-fit biweekly NDVI, we defined SOS as the date at which the SMN was exceeded and EOS as the date at which the NDVI fell below the SMN. We calculated annual growing season length (GSL) as EOS-SOS and then calculated the 1990-1999 average GSL from the nine individual years. We implemented wavelet analysis as described below. Wavelet transformations (see Csillag and Kabos (2002) for detailed description) are conceptually similar to Fourier transformations. In the Fourier approach, sine and cosine functions are used to recreate a signal. The wavelet approach relies on the dilation and translation of a wavelet function wherein the function is non-zero for a finite distance. Wavelet analysis allows determination of the extent to which the signal or image matches the wavelet function at particular locations and resolutions. In a classic example, Torrence and Compo (1998) showed how wavelet analysis could be used to determine the structure and timing of the El Niño Southern Oscillation. For image processing, a dyadic decomposition is used in which the spatial resolution (pixel size) is increased by powers of two. Images should be square with dimensions equal to powers of two. Wavelet transforms allow the determination of the relative contribution of each resolution to the observed data. This is done with wavelet variance:

S(a) =

1 2 ΨD (a,x) ∑ N x

(1)

where S(a) is the wavelet variance at resolution a (2, 4, 8, 16, etc.), ΨD(a,x) are the detailed wavelet coefficients resulting from a two dimensional

458

Phenology: An Integrative Environmental Science

wavelet transform of the original signal f(x) as a function of location x (in this case a vector in two dimensional space), and N is the number of data points in the original signal (Percival 1995). A higher value of S(a) indicates that at the given resolution, the data show a high correspondence to the wavelet function. The plot of wavelet variance vs. the resolution is termed the scalogram (Kumar and Foufoula-Georgiou 1993), and is used to determine the dominant length scale (DLS) observed in the data. The DLS is defined as the scale at which the maximum wavelet variance occurs in the scalogram (Kumar and Foufoula-Georgiou 1997) and represents the length scale of the highest correspondence between the wavelet function and the input dataset. For the purposes of this work, the primary focus is on the determination of the DLS, not on the orientation of that scaling behavior. Therefore, only the detail components in the horizontal and vertical directions of the wavelet transform are used. This results in the inability to analyze anisotropy within the signal (especially within the cospectra where the DLS of the different signals may be occurring in orthogonal directions). However, this was not deemed to be important since the primary purpose of this research is to examine the variability in length scales across regions, and not the directional aspects of the length scales. The wavelet cospectra can be developed in an analogous way. In this case, the wavelet decomposition of two signals f(x) and g(x) of length N are used to compute the cross-scalogram:

C(a) =

1 ΨgD (a,x) ∑ ΨfD (a,x)Ψ N x

(2)

where the overbar designates the complex conjugate. The wavelet cospectra are conceptually similar to regression-based covariance. We used GSL for calculation of wavelet variance and the following three datasets relating to climate, edaphic conditions, and land cover for cospectral analysis. For climate, we used d surface meteorological records from a daily CONUS 1km 1980-1997 (1990-1997 used) dataset (P.E. Thornton, unpublished manuscript developed from methods in Thornton et al. 1997, and described at www.daymet.org). From the daily records of maximum and minimum temperature and radiation we calculated annual potential evapotranspiration (PET) using a Priestley-Taylor approach as described in Kimball et al. (1997). Next, we calculated the annual water deficit (WD) as precipitation - PET and finally calculated the average for the period of record. For edaphic conditions, we derived soil percent sand, silt, and clay

Chapter 7.1: Vegetation Phenology in Global Change Studies

459

and soil depth from a 1km CONUS soils database (Miller and White 1998) and then calculated soil water content at field capacity (SWC) using equations in Cosby et al. (1984). Lastly, we obtained 1992 National Land Cover Data (http://edc.usgs.gov/products/landcover/nlcd.html) reclassified to 1km resolution. From 21 possible classes, we used agricultural (AG, combination of five classes), forest (FO, combination of three classes), shrubland (SH), grassland/herbaceous (GR), barren (BA, combination of three classes), urban (combination of three classes), and water classes. Thus for each 1km pixel and for each of the above land cover classes, a percent coverage was available. All data were either produced at or georectified to the standard EDC Lambert Azimuthal Equal Area (LAZEA) projection. As shown in Plate 1, we extracted three regions from the CONUS GSL image roughly corresponding to west, center, and east regions of the United States where the dimension of each region was 1024×1024 pixels. For each region, we employed Mallat’s two-dimensional dyadic decomposition algorithm (Mallat 1999) with Daubechie’s second order wavelet as implemented in the Interactive Dataa Language (Research Systems Inc., Boulder, CO) to calculate the GSL wavelet scalogram and wavelet cospectra for WD, SWC, and land cover (Plate 1). To assess cospectra at a given length scale, examine the columns in the image portion of each panel. In the west at 2km this shows highest cospectra for SWC. To assess cospectra for one variable at all length scales, examine the rows. In the west, SWC shows highest cospectra at 32km. Scales are different on color bars; colors among panels are thus not directly comparable and should be used only to assess cospectral patterns for the individual regions. All regions showed a DLS at 2km, indicating that variation in GSL was highest at the shortest length scale. The center and east showed a general decline in wavelet variance at longer length scales with a drop at 512km for the east. The west, though, showed higher wavelet variance at 32 to 128km. Overall wavelet variance was dramatically higher in the west. The images in the right side of Plate 1 show that patterns of wavelet cospectra varied strongly by region (urban and water classes showed very low cospectra and are not presented). Consider that for the DLS (2km), the maximum cospectra were SWC in the west, AG & GR in the center, and AG & FO in the east. Dominant cospectra by region were SH, SWC, and FO in the west; AG and GR in the center; and AG & FO in the east. In spite of the patterns of variability, several consistent trends emerged. First, the importance of SWC tended to decline as overall precipitation increased from west to east. Second, for each region, the WD cospectra increased towards longer length scales, indicating that climatic influences on GSL variation were manifest primarily at regional, not local, scales. Third, in the center and east regions, the human-created AG class was a dominant control in the

460

Phenology: An Integrative Environmental Science

cospectra. With the exception of SWC dominance of the 2km and 512km scales in the west, maximum cospectra were associated with land cover and not edaphic (SWC) or climatic (WD) controls. While we present results for WD here, we also investigated cospectra of separate climatic controls including precipitation, maximum and minimum temperature, radiation, and vapor pressure (and the annual amplitude of each). For individual regions and length scales a particular climatic variable could be more dominant than WD, but the overall pattern of low and generally increasing cospectra at longer length scales did not change. The implications of the wavelet analysis for global change science are as follows. First, it is clear that the dominant control of GSL spatial variability is land cover wherein the dominant class(es) usually had the dominant cospectra. That current patterns of GSL in two regions were strongly associated with spatial variability of the AG class indicates that the human presence is already strongly apparent. Second, although our data are from a limited region, it appears that in areas of limited and spatially varied precipitation (as in the west), SWC is an important factor for GSL variability, especially at short and long length scales. Degradation or alteration of soil resources (Imhoff et al. 1997) could therefore cause changes in phenological patterns. Third, climatic variables are remarkably unrelated to spatial GSL variation. Interannual climatic variation controls long- and short-term phenological variation through time at individual locations or regions, but compared to land cover factors, climate was much less important to the spatial variation in GSL. Fourth and perhaps of greatest interest to the GCM community, the DLS occurs at 2km, arguing that even for GCM simulations using fine-scale nested grids, the effect of sub-pixel phenological heterogeneity on climate models should be rigorously considered.

5.

CHALLENGES IN GLOBAL CHANGE PHENOLOGY

Phenology is clearly an important topic within global change science. Global science in fields as diverse as remote sensing (Myneni et al. 1997a), atmospheric CO2 analysis (Keeling et al. 1996), climate modeling (above), and carbon cycle modeling (Lucht et al. 2002) has identified phenology as a crucial component of the Earth system. However, the very cross-disciplinary nature of phenology has created a serious tendency to misinterpret phenological research. The term GSL is especially open to multiple interpretations. We identify the following four possible methods of defining GSL: (1) remote sensing (see Chapter 5.1, this book); (2) weather; (3)

Chapter 7.1: Vegetation Phenology in Global Change Studies

461

ecophysiology; and (4) canopy duration. Each implies a separate definition of the growing season yet in practice the global change community often uses the terms interchangeably. We used the Vegetation/Ecosystem Modeling and Analysis Project (VEMAP, Kittel et al. 1999) dataset to demonstrate differences and similarities in the various approaches. The CONUS VEMAP dataset contains edaphic and daily meteorological records for 1895-1993 at a 0.5°×0.5° resolution. We calculated GSL for the four categories as follows. Remote sensing. We obtained 1981-2001 monthly observations of the fraction of photosynthetically active radiation absorbed by vegetation canopies (FPAR, Myneni et al. 1997b) that were regridded and composited from the original 8km 10-day dataset to the VEMAP resolution. For each pixel, we used a spline function to obtain daily values and estimated GSL for each year as the number of days with FPAR > 0.5 (FPAR-GSL). This is an arbitrary threshold and is used to illustrate the concept of a remotely-sensed estimate of GSL rather than to quantitatively predict a certain vegetative status. These data have been corrected for within- and between-sensor calibration issues and may be used to assess trends and interannual variability. Weather. We used the VEMAP minimum temperature records to calculate the frost-free GSL, i.e. number of days each year with minimum temperatures > 0°C (FF-GSL). Ecophysiology. We used ecosystem simulations from the Biome-BGC model (Nemani et al. 2002) to calculate the carbon uptake period (CUP). CUP indicates the number of days each year in which the assimilation of carbon through photosynthesis exceeded the sum of heterotrophic and autotrophic respiration (CUP-GSL). In other words, CUP refers to the number of days with net ecosystem carbon uptake. For each VEMAP pixel, up to five PFTs were simulated (shrubs, C3 grasses, C4 grasses, evergreen needle leaf forest, and deciduous broad leaf forest). Values were scaled by their proportional land cover within each VEMAP pixel. See Nemani et al. (2002) for a discussion of Biome-BGC VEMAP simulations. Canopy duration. Using the same Biome-BGC simulations, we extracted estimates of canopy duration. For evergreen canopies, canopy duration was defined as 365 days. For deciduous canopies, annual phenology was controlled by the Biome-BGC phenology subroutines developed in White et al. (1997). Pixel canopy duration (CD-GSL) was defined as a weighted average as for CUP. We extracted GSL estimates for the period containing both VEMAP and remote sensing data (1982-1993). Plate 2 shows the mean (left images) and standard deviation (right images) of these estimates. FPAR-GSL was 365 days for much of the southeastern and coastal western US, less than 10 days

462

Phenology: An Integrative Environmental Science

throughout much of the intermountain West, and highly variable in the south-central region. Use of constant thresholds of this type for any geographic region with significant PFT variation will produce regions with perpetual GSL and regions with no GSL whatsoever. This supports the notion of using variable thresholds (White et al. 2002) or information regarding the shape of the vegetation time series (Reed et al. 1994). FF-GSL had similarly high western and southern GSL but a much more homogenous GSL in the continental interior. Compared to other GSL measures, variability was consistently low with highest values in the Pacific Northwest. CUP-GSL was highest in the Southeast and West with low variability in the East and high variability in the West related to interannual variation in precipitation. CD-GSL was spatially similar to CUP-GSL but was generally longer, especially for the evergreen needle leaf dominated regions, where variability was also extremely low or equal to zero. In spite of dramatic differences in the absolute values of GSL estimates (Figure 1a), continental average GSL anomalies were similar (Figure 1b). With the exception of a very low FPAR-GSL in 1982-1983 (possibly related to the El Chichon volcanic eruption), the four GSL metrics tracked the same

Figure 7.1-1. 1982-1993 growing season length (GSL) interannual variability based on days with FPAR > 0.5 (dashed grey); frost-free days (solid grey); carbon uptake period (black dashed); and canopy duration (solid black). (A) Absolute values. (B) Standardized anomalies of (A).

Chapter 7.1: Vegetation Phenology in Global Change Studies

463

interannual patterns with a remarkable degree of coherence (Figure 1b and Table 1). Correlation among the four metrics (Table 1) was lowest for comparisons including FPAR-GSL. FF-GSL was highly correlated with both CUP-GSL and CD-GSL. CUP-GSL and CD-GSL had the highest r2. In summary, the comparison of GSL metrics illustrated that (1) the different methods yielded dramatically different absolute values, (2) spatial patterns were somewhat consistent along the continental boundaries but extensively different in the western US, and (3) at a continental level, all techniques tracked similar patterns of interannual variability. We believe that these results highlight the potential for confusion when using the generic term growing season length but also the surprising ability of all GSL measures to track interannual variability. We argue for more explicit terms such as carbon uptake period or canopy duration. In cases when use of the term growing season length is most appropriate, as for many satellite applications, an explicit attempt should be made to explain how the satellite observations correspond to ground conditions (e.g. White et al. 1999). Table 7.1-1. 1982-1993 CONUS growing season length (GSL) correlations (r2). FPAR, days with FPAR > 0.5; FF, days with minimum temperature > 0°C; CUP, number of days with net ecosystem uptake of carbon; CD=canopy duration (number of days with 1-100% of canopy present). FF CUP CD FPAR FPAR FF CUP CD

6.

1 0.45 0.28 0.54

0.45 1 0.69 0.70

0.28 0.69 1 0.79

0.54 0.70 0.79 1

CONCLUSIONS

Vegetation phenology is a crucial field for global change studies. Other chapters discuss the use of vegetation phenology in monitoring, modeling, remote sensing, and climate/weather. Here, we illustrated three aspects of vegetation phenology and global change. First, we discussed the importance of phenology in the context of climate simulations. Second, using wavelet analysis of U.S. vegetation phenology, we showed that the dominant length scale of spatial variation was 2km. Spatial variation in phenology was in most cases more strongly related to land cover variation than to edaphic or climatic variation, illustrating the potential of humans to rapidly modify landscape phenology patterns. Third, using a comparison of growing season length calculations, we highlighted the potential for confusion when implementing phenological concepts in global change research.

464

Phenology: An Integrative Environmental Science

ACKNOWLEDGEMENTS Michael White was supported by NASA grant NAG5-11282.

REFERENCES CITED Asner, G., A. Townsend, and B. Braswell, Satellite observation of El Niño effects on Amazon forest phenology and productivity, Geophys. Res. Lett., 27, 981-984, 2000. Bonan, G. B., A Land Surface Model (LSM version 1.0) for ecological, hydrological, and atmospheric studies: Technical description user's guide, National Center for Atmospheric Research, Boulder, CO, 150 pp., 1996. Bonan, G. B., S. Levis, L. Kergoat, and K. W. Oleson, Landscapes as patches of plant functional types: an integrating concept for climate and ecosystem models, Global Biogeochem. Cycles, In press, 2002a. Bonan, G. B., S. Levis, S. Sitch, M. Vertenstein, and K. W. Oleson, A dynamic global vegetation model for use with climate models: concepts and description of simulated vegetation dynamics, Global Change Biol., Submitted, 2002b. Botta, A., N. Viovy, P. Ciais, P. Friedlingstein, and P. Monfray, A global prognostic scheme of leaf onset using satellite data, Global Change Biol., 6, 709-726, 2000. Buermann, W., J. Dong, X. Zeng, R. B. Myneni, and R. E. Dickinson, Evaluation of the utility of satellite-based vegetation leaf area index data for climate simulations, J. Climate, 14, 3536-3550, 2001. Claussen, M., and V. Gayler, The greening of the Sahara during the mid-Holocene: results of an interactive atmosphere-biosphere model, Global Ecol. Biogeogr., 6, 369-377, 1997. Coops, N., Improvement in predicting stand growth of Pinus radiata (D. Don) across landscapes using NOAA AVHRR and Landsat MSS imagery combined with a forest growth process model (3-PGs), Photogram. Eng. Remote Sens., 65, 1149-1156, 1999. Cosby, B. J., G. M. Hornberger, R. B. Clapp, and T. R. Ginn, A statistical exploration of the relationships of soil moisture characteristics to the physical properties of soils, Water Resour. Res., 20, 682-690, 1984. Cramer, W., A. Bondeau, F. I. Woodward, I. C. Prentice, R. A. Betts, V. Brovkin, P. M. Cox, V. Fisher, J. A. Foley, A. D. Friend, C. Kucharik, M. R. Lomas, N. Ramankutty, S. Sitch, B. Smith, A. White, and C. Young-Molling, Global response of terrestrial ecosystem structure and function to CO2 and climate change: results from six dynamic global vegetation models, Global Change Biol., 7, 357-373, 2001. Csillag, F., and S. Kabos, Wavelets, boundaries, and the spatial analysis of landscape patterns, Écoscience, 9, 177-190, 2002. DeFries, R. S., L. Bounoua, and G. J. Collatz, Human modification of the landscape and surface climate in the next fifty years, Global Change Biol., 8(5), 438-458, 2002. Dorman, J. L., and P. J. Sellers, A global climatology of albedo, roughness length, and stomatal resistance for atmospheric general circulation models as represented by a simple biosphere model (SiB). J. Appl. Meteorol., 28, 833-855, 1989. Foley, J., S. Levis, M. Costa, W. Cramer, and D. Pollard, Incorporating dynamic vegetation cover within global climate models, Ecol. Appl., 10, 1620-1632, 2000. Foster, D., G. Motzkin, and B. Slater, Land-use history as long-term broad-scale disturbance: regional forest dynamics in central New England, Ecosystems, 1, 96-119, 1998.

Chapter 7.1: Vegetation Phenology in Global Change Studies

465

Imhoff, M., W. Lawrence, C. Elvidge, T. Paul, E. Levine, M. Privalsky, and V. Brown, Using nighttime DMSP/OLS images of city lights to estimate the impact of urban land use on soil resources in the United States, Remote Sens. Environ., 59, 105-117, 1997. Intergovernmental Panel on Climate Change Working Group I, Climate Change 2001: The Scientific Basis, Cambridge University Press, Cambridge, 2001. Keeling, C. D., J. F. S. Chin, and T. P. Whorf, Increased activity of northern vegetation inferred from atmospheric CO2 measurements, Nature, 382, 146-149, 1996. Kiehl, J., J. Hack, G. B. Bonan, B. Bonville, D. L. Williamson, and P. J. Rasch, The National Center for Atmospheric Research Community Climate Model: CCM3, J. Climate, 11, 1131-1149, 1998. Kimball, J. S., S. W. Running, and R. Nemani, An improved method for estimating surface humidity from daily minimum temperatures, Agric. For. Meteorol., 85, 87-98, 1997. Kittel, T. G. F., N. A. Rosenbloom, T. H. Painter, D. S. Schimel, and V. M. Participants, The VEMAP integrated database for modelling United States ecosystem/vegetation sensitivity to climate change, J. Biogeogr., 22, 857-862, 1999. Kumar, P., and E. Foufoula-Georgiou, A multicomponent decomposition of spatial rainfall fields, 1. segregation of large- and small-scale features using wavelet transforms, Water Resour. Res., 29, 2515-2532, 1993. Kumar, P., and E. Foufoula-Georgiou, Wavelet analysis for geophysical applications, Rev. Geophys., 35, 385-412, 1997. Levis, S., J. A. Foley, V. Brovkin, and D. Pollard, On the stability of the high-latitude climate-vegetation system in a coupled atmosphere-biosphere model, Global Ecol. Biogeogr., 8(6), 489-500, 1999. Lloyd, D., A phenological classification of terrestrial vegetation cover using shortwave vegetation index imagery, Int. J. Remote Sens., 11, 2269-2279, 1990. Lu, L. X., and W. J. Shuttleworth, Incorporating NDVI-derived LAI into the climate version of RAMS and its impact on regional climate, J. Hydrometeorol., 3 (3), 347-362, 2002. Lucht, W., I. C. Prentice, R. B. Myneni, S. Sitch, P. Friedlingstein, W. Cramer, P. Bousquet, W. Buermann, and B. Smith, Climatic control of the high-latitude vegetation greening trend and Pinatubo effect, Science, 2966 (5573), 1687-1689, 2002. Mallat, S., A Wavelet Tour of Signal Processing, Academic Press, New York, 577 pp., 1999. Menzel, A., Phenology, its importance to the global change community, Climatic Change, 54, 379-385, 2002. Miller, D. A., and R. A. White, A conterminous United States multilayer soil characteristics dataset for regional climate and hydrology modeling (on line), Earth Interactions, paper 1, 1998. Moulin, S., L. Kergoat, N. Viovy, and G. Dedieu, Global scale assessment of vegetation phenology using NOAA/AVHRR satellite measurements, J. Climate, 10, 1154-1170, 1997. Myneni, R. B., C. D. Keeling, C. J. Tucker, G. Asrar, and R. R. Nemani, Increased plant growth in the northern high latitudes from 1981 to 1991, Nature, 386, 698-702, 1997a. Myneni, R. B., R. R. Nemani, and S. W. Running, Estimation of global leaf area index and absorbed Par using radiative transfer models, IEEE T. Geosci. Remote, 35, 1380-1393, 1997b. Myneni, R. B., and D. L. Williams, On the relationship between FAPAR and NDVI, Remote Sens. Environ., 49, 200-211, 1994. Nemani, R. R., M. A. White, P. E. Thornton, K. Nishida, S. Reddy, J. Jenkins, and S. Running, Recent trends in hydrologic balance have enhanced the terrestrial carbon sink in the United States, Geophys. Res. Lett., 29, art. no. 1468, 2002.

466

Phenology: An Integrative Environmental Science

Percival, D., On estimation of the wavelet variance, Biometrika, 82, 619-631, 1995. Potter, C., E. Davidson, D. Nepstad, and C. Reis de Carvalho, Ecosystem modeling and dynamic effects of deforestation on trace gas fluxes in Amazon tropical forests, Forest Ecol. Manag., 152, 97-117, 2001. Reed, B. C., J. F. Brown, D. VanderZee, T. R. Loveland, J. W. Merchant, and D. O. Ohlen, Measuring phenological variability from satellite imagery, J. Veg. Sci., 5, 703-714, 1994. Schwartz, M. D., B. Reed, and M. A. White, Assessing satellite-derived start-of-season (SOS) measures in the conterminous USA, Int. J. Climatol., 22, 1793-1805, 2002. Thornton, P. E., S. W. Running, and M. A. White, Generating surfaces of daily meteorological variables over large regions of complex terrain, J. Hydrol., 190, 214-251, 1997. Torrence, C., and G. P. Compo, A practical guide to wavelet analysis, Bull. Amer. Meteorol. Soc., 79, 61-78, 1998. Vitousek, P., Beyond global warming: ecology and global change, Ecology, 75, 1861-1876, 1994. White, M. A., R. R. Nemani, P. E. Thornton, and S. W. Running, Satellite evidence of phenological differences between urbanized and rural areas of the eastern United States deciduous broadleaf forest, Ecosystems, 5, 260-273, 2002. White, M. A., M. D. Schwartz, and S. W. Running, Young students, satellites aid U 81, 1,5, 1999. understanding of climate-biosphere link, EOS Trans. AGU, White, M. A., P. E. Thornton, and S. W. Running, A continental phenology model for monitoring vegetation responses to interannual climatic variability, Global Biogeochem. Cycles, 11, 217-234, 1997.

Chapter 7.2 PHENOLOGY OF VEGETATION PHOTOSYNTHESIS Lianhong Gu1, Wilfred M. Post1, Dennis Baldocchi2, T. Andy Black3, Shashi B. Verma4, Timo Vesala5, and Steve C. Wofsy6 1

Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA; Department of Environmental Science, Policy & Management, University of California, Berkeley, CA, USA; 3Faculty of Agricultural Sciences, University of British Columbia, Vancouver, Canada; 4School of Natural Resource Sciences, University of Nebraska, Lincoln NE, USA; 5Department of Physical Sciences, University of Helsinki, Helsinki, Finland; 6 Department of Earth and Planetary Sciences, Harvard University, Cambridge, MA, USA 2

Key words:

1.

Photosynthesis, Ecosystems

Phenological

indices,

Vegetation,

Eddy

covariance,

INTRODUCTION

Traditionally, the study of plant phenology focuses on monitoring and analyses of the timing of life-cycle events. In phenological observations, one of the most important tools has been our eye. Consequently, the life-cycle events that are subject to most phenological investigations have been those that are easily visible, e.g. leaf bud break, first flowering, leaf coloring, leaf fall, etc. Although plant scientists have been always interested in the physiological bases that control the phenological stages of plants (e.g., Mott and Mccomb 1975; Suárez-López et al. 2002; Yanovsky and Kay 2002), analyses and prediction of the timing of these visible events have dominated phenological studies in the past (e.g., Lieth 1974; Podolsky 1984; Peñuelas and Filella 2001). In conjunction with this feature of “visibility”, traditional phenological studies are also characterized by “individuality” in the sense that individual species instead of plant communities are subject to Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 467-485 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

468

Phenology: An Integrative Environmental Science

observations, and “discontinuity” in the sense that the observational time domain is discontinuous (often once a year). These legacies in past phenological studies need a careful look. Our ability to predict a phenological event will eventually depend on how well the biological mechanism is understood. Visible life-cycle events are results and external exhibitions of internal plant physiological processes that operate under changing environmental conditions. They represent the structural aspects of plant activities. Equally revealing about the underlying processes are changes in the functional aspects of plant activities, in particular, the exchanges of mass and energy between plants and their environment (e.g., Fitzjarrald et al. 2001; Freedman et al. 2001; Schwartz and Crawford 2001). These functional aspects of plant activities, although not visible, provide quantitative measures of plant responses to changes in environmental conditions and thus are important for the study of phenology. Shifts in the timing of phenological events serve as powerful biological indicators of global climate change. While targeting individual species makes visual observation easier, phenological responses to environmental changes are species-specific (e.g., Bradley et al. 1999; Spano et al. 1999; Peñuelas and Filella 2001). This makes the comparison of changes in the ontogenetic development of different plants in response to climate changes difficult. Efforts have been made to use cloned plants to obtain comparable phenological data across different regions (Schwartz 1994; Chmielewski 1996; Menzel and Fabian 1999). Although cloned plants help eliminate uncertainty due to genetic variability, species adaptation to local climates may bring additional uncertainty to phenological comparisons in different climatic regions. Eventually the use of cloned plants will be limited by their biogeographical distributions. These potential difficulties with individual plant species point to the need for also using responses of a collection of plant species or plant communities as indicators of climate change. How plants reach from one life-cycle event to another is a topic hard to deal with within the traditional framework of plant phenology because of its focus on discontinuous life-cycle events. However, plant lives are not just composed of individual events but also continuous responses and feedbacks to the ever-changing environmental conditions. To understand precisely how environmental variables affect plant life-cycle events, plant-environment interactions over the whole time domain must be considered. In this regard, continuous measurements of exchanges of mass and energy between plants and their environment provide valuable information about plant activities between life-cycle events. Today’s expanded interest in phenology stems from the realization of the potential roles that phenology can play in monitoring and understanding biospheric responses to global climate change. This new application of

Chapter 7.2: Phenology of Vegetation Photosynthesis

469

phenology calls for new dimensions and new tools in phenological studies. To this end, establishing a linkage between traditional phenology and terrestrial biophysics is of critical importance. Such a linkage is needed because the interactions between vegetation and atmosphere are realized through various biophysical fluxes (carbon dioxide, water vapor, sensible heat, etc) that transfer between them, and these biophysical fluxes are affected by phenological events (Baldocchi et al. 2001; Fitzjarrald et al. 2001; Freedman et al. 2001; Schwartz and Crawford 2001). In this chapter, we take an ecosystem-level functional approach to phenological studies. Instead of focusing on life-cycle events of individual plants, we explore seasonal demarcations in canopy photosynthesis, which contain rich information about plant phenology. Instead of basing our analysis on individual species, we use plant communities (vegetation canopies) as our observational subjects. Such a community-based approach complements the traditional species-based approach in the sense that the former captures the broad vegetational responses while the latter reveals the sensitive changes of individual species. We are interested in not only the timing of individual phenological events but also how communities change from one event to another. Our objective is to develop a systematic methodology that can be used to extract phenological information on photosynthesis from continuous measurements of net ecosystem exchanges (NEE) of carbon dioxide (CO2). We derive a series of phenological indices that can be used to characterize seasonal patterns of vegetation photosynthesis. Our observational tool is the eddy covariance method, a micrometeorological technique that measures trace gas fluxes between the biosphere and atmosphere. Although transitions in surface energy balance are closely related to leaf phenology, and it is possible to extract phenological information from energy flux measurements also (Baldocchi et al. 2001; Fitzjarrald et al. 2001; Schwartz and Crawford 2001), this chapter focuses on the CO2 flux only because of the current emphasis on the global carbon cycle by the global change community. In the following sections, we first introduce the eddy covariance technique and the available flux dataset. Then we outline the model that is used to infer canopy photosynthetic rates from NEE measurements. From there we describe how canopy photosynthetic rates can be used to characterize phenology on a plant community level.

2.

EDDY COVARIANCE FLUX MEASUREMENTS

The eddy covariance technique measures vertical flux densities of scalars (e.g., CO2, water vapor, temperature) between ecosystems and the

470

Phenology: An Integrative Environmental Science

atmosphere by determining the mean covariance between the vertical wind velocity and the respective scalar fluctuations. Typical instrumentation at an eddy covariance site includes a three-dimensional sonic anemometer, to measure wind velocities and virtual temperature, and a fast responding sensor to measure CO2 and water vapor. Scalar concentration fluctuations are measured with open and closed path infrared gas analyzers. Over short canopies, these sensors can be mounted on small poles, while walk-up scaffolding or low-profile radio towers are used over forest canopies. Sampling rates between 10 and 20 Hz ensure complete sampling of the high frequency portion of the flux co-spectrum. The sampling duration must be long enough to capture low frequency contributions to flux covariance, but not too long to be affected by diurnal changes in temperature, humidity and CO2. Adequate sampling duration and averaging period vary between 30 and 60 minutes for most teams. Regular calibrations of gas analyzers are needed to correct potential instrument zero and span drifts. For further information on the eddy covariance method, readers are referred to Baldocchi et al. (1988) and Aubinet et al. (2000). There are now over 200 eddy covariance flux tower sites operating on a long-term and continuous basis globally. At most sites, researchers also collect data on site vegetation, soil, hydrologic, and meteorological characteristics. Regional and global networks have been formed with the aim to understand the mechanisms controlling the flows of CO2, water, and energy to and from the terrestrial biosphere across the spectrum of time and space scales. Data compiled by these flux networks are freely available to the science community through the internet and provide unique opportunities to study biophysical aspects of plant community phenology. Further information on these flux networks and available datasets can be found at http://public.ornl.gov/ameriflux, or http://www-eosdis.ornl.gov/FLUXNET.

3.

INFERRING VEGETATION PHOTOSYNTHESIS

The eddy covariance method measures NEE, which is composed of two components: canopy photosynthetic flux density (gross photosynthesis, Pg) and ecosystem respiration rate ((Re): Ne = Re − Pg

(1)

where Ne denotes NEE of CO2. Re consists of carbon flux from autotrophs (leaves, roots, and shoots of plants) and heterotrophs (microbes, fungi, bacteria, etc.) to the atmosphere. For phenological studies, it is desirable to

Chapter 7.2: Phenology of Vegetation Photosynthesis

471

analyze Pg instead of Ne or Re since Pg is more directly related to plant community activities. We apply the model of Gu et al. (2002) to infer Pg from measured Ne. In Gu et al. (2002), the following function is used to express the dependence of Pg on diffuse photosynthetically active radiation (diffuse PAR, If) and direct photosynthetically active radiation (direct PAR, Ir): Pg =

(

(

)(( ) (

) )I t

(2)

where αf and αr are the initial canopy quantum yield for diffuse and direct PAR, respectively; βf and βr are the closeness to linear response (CLR) coefficient for diffuse and direct PAR, respectively; It is global PAR (= If + Ir). Equation (2) is a generalization to the rectangular hyperbola model commonly used in flux analyses (see references in Gu et al. 2002). The advantage of Equation (2) over the traditional rectangular hyperbola model is that it can describe the differential canopy photosynthetic responses to diffuse and direct PAR. Gu et al. (2002) compared the new model with the traditional rectangular hyperbola for a variety of vegetation sites and found that the new model provides a better fit to observations. The dependence of Re on temperature is described by the following function: Re = c1e c2 [ 3

a

(

3

) s ] + d e d 2Ts 1

(3)

where c1, c2, c3, d1 and d2 are regression coefficients; Ts is soil temperature; Ta is air temperature. With measured Ne and Equations (1)-(3), parameters in (2) and (3) can be estimated through nonlinear regression. To obtain Pg, estimated αf, αr, βf and βr are used in (2). Equation (3) can be used to estimate Re from fitted values of c1, c2, c3, d1 and d2 ((Re is not used in this study).

4.

SEASONAL PHOTOSYNTHETIC PATTERNS OF CONTRASTING TERRESTRIAL ECOSYSTEMS

We have used the method described in the previous section to infer canopy photosynthetic rates for five sites with contrasting vegetation structures and climatic conditions. These sites are a Scots pine forest in Hyytiälä, Finland (61°51'N, 24°17′E, data from 1997), an aspen forest in

472

Phenology: An Integrative Environmental Science

Prince Albert National Park, Saskatchewan, Canada (53°63′N, 106°20′W, 1996), a mixed deciduous forest in Walker Branch Watershed, Tennessee, USA (35°58′N, ′ 84°17′W, 1996), a mixed hardwood forest in Massachusetts, USA (Harvard Forest, 42°32' N, 72°10'W, 1992), and a native tallgrass prairie in Oklahoma, USA (36°56′N, 96°41′W, 1997). Detailed site information can be found elsewhere (Goulden et al. 1996; Gu et al. 2002, and references therein). Since diurnal variations in canopy photosynthetic rates do not provide phenological information, we focus on seasonal patterns. Seasonal patterns are characterized by daily maximum canopy photosynthetic rates ((P Pm). To determine Pm, canopy photosynthetic rates Pg (time step is hourly at the Harvard Forest site, and half-hourly at other sites) in each day are ranked, and the largest value is taken as Pm. For a given site, daily maximum canopy photosynthetic rates form the upper boundary in the scatter plot of canopy photosynthetic rates against time. A few daily maximum values clearly fall out of seasonal patterns due to excessive cloudiness. They can be easily picked up visually. The total number of these points is very small, and they are excluded from the analysis. Figure 1 depicts the seasonal changes of Pm at these five sites (unusually small Pm points have been already removed). Also shown in Figure 1 are values calculated from the Weibull function fitted to these data. This curve fitting is necessary for the determination of phenological indices in the next section. The Weibull function is given as: ⎡ ⎧ ⎛ ⎢ − t −t0 + b ( ⎜ ⎢ b ⎪ ⎣ ⎪ Pm ( ) = y0 + a⎜1 − e ⎜⎜ ⎨ ⎝ ⎪ ⎪ ⎩ Pm ( )

)1 / c

c⎤

⎥ ⎥ ⎦

⎞ ⎟ ⎟, ⎟⎟ ⎠

t ≥t 0 −b(

0

)1/ c (

(4)

)1/ c

t is the daily maximum canopy photosynthetic rate for day t; y0, where Pm(t) t0, a, b and c are regressional coefficients. Because the Weibull function is monotonic (that is, Pm keeps increasing with t), the fitting has to be done separately for spring and fall. The choice for the end day of the spring portion or the start day for the fall portion does not critically affect the fitting parameters as long as they are within the period in which Pm is relatively constant in the middle growing season. In the fall fitting, the number of days from the end of the calendar year is used as the independent variable. SigmaPlot has a built-in nonlinear regression for the Weibull function and is used in this study. Our choice of the Weibull function results from many trials and failures. This function fits to observations better than other ones we have tried (e.g., sigmoid). It provides a very good fit at all five sites even through the sites

Chapter 7.2: Phenology of Vegetation Photosynthesis

473

are quite different in vegetation types. The R2 of the spring fitting and fall fitting are 0.96 and 0.94 for Harvard Forest, 0.99 and 0.99 for the tallgrass prairie, 0.98 and 0.97 for the Scots pine forest, 0.98 and 0.97 for the mixed forest in Tennessee, and 0.98 and 0.98 for the aspen forest, respectively.

Daily maximum canopy photosynthetic rate -2 -1 (μmol m s )

40

a

Tallgrass prairie

30

20

10

Scots pine forest

Daily maximum canopy photosynthetic rate -2 -1 (μmol m s )

0 40

Mixed forest (Tennessee)

b

30

Mixed forest (Harvard Forest)

20

10

Aspen forest 0 0

100

200

300

Day of year

Figure 7.2-1. Seasonal patterns of daily maximum canopy gross photosynthetic rates at five eddy covariance flux sites.

474

Phenology: An Integrative Environmental Science

The seasonal patterns in Pm show both differences and similarities among the sites. For example, the growing season of the Scots pine forest almost coincides with that of the tallgrass prairie even through the two sites are nearly 25 degrees (in latitude) apart from each other. However, photosynthesis in the tallgrass prairie increases faster in the spring and reaches a higher maximum rate than does the Scots pine forest. The former starts to decline earlier, leading to a shorter middle growing season when Pm is relatively stable. The three deciduous forests have different growing season lengths, with the aspen forest being the shortest. This is not surprising since the aspen forest site is the most northern of the three sites. However, photosynthesis in the aspen forest increases rapidly in the spring, leading to a middle-growing season photosynthetic peak similar to Harvard Forest’s. These seasonal photosynthetic patterns reflect the responses of plant communities to seasonal cycles in climate conditions. Studying the relationships between the seasonal photosynthetic patterns and climate conditions can yield important insights into how terrestrial ecosystems will respond to global climate changes. However, currently we lack systematic approaches to objectively quantify these seasonal patterns. For example, growing season length has been found to be an important indicator of annual carbon sequestrations (Baldocchi et al. 2001). But how is the growing season defined? As shown in Figure 1, entering into the growing season is a gradual process. Determining the initiation or end day of the growing season can be subjective. Also, two plant communities may have the same growing season length, but differ dramatically in their canopy photosynthetic rates within the growing season (Figure 1a). Therefore, it is desirable to develop a methodology and a set of phenological indices that can be used to objectively and quantitatively characterize the overall patterns of seasonal photosynthetic dynamics of plant communities.

5.

PHENOLOGICAL INDICES OF VEGETATION PHOTOSYNTHESIS

Here we introduce a number of indices to quantify seasonal photosynthetic patterns. These indices include: spring photosynthesis development velocity, fall photosynthesis recession velocity, growing season initiation day, growing season termination day, center day of the growing season, length of the growing season, effective length of the growing season, effective daily maximum canopy photosynthetic rate, and seasonal carbon dioxide assimilation potential index. Figure 2 uses Harvard Forest (Figure 2a) and Scots pine forest (Figure 2b) as examples to illustrate these indices.

Chapter 7.2: Phenology of Vegetation Photosynthesis

5.1

475

Spring Photosynthesis Development and Fall Photosynthesis Recession Velocities

Daily maximum canopy photosynthetic rates, Pm, tend to change linearly with time in spring and fall at all five sites (Figures 1, 2). These linear trends make it possible to define a spring photosynthesis development velocity (V Vs, μmol/m2 s day) and a fall photosynthesis recession velocity (V Vf, μmol/m2 s day), which represent the slopes of the two time periods, respectively. These two pace parameters should contain important information about the response of vegetation to changes in climate conditions. The key is to determine when the linear segments start and when they end. To do so, we note that the seasonal pattern of Pm shows four sharp turning points with two points in the spring and the other two in the fall. The first sharp turning point in the spring represents the day when canopy photosynthesis development starts to accelerate in response to the rapid improvement in meteorological conditions for plant growth. This day is termed photosynthesis upturn day (PUD, denoted by Du). The second sharp turning point in the spring corresponds to the day when the development process of canopy photosynthesis starts to slow down and gradually move towards stabilization in which Pm reaches its peak. This day is termed photosynthesis stabilization day (PSD, denoted by Ds). The spring linear segment lies between PUD and PSD. The first sharp turning point in the fall corresponds to the day when canopy photosynthesis enters into a period of quick decline in response to the deterioration in meteorological conditions for plant growth. This day is termed photosynthesis downturn day (PDD, denoted by Dd). As the recession approaches the end of the growing season, the speed of recession tends to decrease, possibly due to residual photosynthesis by some leaves or conifer species in the plant community, leading to a second sharp turning point in the fall. After the second sharp turning point in the fall, photosynthesis of deciduous trees proceeds to stop while evergreen species may still maintain a low level, limited photosynthetic rate. This second turning point is termed photosynthesis recession day (PRD, denoted by Dr). These sharp turning points can be located by calculating the radius of curvature of Pm(t). t The radius of curvature measures the sharpness that a curve turns. For a given point on a curve, the radius of curvature is the radius of the circle that fits or “kisses” the curve at that point. The more sharply the curve turns, the smaller the radius of curvature is, and vice versa. Mathematically, for curve f = Pm(t), t its radius of curvature is ρ(tt). The

476

Phenology: An Integrative Environmental Science 2 ⎡ ⎛ dP P ⎞ ⎤ ⎢1 + ⎜ m ⎟ ⎥ ⎢ ⎝ dt ⎠ ⎥⎦ ȡ= ⎣ d 2 Pm dt 2

1.5

(5)

t sharpest turning points of curve Pm(t) correspond to the local minima of ρ (t) (for details on radius of curvature, consult a calculus textbook such as Thomas and Finney 1998, p. 881-893). It is possible to determine the radius of curvature from the calculated Pm directly. However this would require a rigorous smoothing procedure to remove the noises in the numerical first and second order derivatives of the calculated Pm. Here we use the fitted Weibull function (4) as shown in Figures 1 and 2. The first and second derivatives of Pm(t) are given by:

dPm ac − k c c −1 = e k dt b

d 2 Pm dP = dt 2 dt

if k =

(6a)

(

t − t 0 + b( b

)bk1 )1 / c

(6b)

≥ 0 ; and

dP Pm =0 dt

(6c)

d 2 Pm =0 dt 2

(6d)

if k < 0.

Substituting Expressions (6) into (5) leads to ρ as a function of day of year. Although we could determine the minima of ρ by letting dρ/dt = 0 and solving the resulted complicated third order differential equation, it does not have to be done in this way. We can determine the days when the minimum

Chapter 7.2: Phenology of Vegetation Photosynthesis

477

ρ occurs by simply ranking the ρs calculated for all days in the year and selecting those days when ρ reaches a minimum. The two points ((P Pm(D ( u), Du) and ((P Pm(D ( s), Ds) form a line, which we call “Spring Development Line.” The slope of the spring development line, that is, the spring photosynthesis development velocity is given by: Vs =

Pm (

s

)

m

( u)

(7)

Ds - Du

Similarly, the “Fall Recession Line” is formed by the two points ((Pm(D ( d), Dd) and (P (Pm(D ( r), Dr). The fall photosynthesis recession velocity is determined by: Vf =

5.2

Pm (

d

)

m

( r)

(8)

Dr - Dd

Growing Season Initiation and Termination Days

We define the intersection between the spring development line and the time (day of year) axis as the initiation of the growing season (see Figure 2). It is easy to show that the growing season initiation day ((Di) is: Di =

Du * P ( P(

) ) s s

s

( u)

( u)

(9)

In a similar fashion, we define the growing season termination day (D ( t) as the intersection between the fall recession line and the time axis: Dt =

Dr * P( P(

) d)

d

d

( r)

( r)

(10)

The length of the growing season ((L) is simply: L = Dt − Di

(11)

Daily maximum canopy photosynthetic rate -2 -1 (μmol m s )

478

Phenology: An Integrative Environmental Science

Harvard Forest a

30

B PSD

PDD

20 Center of growing season (Dc)

10 PRD

PUD

0 0

100 Growing season starts

25

Daily maximum canopy photosynthetic rate -2 -1 (μmol m s )

C

Spring development line

Fall recession line

A

200

D

300

Day of year

Growing season ends

B

Fall recession line

Scots Pine Forest b 20

Spring development line

C

PDD

PSD 15

Center of growing season (Dc)

10

PUD

5

PRD

0 0

100 Growing season starts

A

200

Day of year

D

300 Growing season ends

Figure 7.2-2. Determination of phenological indices of vegetation photosynthesis. See text for explanation.

Chapter 7.2: Phenology of Vegetation Photosynthesis

5.3

479

Center Day and Effective Length of the Growing Season

We define the center day of the growing season ((Dc) as the “center of gravity” of the seasonal photosynthesis curve Pm(t):

1 w2

Dc =

∫

Dt

Di

tPm2 dt

(12)

where: Dt

w2 = ∫ Pm2 dt

(13)

Di

Center day of the growing season is marked in Figure 2. The effective length of the growing season ((Le) is defined as the standard deviations from the center day of the growing season Dc: Le =

2 3 ⎡ Dt ( w ⎢⎣ ∫Di

2 2 ⎤ c ) Pm dt ⎥

⎦

1/ 2

(14)

where the factor 2 3 is introduced so that the effective length of the growing season would equal exactly the width of the rectangle if the seasonal pattern of vegetation photosynthesis is perfectly rectangular. In Figure 2(a and b), the number of days between the two points A and D is the effective length of the growing season.

5.4

Seasonal Carbon Assimilation Potential Index and Effective Maximum Canopy Photosynthetic Rate

Let u = ∫ Pm ( )dt Dt

(15)

Di

where u is simply the area under the curve Pm(t) within the growing season (see Figure 2). It is an indicator of seasonal carbon dioxide assimilation potential by a canopy under a give climate condition. We call it the seasonal carbon assimilation potential index.

480

Phenology: An Integrative Environmental Science

The effective daily maximum canopy photosynthetic rate (Pe) is calculated from the effective length of the growing season Le and the seasonal carbon assimilation potential index u: Pe =

u Le

(16)

In Figure 2(a and b), the height of the rectangular ABCD is Pe. In this section, we have introduced a series of indices and the methods to determine their values without much elaborating their usefulness. This is the task of the next section.

6.

EVALUATION OF PHENOLOGICAL INDICES OF VEGETATION PHOTOSYNTHESIS

With the phenological indices introduced above, seasonal photosynthetic patterns can be quantified and compared among different vegetation types (Table 1). Of course, these indices can change from year to year, in response to interannual climatic variability. Here we do not intend to study such variability, but rather we compare the values of these indices for different sites to illustrate their usefulness. For the three deciduous forests, the time when the growing season starts and terminates follows the order of their latitudes. However, this pattern does not hold when different functional types are considered. The Scots pine forest in Finland starts to grow almost as early as does the mixed deciduous forest in Tennessee, USA, although the former is a more northern site. The tallgrass prairie site in Oklahoma, USA and the mixed deciduous forest in Tennessee have similar latitudes. However, the growth of the tallgrass prairie is initiated eight days earlier and ends only three days earlier than the mixed deciduous forest in Tennessee. Clearly for different vegetation functional types, the growing season length is not an indicator of carbon assimilation capacity. The tallgrass prairie in Okalahoma has a longer growing season than the mixed deciduous forest in Tennessee. However, the latter has a higher growing season assimilation potential index than the former even through their effective daily maximum canopy photosynthetic rates are similar. In this regard, the effective length of the growing season depicts the difference between the two sites better. Although the prairie site has a longer growing season, its effective length of growing season is shorter than is the mixed deciduous

Chapter 7.2: Phenology of Vegetation Photosynthesis

481

forest in Tennessee. Nevertheless for the three deciduous forests, both the length of the growing season and the effective length of the growing season follow the same rank order. Table 7.2-1. Phenological indices of vegetation photosynthesis determined for the five sites under investigation Mixed Scots Tall Mixed Aspen Vegetation Pine grass forest forest Forest Forest Prairie (TN) (Harvard) Growing season initiation day (Di) 116 148 105 113 131 Growing season termination day (Dt) 303 276 308 311 299 187 128 203 198 168 Growing season length (L, days) 197 202 211 Growing season center (D ( c) 210 205 Growing season assimilation 4397.6 4760.3 3226.2 2290.5 2367.7 potential index (u, μmol m-2 s-1 day) Effective growing season length (L ( e, 123 136 119 124 96 days) Effective daily maximum canopy 35.8 35.2 27.1 18.5 24.7 photosynthetic rate ((Pe, μmol m-2 s-1) Photosynthesis development velocity 0.5999 0.8854 0.7259 0.2993 1.2801 (V Vs, μmol m-2 s-1 day-1) Photosynthesis recession velocity 0.3800 0.4786 0.5384 0.3271 0.7895 (V Vf, μmol m-2 s-1 day-1) 105 116 140 Photosynthesis upturn day (PUD) 133 149 Photosynthesis stabilization day 149 143 164 169 166 (PSD) 44 27 24 Spring linear period (days) 36 17 225 257 255 Photosynthesis downturn day (PDD) 255 261 296 307 288 Photosynthesis recession day (PRD) 293 275 71 50 33 Fall linear period (days) 38 14

The capacity of vegetation to assimilate CO2 over the growing season, as indicated by the growing season assimilation potential index (u), depends on not only the length of the growing season, but also how fast the vegetation can grow to its peak assimilation status and how long the vegetation can stay in its peak status. Comparing the Scots pine forest in Finland with the Aspen forest in Canada and Harvard Forest, the Scots pine forest has a smaller capacity to assimilate CO2 during the growing season than either the aspen forest or Harvard Forest. However, it has a longer growing season: 187 days vs. 128 days for the aspen forest and 168 days for Harvard Forest. Its seasonal carbon assimilation capacity is limited by its small effective daily max. canopy photosynthetic rate: 18.5 μmol/m2 s vs. 24.7 μmol/m2 s for the

482

Phenology: An Integrative Environmental Science

aspen forest and 27.1 μmol/m2 s for Harvard Forest. Such differences between forests dominated by deciduous trees and those with evergreen species may reflect different strategies for maximizing potential use of environmental conditions. It is interesting to note that while the length of the growing season as well as the days when the growing season starts or terminates differ a lot among different sites, the center of the growing seasons are quite close to each other among different sites. For example, the growing seasons start 43 days apart and end 32 days apart and their lengths of the growing season differ in 65 days, but their center of the growing season is only eight days apart. Of fundamental importance are the vegetation photosynthesis development velocity in the spring and recession velocity in the fall. They indicate how quickly the vegetation can respond to changes in climatic conditions. For different vegetation types, the two velocities tend to be linearly related to each other ( V f = 0.1484 + 0.4673Vs , R2 = 0.88, Figure 3), indicating that the rate of decline in the fall is proportional to the rate of development and expansion in the spring. In addition, vegetation photosynthesis tends to develop faster than it recesses. 0.9

Fall photosynthesis recession velocity (μmol m-2 s-1 day-1)

2

R = 0.88

Aspen forest

0.8

0.7

Mixed forest (Harvard Forest)

0.6

0.5

Mixed forest (Tennessee)

Scots pine forest

0.4

Tallgrass prairie 0.3

0.2 0.2

0.4

0.6

0.8

1.0

1.2

1.4

Spring photosynthesis development velocity (μmol m-2 s-1 day-1) Figure 7.2-3. Relationship between photosynthesis development velocity and photosynthesis recession velocity.

Chapter 7.2: Phenology of Vegetation Photosynthesis

7.

483

CONCLUSIONS

In this chapter, we expanded the traditional research area of plant phenology by characterizing plant community growth activities from vegetation photosynthesis. We introduced a systematic approach to objectively determine the various stages in the seasonal march of vegetation photosynthesis. We derived a suite of phenological indices that can be used to describe seasonal vegetation photosynthetic activities for a variety of vegetation types. These indices make it possible to quantify the differences and similarities between different vegetation types in their responses to changes in climatic conditions. Our study indicates that vegetation functional types may be considered as basic units in phenological studies. We found that the pace of photosynthesis development in the spring is related to the pace of photosynthesis recession in the fall, and development is in general faster than recession. The potential applications of these indices are tremendous. Studies indicate that global warming has been lengthening the growing season on global scales (e.g., Peñuelas and Filella 2001, Myneni et al. 1997) and variations in biospheric growth activities have been used to explain the interannual variability in atmospheric CO2 concentration (Keeling et al. 1996). In these studies, biospheric growth activities are generally defined by changes in plant organs or vegetation greenness conditions. As we have demonstrated here, a longer growing season does not necessarily means larger CO2 assimilation by vegetation. How fast plants reach their peak CO2 assimilation potentials after released from winter dormancy is also important. If global warming reduces the velocity of spring photosynthesis development, the enhancement of annual carbon assimilation due to a lengthening growing season may be compromised. If the spring photosynthesis development velocity is reduced too much, then annual carbon assimilation can be adversely affected. In this case, the effective length of the growing season would provide a better measure for the effects of global warming on plant growth activities. Future phenological research activities for the canopy photosynthetic indices proposed in this chapter would be to study how variable climatic conditions affect the expression of the indices. This would allow us to develop better models to predict impacts of global climate change on vegetation activities. Also there is a need to establish relationships between phenology of vegetation photosynthesis and traditional plant phenology. With such relationships, we can use the historical data of plant phenological observations to examine how vegetation photosynthesis has changed in the past.

484

Phenology: An Integrative Environmental Science

This chapter essentially defines the growing season in the context of plant community CO2 assimilation. Traditionally, growing seasons have been determined based on structural changes in plants (bud break, leaf-out, leaf fall, etc.). These two approaches might not always agree with each other as environmental factors such as frost, drought, insect activities, etc. affect plant photosynthesis. Clearly, however, to explain photosynthetic patterns as characterized by the indices developed in this chapter, it is necessary to link them with dynamics in canopy structures such as leaf area development. This is an issue that future studies should also address.

ACKNOWLEDGEMENTS This study was initially conceived when L. Gu was working for Fluxnet at University of California at Berkeley. The bulk of the research was carried out at Oak Ridge National Laboratory (ORNL) with support from the U.S. Department of Energy, Office of Science, Biological and Environmental Research Program, Environmental Science Division, Terrestrial Carbon Program. We would like to thank Peter Curtis, Paul Hansen, Stan Wullschleger and Rich Norby for their comments on a draft of this chapter. ORNL is managed by UT-Battelle, LLC, for the U.S. Department of Energy under contract DE-AC05-00OR22725.

REFERENCES CITED Aubinet, M., A. Grelle, A. Ibrom, Ü. Rannik, J. Moncrieff, T. Foken, A. S. Kowalski, P. H. Martin, P. Berbigier, C. Bernhofer, R. Clement, J. Elbers, A. Granier, T. Grünwald, K. Morgenstern, K. Pilegaard, C. Rebmann, W. Snijders, R. Valentini, and T. Vesala, Estimates of the annual net carbon and water exchanges of forests: The EUROFLUX methodology, Advances in Ecological Research, 30, 114-175, 2000. Baldocchi, D. D., E. Falge, L. Gu, R. Olson, D. Hollinger, S. Running, P. Anthoni, C. Bernhofer, K. Davis, J. Fuentes, A. Goldstein, G. Katul, B. Law, X. Lee, Y. Malhi, T. P. Meyers, J. W. Munger, W. Oechel, K. Pilegaard, H. P. Schmid, R. Valentini, S. Verma, T. Vesala, K. Wilson, and S. Wofsy, FLUXNET: A new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor and energy flux densities, Bull. Amer. Meteorol. Soc., 82, 2415-2435, 2001. Baldocchi, D. D., B. B. Hicks, and T. P. Meyers, Measuring biosphere-atmosphere exchanges of biologically related gases with micrometeorological methods, Ecology, 69, 1331-1340, 1988. Bradley, N. L., A. C. Leopold, J. Ross, and W. Huffaker, Phenological changes reflect climate change in Wisconsin, Proc. Natl. Acad. Sci. (USA), 96, 9701-9704, 1999. Chmielewski, F. -M, The International Phenological Gardens across Europe: Present state and perspective, Phenol. Seasonality, 1, 19-23, 1996.

Chapter 7.2: Phenology of Vegetation Photosynthesis

485

Fitzjarrald, D. R., O. C. Acevedo, and K. E. Moore, Climatic consequences of leaf presence in the eastern United States, J. Climate, 14, 598-614, 2001. Freedman, J. M., D. R. Fitzjarrald, K. E. Moore, and R. K. Sakai, Boundary-layer clouds and vegetation-atmosphere feedbacks, J. Climate, 14, 180 – 197, 2001. Goulden, M. L., J. W. Munger, S. –M. Fan, and B. C. Daube, Measurements of carbon sequestration by long-term eddy covariance: methods and a critical evaluation of the accuracy, Global Change Biology, 2, 169-182, 1996. Gu, L., D. D. Baldocchi, S. B. Verma, T. A. Black, T. Vesala, E. M. Falge, and P. R. Dowty, Advantages of diffuse radiation for terrestrial ecosystem productivity, J. Geophysical Research, 107(D6), DOI 10.1029/2001JD001242, 2002. Keeling, C. D., J. F. S. Chin, and T. P. Whorf, Increased activities of northern vegetation inferred from atmospheric CO2 measurements, Nature, 382, 146-149, 1996. Lieth, H. (editor), Phenology and Seasonality Modeling, Spring-Verlag, New York, 444 pp., 1974. Menzel, A., and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999. Mott, J. J., and A. J. Mccomb, Role of photoperiod and temperature in controlling phenology of 3 annual species from an arid region of western-Australia, J. Ecology, 63, 633-641, 1975. Myneni, R. B., C. D. Keeling, C. J. Tucker, G. Asrar, and R. R. Nemani, Increased plant growth in the northern high latitudes from 1981 to 1991, Nature, 386, 698-702, 1997. Peñuelas, J., and I. Filella, Responses to a warming world, Science, 294, 793-795, 2001. Podolsky, A. S., New Phenology: Elements of Mathematical Forecasting in Ecology, John Wiley & Sons, New York, 504 pp., 1984. Schwartz, M. D., and T. M. Crawford, Detecting energy balance modifications at the onset of spring, Phys. Geography, 22, 394-409, 2001. Schwartz, M. D., Monitoring global change with phenology: The case of the spring green wave, Int. J. Biometeorol., 38, 18-22, 1994. Spano, D., C. Cesaraccio, P. Duce, and R. L. Snyder, Phenological stages of natural species and their use as climate indicators, Int. J. Biometeorol., 42, 124-133, 1999. Suárez-López, P., K. Wheatley, F. Robson, H. Onouchi, F. Valverde, and G. Coupland, CONSTANS mediates between the circadian clock and the control of flowering in Arabidopsis, Nature, 410, 1116-1120, 2001. Thomas, G. B., and R. L. Finney, R. L., Calculus and Analytic Geometry, 1139 pp., AddisonWesley Publishing Company, Reading, Massachusetts, 1998. Yanovsky, M. J., and S. A. Kay, Molecular basis of seasonal time measurement in Arabidopsis. Nature, 419, 308-312, 2002.

Chapter 7.3 RADIATION MEASUREMENTS Phenological Effects Jie Song Department of Geography, Northern Illinois University, Dekalb, IL, USA

Key words:

1.

Radiation, Albedo, Canopy, Satellite measurements, Algorithms

INTRODUCTION

Surface albedo, the ratio of the upwelling solar irradiance to the downwelling irradiance at the surface of the Earth, is a significant indicator of environmental conditions and a key parameter for calculation of the radiation budgets of the atmosphere and surface. By knowing the surface albedo (α), one can estimate the amount of heat produced at the surface by the absorption of incident solar radiation. The albedo has strong temporal and spatial variations that influence the regional and global surface energy budget. For a vegetated surface, α can change with plant phenology, soil moisture content, and fractional canopy cover, as well as solar zenith angle. Uncertainties in α translate almost directly into errors in the calculation of net radiation and energy fluxes and have been investigated in many studies. In a review of studies on the sensitivity of climate models to surface albedo changes, Henderson-Sellers and Wilson (1983) concluded that an accuracy of ±0.05 in α was needed for climate modeling purposes. To evaluate in detail the variations in α on diurnal and seasonal scales, ground-based measurements are usually essential. In principle, factors internal to the vegetative canopy should be examined separately from external factors, but albedo measurements made above vegetation usually do not distinguish between canopy and noncanopy effects. Hence, the role of noncanopy Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 487-503 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

488

Phenology: An Integrative Environmental Science

surfaces often has to be inferred, which requires careful analysis. In this chapter, the diurnal asymmetry in α is investigated using field observations above the canopy. The contributions of the phenological effects on α excluding the influence of solar zenith angle are examined. In addition, phenological effects on the conversion of narrowband satellite data to the broadband albedo α are explored.

2.

DIURNAL VARIATION IN SURFACE ALBEDO

In most radiation transfer models, the diurnal variation of α is assumed to be nearly parabolic and symmetric about solar noon. Many vegetated surfaces during the late growing phase, however, do not display such symmetry, even for uniform vegetation in flat terrain. For example, morning-to-afternoon albedo differences have been observed to be as high as 0.1 at the Cloud Atmospheric Radiation Testbed site in northern Oklahoma in the fall (Minnis et al. 1997). At Matador, Canada, values of α were found to be considerably larger during the afternoon than during the morning in the fall over grassland (Ripley and Redmann 1976). Ripley and Redmann attributed this phenomenon to the alignment of the vegetation in response to the strong prevailing west-northwest winds. An eastward alignment of the vegetation could be seen clearly in a photograph presented in their paper. In another case, Song (1995) found larger values of α during the morning (rather than in the afternoon) in an ungrazed tallgrass prairie were associated with strong prevailing winds from the southeast, at the first International Satellite Land Surface Climatology Project (ISLSCP) field experiment (FIFE) conducted in northern Kansas. The amounts of diurnal albedo asymmetry were less obvious for bare and grazed sites at FIFE.

2.1

Field Investigation in Asymmetric Albedo

Radiation and wind data from the Surface Radiation Budget Network (SURFRAD) can be examined to study the link between albedo asymmetry and wind (e.g., Song 1998). At each SURFRAD station, upwelling and downwelling solar and infrared irradiances are measured together with wind direction and speed (Hicks et al. 1996). Song (1998) studied observations made from early April to the end of October in 1997 in natural grassland near Bondville, Illinois. This site was chosen for examination because the grass at the station was not mowed, which allowed the grass to reach sufficiently tall (1.0-1.3 m) to be moved about by the ambient wind. Diurnal variations in α on cloudless days were found to be asymmetric, with lower values in the morning than in the afternoon, whenever there were strong

Chapter 7.3: Radiation Measurements

489

west or southwest winds (greater than 6 m/s at 10-m height) and almost symmetric whenever the winds were weak (less than 3 m/s at 10-m height). The maximum difference on average was found to be 0.04. Findings for the winds from the east were very limited because so few cloudless days occurred with east winds. To further illustrate the influence of wind on albedo variation, field observations taken during cloudless summer days of 1997 above winter wheat and corn (maize) fields near DeKalb, Illinois, can be utilized (Song 1998). The observations included incident and reflected solar irradiances measured using an albedometer consisting of two pyranometers; canopy reclination angles measured with a protractor between plant stems and the vertical; and wind speed and direction were observed at a 10-m height at a nearby weather station. Significant diurnal asymmetries occurred in α on four of 12 cloudless days when the winds were moderate to strong and wind directions were persistent. Representative examples can be found in the data for 9 June in the wheat field and 16 July in the cornfield (Figure 1).

Figure 7.3-1. Variation of surface albedo with incident solar radiation for two days in 1997. The arrows indicate increasing time. Reprinted from Song (1998, Figure 1, p. 183), © Elsevier Science, Oxford, UK, used with permission.

On 9 June, the observed albedos in the wheat field were systematically higher in the morning than in the afternoon, and the minimum α occurred about an hour after solar noon. Wind directions in early June were primarily from the southeast and northeast, due to the influence of low-pressure systems south and southeast of the study area. The leaves and stems of the wheat canopy on 9 June were observed to lean or recline about 10° toward the southwest. Dew was light in the early morning and had evaporated

490

Phenology: An Integrative Environmental Science

almost completely by the time measurements started. If the diurnal variation in α had been influenced only by morning dew, the α would have been symmetric over the rest of the day, and the minimum α would have occurred at solar noon. Therefore, the optical effects of dew cannot explain the systematic morning-afternoon albedo difference, and it is likely that an evident westward tilt of the canopy was the major cause. For the cornfield on 16 July, the observed albedos were consistently lower in the morning than in the afternoon, with an average difference of about 0.01, and the minimum albedo occurred at a little less than an hour before solar noon (Figure 1). Wind directions were predominately from the west on July 16 and the previous 2 days due to a high-pressure system located to the south of the study area. The bulk of the corn canopy leaves were observed to be tilted about 5° toward the east. Dew was not observed during any of the measurements above corn. The eastward-inclined canopy appears to be the sole explanation for the albedo asymmetry.

2.2

Asymmetric Albedo–Reclined Canopy

Many studies assume that the trunks and stems are normal to the surface and the leaves are distributed symmetrically along these vertical columns. For the overall, bulk canopy, the leaves are assumed to be randomly arranged with regard to azimuth. The solar angle relative to the canopy is often assumed to be the same as the solar zenith angle relative to a horizontal surface. However, if the prevailing wind forces the canopy to recline to the west (or to the east), the solar angle relative to the reclined canopy, will be larger (or smaller) in the morning than in the afternoon. If the wind forces the canopy to recline to the north (or to the south), the solar angle relative to the canopy would remain symmetric around the local vertical, but it would be larger (or smaller) than for an unreclined canopy. To develop parameterizations that model the interactions of solar radiation with vegetative canopies, a correct interpretation of the solar angle is needed. Z that is used to obtain the interception Accordingly, the solar zenith angle (Z) parameter in the canopy radiative transfer models should be replaced by Zƍ, the angle between the incident solar rays and a line perpendicular to the bulk canopy leaf surfaces (Figure 2). According to spherical trigonometry, Zƍƍ is related to Z, the solar azimuth (a), the canopy reclination angle (i), and the canopy azimuth (aƍ) ƍ by the following equation:

cos Z ′ = cos Z cos i + sin Z sin i cos(a − a ′ )

(1)

Both a and aƍƍ are considered negative when directed east of south and positive when directed west of south. The canopy azimuth (aƍ) ƍ is expected to

Chapter 7.3: Radiation Measurements

491

be determined by the wind direction. The value of aƍƍ is the wind direction when the wind direction is less than 180° and is 180° minus the wind direction when the wind direction is larger than 180°. The canopy reclination angle (i) is expected to depend on the wind speed, canopy type, and vegetation phenology. Senesced canopies are particularly susceptible to wind-blown reclination. Song (1998) illustrated the behavior of the diurnal albedo asymmetry simulated with analytical expressions for various reclination orientations and values of the Leaf Area Index (LAI).

Figure 7.3-2. Relation of the angle Z' between the direct-beam solar radiation and canopy direction for a reclined canopy. Reprinted from Song (1998, Figure 3, p. 186), © Elsevier Science, Oxford, UK, used with permission, and modified after Sellers (1965).

3.

SEASONAL VARIATION IN SURFACE ALBEDO

The albedo of vegetated surfaces is subject not only to diurnal and seasonal changes in Z but also to temporal variability due to differences in the phenology of various species. Isolation of the role of phenology from the effects of changes in Z is complicated when the temporal and geographical coverage of observations is extensive. For example, Taylor and Stowe (1984) used NIMBUS-7 satellite data to generate tables illustrating the variation in α with Z for a limited number of land surface types. Their work showed that α was strongly affected by variations in surface cover as well as by changes in Z. To study the seasonal variation of α caused by phenological variations, the contributions of changes in Z should be excluded.

492

Phenology: An Integrative Environmental Science

Ripley and Redmann (1976) tried to make adjustments to exclude the effects of changes in Z on seasonal variations in α for grassland located near Matador, Canada. Mean daily values of the α were plotted against daily minimum solar zenith angles for 24 days during the 1971 growing season. A straight line was fitted to the points and used to adjust the α values for the seasonal variation in Z. Values with this variation removed were plotted against the date from May to September. They noticed a gradual decrease in the α during the season and the slope of the scatter plot was -0.00021 per day. The assumption behind their adjustment was that plant biophysical characters did not change during the 24-day period, which might not be realistic, especially during the growing season. Thus the inferred changes in the mean daily α during these 24 days might have included the effect of changes in solar zenith angle as well as in phenology. Another difficulty is that the minimum values of the diurnal solar zenith angle within the 24-day period (used to adjust the α estimates) were rather limited relative to the range of solar zenith angles during the summer. Without adjustment for the influence of Z, Gutman et al. (1989) derived the seasonal variation of albedos from data collected with advanced very high resolution radiometers (AVHRRs) on National Oceanic and Atmospheric Administration (NOAA) satellites. The derived α on clear days when the NOAA-9 satellite passed over the sites between 1400 h and 1500 h local standard time (LST) were used to represent the mean diurnal value of α. Gutman et al. (1989) found that α for wheat cropland increased from early spring (March) to early summer (May), when vegetative greenness reached a maximum, and decreased thereafter. The seasonal variations in α caused by phenological changes, however, are not distinguished from those caused by seasonal changes in Z between 1400 h and 1500 h LST. To make α measurements useful in detecting phenological differences, one should compare albedos at the same value of Z.

3.1

Albedo Observations

Surface albedos were measured in 1987 during the FIFE at several locations in the Konza Prairie, which is located a few kilometers south of Manhattan, Kansas (Sellers et al. 1988). Meteorological and biophysical variables were observed during several phases of the development and the decline of the vegetation (Strebel et al. 1994). Two ungrazed and unburned tallgrass prairie sites in the northwest corner of the 15-km by 15-km area of the Konza Prairie were selected for this study. Station 5 (sitegrid ID = 2123), one of the Super Portable Automated Mesonet Stations, was located at the top of a ridge, while station 2 (sitegrid ID = 1916), one of the Bowen Ratio

Chapter 7.3: Radiation Measurements

493

surface flux observation stations, was located in a valley. Eppley precision spectral pyranometers were mounted over the sites to measure incoming (upright position) and reflected (inverted position) solar irradiances in the 0.285-2.800 μm spectral region. Figure 3a shows the diurnal variation in α from spring to winter at station 5. Solar noon always occurred after 1200 h LST. The values of α decreased from spring (day of year 121) to summer (day 251) and increased from summer to winter (day 345). However, these trends in α do not necessarily represent phenological changes because daily minimum Z ranged from 15.5° to 62° throughout the seasons.

Figure 7.3-3. Observed diurnal variations in albedo at (a) station 5 in the Konza Prairie grassland (Kansas) during 1987, (b) a maize field in northern Illinois, and (c) a winter wheat field in northern Illinois during 1997. Reprinted from Song (1999, Figures 1 and 2, p. 154), © Springer-Verlag, Heidelberg, Germany, used with permission.

The 1997 field observations noted previously for corn (maize) and winter wheat fields in northern Illinois produced data useful for examining relationships between α variation and crop phenology. An albedometer, a ceptometer, and a hand-held multispectral radiometer were used to measure

494

Phenology: An Integrative Environmental Science

α, LAI, and normalized difference vegetation index (NDVI), respectively. Measurements in the maize field (Figure 3b) started on 23 June (day of year 173), when the maize was 30 cm tall and 80% of the soil surface was exposed, and ended on 5 July (day 197) just before pollination. The albedo increased during the growing stage as the foliage gradually covered the ground. Measurement in the wheat field (Figure 3c) started on 9 June (day 160) during peak greenness, and ended on 15 July (day 196), after senescence and just before harvesting. The albedos showed a decreasing trend from the peak green stage to ripening in the wheat field. Again, the observed α variation is partly due to the solar zenith angle variation with day of the year. Seasonal variations in α caused by the phenological changes should be studied at fixed solar zenith angles.

3.2

Obtaining Albedo at Constant Z

In the evaluation of α for fixed solar zenith angles, three types of fields are examined here. For ungrazed, unburned grassland at the Konza Prairie, the relationships between α and Z are displayed in Figure 4a for days of year 157 (late spring), 187 (early summer), 251 (autumn), and 315 (early winter) at station 5 of FIFE. A curve showing the variation of α for bare soil (Idso et al. 1975) is shown in the lower portion of the figure. As is typical, α for the grassland increased sharply with Z as a result of rapid increases in leaf spectral reflection (e.g., Monteith and Unsworth 1990). The similarity of the variation in α on different days addressed in Figure 4a might be due to the canopy being sufficiently full to cover the soil surface for all the days shown. Figures 4b and 4c show values of α measured in the corn (maize) and winter wheat fields. The rate of change in α versus Z for young maize plants (LAI < 3.5) was almost the same as that for bare soil (Figure 4b), but is steeper for mature maize plants (LAI = 5). In the early stage of crop growth, a large fraction of soil was exposed directly to direct-beam solar radiation. At the peak stage, the foliage covered most of the ground, and α increased sharply with Z as a result of greater leaf specular reflection. For wheat (Figure 4c) during peak green to senescent stages, α increased faster with Z than that for bare soil; very little of the soil surface under the wheat canopy was visible. The value of Z = 20° seems to be a reasonable reference value at which changes in α due to phenological variations can be observed in the midlatitudes during the summer growing season. Such a relatively small value is desirable to minimize the effects of azimuthal asymmetries associated with canopy structure and also potential inaccuracies associated with leveling of sensors used to measure α. In the spring, fall and winter, however, minimum Z is greater than 20°, and the value of the α at a reference of 20° must be

Chapter 7.3: Radiation Measurements

495

inferred. One method is to extrapolate changes in α to Z = 20°, by pairing one day with a wide range of Z values (including 20°) with another day that had minimum Z greater than 20° but otherwise had a rate of change in α versus Z that was nearly identical to the rate for the first day. This approach assumes that the variation in α within each of the two days is dominated by

Figure 7.3-4. Variations of surface albedo with solar zenith angle at (a) station 5 in the Konza Prairie grassland, (b) a maize field, and (c) a winter wheat field. Reprinted from Song (1999, Figure 3, p. 155), © Springer-Verlag, Heidelberg, Germany, used with permission.

changes in Z, with negligible phenological effects. A best-fit polynomial can be generated and applied as follows to implement this approach:

α (Zo ) =

1 n ∑ {α (Zi ) − [ f (ZZi ) − f (ZZo )]} n i =1

(2)

Here Z0 = 20°, n is the number of α observations on a day with minimum Z larger than 20°, α(Zi) is the albedo observed at Zi, and f(Z) Z is the polynomial function for albedo. The difference [f(Z Zi) - f(Z Z0)] represents the amount of albedo adjustment relative to Z0 at each Z. As already noted, the above equation can be applied only to pairs of days when the rates of change in α versus Z appear to be nearly identical. For example, α adjustment for

496

Phenology: An Integrative Environmental Science

day 157 was assumed to be sufficiently representative to use in Eq. (2) for days 251 and 315.

3.3

Albedo Variation Caused by Phenology

Seasonal variations in α viewed at a constant Z can be attributed mostly to phenological changes associated with the stage of plant growth and environmental conditions. During 1987 FIFE, climate at the Konza Prairie is near normal. At the two Konza prairie grassland sites (stations 5 and 2), α for Z = 20° decreased almost linearly, for example, from 0.21 (day 121, spring) to 0.15 (day 345, early winter) at station 5 (Figure 5) even though the minimum α each day increased in the later season as shown in Figure 3. The reason is that in the prairie field, soil is covered densely by senesced grass of the previous year in early spring, which tends to have higher surface albedo because of lighter color and horizontal inclination of the dried grass. During the spring growing season, higher temperature and abundant moisture promote decomposition of the senesced grass, which tends to lower the surface albedo with day of the year. Despite greening leaves of the prairie grass increasing reflectance in the near-IR, the vertical growth of the stalks tend to trap the incident solar radiation, thus decreases the albedo as well. During the summer, prairie grass is in heat stress due to higher potential evapotranspiration demand, which tends to decrease the reflectance of the near-IR, and thus to decrease surface albedo further. As a result, the general trend of surface albedo at a prairie grassland shows a negative slope with day of the year (Figure 5).

Figure 7.3-5. Seasonal variations of albedo at 20° zenith angle in the Konza Prairie grassland during 1987. Reprinted from Song (1999, Figure 4, p. 155), © Springer-Verlag, Heidelberg, Germany, used with permission.

Chapter 7.3: Radiation Measurements

497

The (decreasing) rate of -0.00026 per day estimated from Figure 5 is steeper than the slope (-0.00021 per day) derived from Ripley and Redmann (1976). The less steep slope by Ripley and Redmann could be due to the partial removal by their correction procedure for the seasonal variation in α with Z. In addition, soil water content can influence the progression of vegetation phenology. Relatively wetter fall conditions in 1987 may contribute to the lower albedo due to darker soil and delay in the dry-down and senescence stage. More water content in vegetation can further lower surface albedo due to the strong absorption of near-IR by water. The albedo at station 2 (Figure 5) was smaller than the albedo at station 5 but had about the same rate of change with time. Station 2 was located at the bottom of a valley where the soil was darker and the grass was denser than at station 5, which was on a ridge. The albedo for Z = 20° in the corn (maize) field in northern Illinois (Figure 6a) increased from green-up (LAI = 1.8) to peak green (LAI = 3.5) stages and decreased slightly just before pollination (LAI = 5). Although the reflectance of visible (red) radiation varied little, the reflectance of nearinfrared radiation dropped significantly during an ongoing local drought, which caused α to decrease slightly before pollination. The total precipitation for June and July 1997 in the area was about half of the climatic average. In the winter wheat field, α decreased from the peak green stage to harvesting stage (Figure 6b), as soil moisture content decreased and the plant reached senescence in the late spring. The wheat leaves gradually lost turgidity, which caused a consistent decrease of reflectance in the nearinfrared reflectance. In addition, the leaf shape and distribution angle changed so that more soil was exposed. In crop fields after harvesting, the soil surface is usually dark and exposed due to less residue cover. During the growing season, LAI increases as leaves gradually cover the soil surface. As vegetative surface reflects more radiation than soil surface, thus, surface albedo increases as LAI increases, and vice versa (Figure 6). The results of this work show that α for vegetated surfaces is influenced strongly by both phenology and solar zenith angle. The variations in α caused by phenology should be examined at a chosen reference solar zenith angle. The relationship between α and surface phenology is important in environmental monitoring as well as in land surface parameterizations used for estimating the atmospheric radiation budget in climate models. By using methods of analysis that reduce the confounding effects of variations in Z on estimating α, more accurate and comprehensive evaluations of phenological influences on α are possible.

498

4.

Phenology: An Integrative Environmental Science

SATELLITE-BASED ALBEDO MEASUREMENT

Spatial variations in α depend primarily on reflectances in the visible and near-infrared portions of the solar spectrum. Various methods have been developed for estimating spatially representative values of α from highresolution remote sensing measurements, such as those made with AVHRRs on satellites. Satellite data have the potential to provide high-resolution but area-representative estimates of the α, provided that atmospheric effects on remote sensing measurements for individual narrow bands can be corrected with reasonable confidence and a reliable method for converting narrowband

Figure 7.3-6. V Variations of albedo, red and near-infrared reflectances at 20° zenith angle, and leaf area index (LAI) during the growing season in (a) maize field and (b) winter wheat field during 1997. Reprinted from Song (1999, Figure 5, p. 156), © Springer-Verlag, Heidelberg, Germany, used with permission.

reflectances to (intrinsically broadband) α values can be established (e.g., Pinker 1985). However, uncertainties can be introduced in the conversion from satellite-derived narrowband reflectances to α, because coefficients associated with the narrowband-to-broadband (NTB) conversion depend on land cover and plant phenology. A simulation study by Vulis and Cess (1989) indicated that NTB conversion coefficients are larger for vegetated surfaces than for ocean or desert. Thus, a reliable algorithm for NTB conversion suitable for major land-use types should include the description of vegetative conditions. Regardless, constant NTB conversion coefficients have been suggested in many studies for the use of data from AVHRR channels 1 and 2, as summarized in Table 1. The new algorithm should have the NTB conversion coefficients expressed as functions of NDVI, because the normalized difference vegetation index (NDVI) is a good

Chapter 7.3: Radiation Measurements

499

quantitative indicator for the land cover and vegetative conditions and is easily calculated from reflectances in AVHRR channels 1 and 2. Table 7.3-1. Summary of empirical coefficients (c1, c2, and d) used to estimate surface albedo (α) from AVHRR data of channels 1 and 2, by using the empirical function of α = c1 ρ1 + c2 ρ2 + d Source c1 c2 d Brest and Goward (1987) 0.526 0.418 (veg.) / 0.474 (soil) 0 He et al. (1987) 0.332 0.678 0 0.5 0.5 0 Saunders (1990) Potdar and Narayana (1993) 0.798 0.188 0.051 Valiente et al. (1995) 0.545 0.320 0.035 Russell et al. (1997) 0.441 0.670 0.044

4.1

Development of a New Algorithm

The surface albedo can be expressed as λ

α=

∫λ12 ρ (λ ) S (λ ) dλ λ

∫λ12 S (λ )dλ

(3)

where α(λ) and S((λ) represent the spherical reflectance and the downwelling solar irradiance, respectively, at the wavelength λ. Wavelength integration from λ1 = 0.3 μm to λ2 = 2.8 μm covers most of the solar spectrum. To use the reflectances within a few narrow wavebands observed by AVHRR channel 1 (λ = 0.58-0.68 μm) and channel 2 (λ = 0.725-1.10 μm) to estimate the total reflectance within the whole solar spectrum, an improved understanding is required of the relationship between the reflection characteristics of the narrow bands and that of the broad band for various surface types. In this study, the solar spectrum is divided into two subregions: the VIS subregion (0.3-0.685 μm) and the NIR subregion (0.685-2.8 μm). The reflectance relationships between AVHRR channel 1 and the VIS subregion and between AVHRR channel 2 and the NIR subregion are examined. The wavelength of 0.685 μm was chosen to separate VIS and NIR, to be consistent with the wave band arrangement used by the multispectral (narrowband) instrument that provided the data for this analysis. With the definition of the VIS and NIR spectral subregions, Eq. (3) can be converted into the following discrete form, in which the first term in the numerator is the reflected radiation for the VIS subregions, and the second term is the reflected radiation for the NIR subregion:

500

Phenology: An Integrative Environmental Science m′

∑ ρi Si Δ λi +

α=

i=1

m

m

∑ ρi Si Δλi

i= m ′ +1

∑ Si Δλ i

i=1

(4)

Here the VIS and NIR spectral subregions are divided at the waveband m'' for the numerical integration, and m denotes the total number of wavebands used in the numerical integration. It is assumed that the VIS and NIR reflected radiation represented by each term in the numerator of Eq. (4) can be parameterized by a corresponding narrowband reflected radiance accompanied by appropriate conversion factors f1 and f2, respectively, in the following form: α = ρ v f1 −1

S v Δλ v m

∑ S i Δλ i

+ ρ n f 2 −1

i =1

S n Δλ n m

∑ S i Δλ i

i =1

(5)

Here the subscripts v and n represent the narrow wavebands in the VIS and NIR channels, respectively. Now the problem of estimating α from ρv and ρn yields to determining the new conversion factors f1 and f2 as well as the fractions of the downwelling solar radiation within the specified narrow bands, VIS and NIR, respectively. These fractions are independent of surface conditions and were estimated to be 0.1464 and 0.1050, respectively, according to the solar radiation spectrum calculated by using the LOWTRAN 7 model. Comparing Eq. (5) with Eq. (4) indicates that the values off f1 and f2 are determined by the following equations: f1 =

ρ v S v Δλ v m'

∑ ρ i S i Δλi

i =1

,

f2 =

ρ n S n Δλ n m

∑ ρ i S i Δλi

i = m ' +1

(6)

The values of f1 and f2 represent the ratio of the reflected radiances within the narrow bands to the reflected radiances for the VIS and NIR broad bands, respectively, or the degree of the representativeness of the narrowband reflectance for the broadband reflectance. Analyses by Song and Gao (1999) based on simultaneous measurements of both narrowband reflectances and of α over many types of surfaces reveal that NTB conversion factors change with surface conditions, particularly the fractional coverage and phenology of vegetation. The relationships between the two NTB conversion factors, f1 and f2, and NDVI were established from

Chapter 7.3: Radiation Measurements

501

analysis of data taken over various surfaces that are described in their study. The f1-NDVI and f2-NDVI relationships developed were then employed to infer the NTB conversion coefficients in Table 1 for the new algorithm (with d=0). c1 = 0.494 NDVI 2 − 0.329 NDVI + 0.372,

(7a)

c 2 = −1.439 NDVI 2 + 1.209 NDVI + 0.587.

(7b)

Using the new algorithm with NTB conversion coefficients varying with NDVI, Song and Gao (1999) found that α calculated from reflectances measured at AVHRR channels 1 and 2 show a better agreement with in situ and aircraft measurements than using the other empirical methods.

5.

PHENOLOGICAL IMPLICATIONS

Understanding the influence of phenological characteristics on surface radiative properties requires detailed field measurements under a variety of environmental conditions. This chapter emphasizes the use of observations made in situ in the field and remotely from satellites to study the effect of vegetation phenology on broadband surface albedo α. This work demonstrated that canopy reclination due to prevailing winds can cause diurnal asymmetries in α, especially during senescence. Results of analyses indicated that the influence of phenology on seasonal α variations should be studied under a constant solar zenith angle. Methods were summarized to evaluate α from the narrowband reflectances in VIS and NIR. It is found from multiple field samples that narrowband to broadband conversion coefficients are associated with vegetation cover and phenology, and can be expressed as functions of NDVI, which can be estimated from satellite remote sensing data and is commonly used to quantify the phenological progression of vegetative properties. The effect of phenological variation in trees on albedo has not been studied yet. It is expected that albedo over trees during spring greening-up will initially increase due to increased reflectance in the near-IR. During peak green stage, albedo may level off with increasing LAI because the multiple reflection of the canopy leaves tends to lower the albedo, which partially cancels the higher near-IR reflectance effect. During dry-down, albedo may decrease with time because lower reflectance in the near-IR will compound the multiple reflection effect.

502

Phenology: An Integrative Environmental Science

ACKNOWLEDGEMENTS This work was supported in part by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research, Water Cycle Pilot Study. Lenny Walther of the Northern Illinois University cartography laboratory provided valuable assistance with the figures. The author is especially in debt to Dr. Marvin L. Wesely, who had helped in revising the manuscript before his tragic departure.

REFERENCES CITED Brest, C. L., and S. N. Goward, Deriving surface albedo measurements from narrow band satellite data, Int. J. Remote Sensing, 8, 351-367, 1987. Gutman, G., G. Ohring, D. Tarpley, and R. Ambroziak, Albedo of the US Great Plains as determined from NOAA-9 AVHRR data, J. Climate, 2, 608-617, 1989. He, C., K. Kittleson, and J. Bartholic, Evapotranspiration monitored from satellite as an indication of shift and impact of vegetation change, Preprints of the Twenty-First International Symposium on Remote Sensing of Environment, Ann Arbor, MI, 1987. Henderson-Sellers, A., and M. F. Wilson, Surface albedo for climatic modeling, Reviews of Geophysics and Space Physics, 21, 1743-1778, 1983. Hicks, B. B., J. J. DeLuisi, and D. R. Matt, The NOAA Integrated Surface Irradiance Study (ISIS)–A new surface radiation monitoring program, Bull. Amer. Meteorol. Soc., 77, 28572864, 1996. Idso, S. B., R. D. Jackson, R. J. Reginato, B. A. Kimball, and F. S. Nakayama, The dependence of bare soil albedo on soil water content, J. Appl. Meteorol., 14, 109-113, 1975. Minnis, P., S. Mayor, W. L. Smithn Jr., and D. F. Young, Asymmetry in the diurnal variation of surface albedo, IEEE Transactions on Geoscience and Remote Sensing, 35, 879-891, 1997. Monteith, J. L., and M. H. Unsworth, Principles of Environmental Physics, 2nd ed., Arnold, New York, 291 pp., 1990. Pinker, R. T. Determination of surface albedo from satellites, Advanced Space Research, 5, 333-343, 1985. Potdar, M. B., and A. Narayana, Determining short-wave planetary albedo from spectral signatures of land-ocean features and albedo mapping using NOAA AVHRR data, Acta Astronautica, 29, 687-690, 1993. Ripley, E. A., and R. E. Redmann, Grassland, in Vegetation and the Atmosphere, 2, Case Studies, edited by J. L. Monteith, pp. 351-398, Academic Press, London, 1976. Russell, M. J., M. Nunez, M. A. Chladil, J. A. Valiente, and E. Lopez-Baeza, Conversion of nadir, narrowband reflectance in red and near-infrared channel to hemispherical surface albedo, Remote Sensing Environ., 61, 16-23, 1997. Saunders, R. W., The determination of broad band surface albedo from AVHRR visible and near-infrared radiances, Int. J. Remote Sensing, 11, 49-67, 1990. Sellers, W. D., Physical Climatology, The University of Chicago Press, Chicago, 272 pp., 1965.

Chapter 7.3: Radiation Measurements

503

Sellers, P. J., F. G. Hall, G. Asrar, D. E. Strebel, and R. E. Murphy, The first ISLSCP field experiment (FIFE), Bull. Amer. Meteorol. Soc., 69, 22-27, 1988. Song, J., Influence of heterogeneous land surfaces on the surface energy budget at meso- and large scales, Ph.D. Dissertation, University of Delaware, Newark, 1995. Song, J., Diurnal asymmetry in surface albedo, Agricult. Forest Meteorol., 92, 181-189, 1998. Song, J., Phenological Influences on the albedo of prairie grassland and crop fields, Int. J. Biometeorol., 42, 153-157, 1999. Song, J., and W. Gao, An Improved method to derive surface albedo from narrowband AVHRR satellite data: Narrowband to broadband conversion, J. Appl. Meteorol., 38, 239249, 1999. Strebel, D. E., D. R. Landis, K. F. Huemmrich, and B. W. Meeson, Collected data of the first ISLSCP field experiment, Vol. 1, Surface observations and non-image data sets, CD-ROM series, NASA Goddard Space Flight Center, 1994 [Data also available online at the following address: http://www-eosdis.ornl.gov/FIFE/FIFE_Home.html]. Taylor, V. R., and L. L. Stowe, Reflectance characteristics of uniform earth and cloud surfaces derived from Nimbus-7 ERB, J. Geophys. Res., 89(D4), 4987-4996. Valiente, J. A., M. Nunez, E. Lopez-Baeza, and J. F. Moreno, Narrow-band to broad-band conversion for Meteosat-visible channel and broad-band albedo using both AVHRR-1 and -2 channels, Int. J. Remote Sensing, 16, 1147-1166, 1995. Vulis, I. L., and R. D. Cess, Interpretation of surface and planetary directional albedo for vegetated regions, J. Climate, 2, 986-996, 1989.

Chapter 7.4 PHENOLOGY AND AGRICULTURE Frank-M. Chmielewski Subdivision of Agricultural Meteorology, Institute of Crop Sciences, Faculty of Agriculture and Horticulture, Humboldt-University Berlin, Berlin, Germany

Key words:

1.

Agriculture, Horticulture, Climate change, Impacts, Growing season

INTRODUCTION

This chapter deals with both traditional aspects of phenology in agriculture (length of growing season and different applications of phenological data in agriculture) as well as modern aspects, which focus on impacts of climate change on phenophases of field crops and fruit trees. Generally, phenology has a long tradition in agriculture and horticulture. The knowledge of the annual timing of phenophases and their variability can help to improve the crop management which leads finally to higher and more stable crop yields and to an improved food quality. Phenological observations are essential for many aspects in practical agriculture. The data can be used to define the growing season length in a region. On the basis of the available time in the year, cropping schedules can be developed which include suitable crops and varieties, the organization of crop rotation, and catch cropping. Phenological observations also play an important role in processes that are relevant in practical agriculture, such as the timing of irrigation, fertilization, and crop protection. The data are also necessary to evaluate the risk of frost damage and to make forecasts of plant development and harvest dates. In agrometeorological studies, phenological data are used to analyze crop-weather relationships and to describe or model the phytoclimate. The individual sections of this chapter give some examples of Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 505-522 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

506

Phenology: An Integrative Environmental Science

the uses of phenological data in agriculture and horticulture. The chapter ends with more recent aspects in plant phenology, showing that the relatively small changes in air temperature have had already distinct impacts on plant development of fruit trees and field crops.

2.

THE LENGTH OF GROWING SEASON AS AN ENVIRONMENTAL LIMIT FOR CROP PRODUCTION

The average length of the growing season in a region sets the environmental limits for plant production. Each crop needs a certain time for growth, development, and yield formation. In mid-latitudes the average growing time of agricultural crops, which is the time from sowing to ripeness or to harvest, ranges between about three and six months. Some differences exist among the individual varieties. In regions with a long growing season it is always possible to choose the optimal period for cropping. This means to select the time of the year when the environmental conditions, mainly the climate, is the best for the plant. For crops with a relatively short growing time this is even possible in regions with a limited growing season. Spring cereal such as barley is an early maturing crop and needs a relatively short period from sowing to harvest. For sub-arctic adapted cultivars, the growing time can be less than 80 days. In contrast to this, sugar beets have a relatively long growing time, of more than six months (Table 1), which is why sugar beet cropping is limited to regions with a relatively long growing season. The length of growing season also determines how much time is available to cultivate crops in spring or autumn as well as to harvest them at the end of summer. This influences the work sequence and the labor load of farms. Table 7.4-1. Average growing time of selected crops in Germany, Experimental Station Berlin-Dahlem, 1953-2000. Crop Seeding (day.month) Harvest (day.month) Growing time (days) Winter rye Spring barley Oats Maize Potato Sugar beet Field bean Lupine

28.09 04.04 04.04 04.05 26.04 16.04 06.04 08.04

01.08 31.07 06.08 16.10 20.09 24.10 15.08 23.08

192 118 124 165 147 191 131 137

Chapter 7.4: Phenology and Agriculture

2.1

507

Definitions of Growing Season

Generally, the growing season is the time of the year in which plants germinate, grow, flower, fructify, and ripen. Since most biological processes are bound to water, growth starts above the freezing point, usually at approximately 4-5°C for C3-plants. With increasing temperature the biochemical processes are accelerated (rapidly above 10°C) up to the temperature where enzyme systems are destroyed and cells die (Hörmann and Chmielewski 2001). There are a lot of ways to define the length of growing season in midand higher latitudes. The climatic growing season is defined in terms of climatic variables. For instance, thresholds of air temperature can define it, with common values between 0 and 10°C. According to Chmielewski and Köhn (2000), the beginning of growing season was defined as the day of the year on which the average daily air temperature was ≥ 5.0°C (1), with the assumption that on the following days (i) the sum of differences remains positive. ∑ i (Ti – 5°C) > 0°C, (i=2,3,…, 30)

(1)

Correspondingly, the end of the growing season was defined as the day of the year on which the average daily temperature was < 5.0°C, under the condition that: ∑ i (Ti – 5°C) < 0°C, (i=2,3,…, end of year).

(2)

The threshold of 5°C is often used to fix the general growing season for plants in temperate zones. Sometimes for winter cereals a threshold of 3°C is recommended (Reiner et al. 1979). A similar definition was used by Mitchell and Hulme (2002), who suggested introducing the length of growing season as an indicator of climatic changes. They defined the beginning of growing season as the start of a period when the daily mean air temperature is greater than 5°C for five consecutive days. The period ends on the day prior to the first subsequent period when daily mean air temperature is less than 5°C for five consecutive days. For crops the frost-free season (Goodrich 1984), or the time between the last killing frosts in spring and the first killing frost in autumn (Critchfield 1966; Brown 1976), is also used. The length of growing season can also be calculated using phenological events. Schnelle (1955, 1961) defined this period as the number of days between sowing of spring cereals and winter wheat. For the beginning of growing season, the timing of phenological events of natural vegetation like

508

Phenology: An Integrative Environmental Science

bud burst, leafing, and flowering are often used. Accordingly, the end of growing season is then defined by autumn coloring and leaf fall of trees and shrubs.

2.2

The Length of Growing Season in Europe

Using phenological observations of the International Phenological Gardens across Europe: (http://www.agrar.hu-berlin.de/pflanzenbau/agramet/ipg.html) a phenological maps of the beginning, end, and length of growing season were calculated (Rötzer and Chmielewski 2001). The beginning of the growing season was represented by the average date of leaf unfolding of Betula pubescens, Prunus avium, Sorbus aucuparia, and Ribes alpinum. For the end of growing season, the average dates of leaf fall of Betula pubescens, Prunus avium, Salix smithianta, and Ribes alpinum were used. The difference between the end and the beginning of growing season was defined as the length of growing season. On average, the growing season in Europe starts in most regions between April 10 and April 25. The British Isles (with the exception of Scotland), Belgium, the Netherlands, the northern part of France as well as Hungary, Croatia, and the former Republic of Yugoslavia show a beginning of growing season between April 5 and 15 (an earlier beginning before April 5 was calculated for southern France, northern Portugal and Spain and for most of the coastal regions of the Mediterranean Sea). A late beginning of the growing season (between April 15 and 25) was found in most parts of Denmark, Germany, the Czech Republic, and Poland. The growing season starts in Scandinavia and in the Baltic Sea region after April 25. In mountainous regions like the Alps or the Dalmatian Mountains, great differences of the beginning of growing season can be seen. While alpine valleys often show a beginning of growing season before April 15, at high altitudes (above about 1500 m) growing season starts nearly four weeks later. The latest beginning (after May 30) is observed in the highlands of Norway, and north of the Arctic Circle. The end of growing season does not have the broad range of annual timings as the beginning of growing season does. While in eastern Europe, most parts of Scandinavia, southern Germany, Switzerland, Austria and the Czech Republic growing season ends between October 25 and 30, the western part of Europe, northern Germany, Hungary, and southern Europe show slightly later dates (between October 30 and November 4). In high altitudes like the Alps, the Carpathian Mountains, the Dalmatian Mountains, or the Scandinavian Mountains the end of growing season can be observed before October 25.

Chapter 7.4: Phenology and Agriculture

509

Since the end of growing season shows relatively small variations all over Europe, the length of growing season mainly depends on its beginning. The longest growing seasons (with over 220 days) is found in the southern part of France and in the coastal regions of southern Europe (Plate 1). In large parts of Ireland, southern England, the Netherlands, Belgium, most parts of France, Hungary, and in southern Europe (excluding the mountainous regions) growing season lasts between 200 and 220 days (see Table 2). In Scotland and Denmark, in most regions of Germany, in Switzerland, Austria, the Czech Republic, Slovenia, Poland, and in the southern part of Sweden growing season lasts less than 200 days but more than 180 days. Shorter growing seasons with less than 180 days are calculated for high altitudes as well as for nearly all of Scandinavia. High altitudes in Scandinavia as well as the regions north of the polar circle show growing seasons less than 150 days. Table 7.4-2. Average beginning (B), end (E), and length (L) of growing season in different natural regions of Europe, 1969-2000, DOY = day of the year. E (DOY) L (days) Natural Region in Europe B (DOY) British Isles/Channel Coast 101 301 200 North Sea / Central European Lowlands 104 302 198 119 310 191 Baltic Sea Region North Atlantic Mountain Region 127 299 172 North Scandinavia 143 282 139 Northern Central European Highlands 105 305 200 108 305 197 Southern Central European Highlands North Alpine Foreland 110 303 193 119 296 177 Bav.-Bohemian Highlands / Carpathian Mountains 101 303 202 Great Hungarian Lowlands / Danube-Save-Region Dinaric Mountain Region / Dalmatia 108 300 192 Portugal 84 -

3.

PHENOLOGICAL OBSERVATIONS IN AGRICULTURE AND HORTICULTURE

Phenological observations in agriculture and horticulture are common and have great value (see section 4). They include such principal growing stages or field work as seeding / planting, germination / bud development, leaf development, formation of side shoots / tillering, stem elongation / rosette growth / shoot development, booting / development of harvestable parts of vegetative plants, inflorescence emergence / heading, flowering, development of fruit, ripening and maturity of fruit and seed, senescence, beginning of dormancy, and harvest.

510

Phenology: An Integrative Environmental Science

In Germany, the German Weather Service (DWD) is running a network of about 1500 phenological observers. The phenological observations cover natural vegetation species, crops, fruits, and vines. Observed crops are sugar beets, permanent grassland, oats, maize, sunflowers, and winter cereals such as barley, rape, rye, and wheat. Observations of fruits include apple, pear, red currant, sour cherry, sweet cherry, and gooseberry species. For vines the varieties Mueller-Turgau and Riesling are observed (see also Chapter 7.5). In order to have comparable observations it is absolutely necessary to define the phenological phases exactly. For phenological observations, different scales were developed in the past. In agriculture the Feekes-scale was introduced (Feekes 1941). This scale was based on 23 phenological phases for winter wheat between germination and ripeness. By illustrations of Large (1954) the scale became known worldwide (see also Clive-James 1971). In subsequent years the Feekes-scale was modified by many authors for phenological observations of cereals (Keller and Baggiolini 1954; Petr 1966; Broekhuizen and Zadoks 1967). Later, Zadoks et al. (1974) developed a phenological decimal-coded system for cereals, including rice. This scale was published by the European Association for Plant Breeding (EUCARPIA) and is still known as the Eucarpia (EC)-scale. In Germany the EC-scale was adopted by the German Federal Biological Research Center for Agriculture and Forestry (BBA). In 1981, the BBA also published a decimal code for maize, which was not considered in the ECscale before. Today the extended BBCH scale (Strauss et al. 1994) is recommended for phenological observations (see Chapter 4.4). This decimal code, which is divided into principal and secondary growth stages, is based on the well-known cereal code developed by Zadoks et al. (1974). Figure 1 shows the average timing of selected phenophases for three cereal species (oats, spring barley, and winter rye) at the Experimental Station at Berlin-Dahlem, according to the BBCH scale. Apart from the phenological macro-stages, some micro-phenological stages are also shown They describe the which are only visible under the microscope. development of the growth apex and help to define the beginning of the generative development of cereals. A scale for micro-phenological observations was also developed by Nerson et al. (1980). In Figure 1, the timing of three developmental stages at the growth apex (D*, A*, B*) is shown.

Chapter 7.4: Phenology and Agriculture

4.

511

APPLICATIONS OF PHENOLOGY IN AGRICULTURE

Phenological observations in agriculture and horticulture provide basic information for farmers. Knowing the valuable phenological information helps decision-making for farmers, i.e. to correctly time operations such as planting, fertilizing, irrigating, crop protection and to predict phenophases. Phenological data are also helpful for scientific applications, such as investigations of crop-weather relationships and to describe the microclimate of crop stands. The following paragraphs give some examples.

Figure 7.4-1. Average timing of phenophases and the duration of phenological stages of cereal species between sowing and full ripeness at the Experimental Station Berlin-Dahlem, 1962-1996. Phenological stages according to the BBCH code: 00: sowing, 12: 2-leaf-stage, 21: beginning of tillering, 31: beginning of stem elongation, 61: beginning of flowering, 69: end of flowering, 75: milky-, 85 dough-, 87: yellow-, 89: full ripeness and micro-phenological stages. D*: beginning of development at the growth apex, A*: development of spikelets, B*: formation of floret primordial.

4.1

Selection of Growing Zones

Phenological observations can help to select favorable and unfavorable areas for agricultural and horticultural production. The data can be used for site classifications and to find crops and varieties that grow well under the given climatic conditions. Mainly, fruit and vegetable growing is very sensitive to the site selection. Therefore, phenological observations in horticulture have a great value to define optimal cropping areas, to evaluate the cultivability of fruit trees and vegetable in a given region, and to select

512

Phenology: An Integrative Environmental Science

individual species and varieties. Poorly adapted varieties show higher yield fluctuations and even yield depressions. For outdoor vegetable growing in mid-latitudes, mainly the warmest areas with an early spring development are recommended. These areas are easier to find with phenological maps. Phenological observations are therefore helpful to sketch regional and local cropping plans. For instance, maps of the beginning of growing season in Germany clearly show the favorable areas for fruit, vine, and vegetable growing. The regions with a very early beginning of growing season are the warmest areas in spring. In these regions, plant development is clearly advanced, compared with the spatial average. Regional studies for the cultivability of crops in Bavaria are represented by Rötzer et al. (1997).

4.2

Crop Management and Timing of Field Work

For sustainable crop management, phenological data are essential to meet the right dates for irrigation, fertilizing, and crop protection. For example, maize can grow well in regions where the mean air temperature from May to September is above 15°C, but in some regions the rainfall is a limiting factor for growth, so that irrigation becomes necessary. The highest water demand for maize is between the beginning of stem elongation and flowering (BBCH 31 - BBCH 69). Irrigation at the right time is a prerequisite for a sufficient grain yield. At the same time, the demand on nutrients is also very high. In just five weeks around the period of heading, 75% of nutrients are taken up. Other cereals as wheat, barley, and rye have similar demands on water and nutrient supply. According to the nutrient requirement, N-fertilization of cereals is recommended at the beginning of the growing season (to promote the process of tillering as well as the formation of spikelets), in the period of stem elongation (to moderate the reduction of tillers and spikelets as well as florets), and if necessary shortly before heading (to encourage the grain size and protein content in the grain). The application of herbicides for weed control is possible after emergence, for example at BBCH 13 (three-leaf-stage). With the appearance of the last leaf, the flag leaf (BBCH 37), growth regulators can be applied to avoid stalk breakage and the risk of grain lodging. These examples show how phenological information is important for agriculture. Phenological data are prerequisites for adapted and sustainable crop management and to obtain sufficient crop yield. They help farmers to trace the plant development, to monitor the yield formation processes and to find out the optimal time for cultural practices.

Chapter 7.4: Phenology and Agriculture

4.3

513

Risk of Frost Damage

Fruit growers need phenological information mainly in the flowering period and during ripeness. For these individuals it is important to know the date of latest frosts in a region, especially those that occur after first bloom. Early frosts before the beginning of blossom may cause masked injuries in flower buds, but the damage is not as great as it would be in the period of blossom. Frost during the flowering period can harm the blossoms, so that total crop failures can occur. Phenological observations can help to evaluate the frost risk in a region (see next paragraph). For practical aspects in agriculture and horticulture it is sometimes necessary to estimate the timing of phenological events.

4.4

Forecast of Phenological Events

Phenological models use different approaches, such as effective temperature sums or chilling and forcing units to predict the timing of phenophases (see Chapter 4.1). In regions where plant development is mainly forced by temperature, there is another simple way to make phenological forecasts using the following equation:

P2 = P 1 +

Ts (T - TB)

(3)

where P2 is the phenophase to be estimated, P1 is the onset of a previous phase, TS is the effective temperature sum between P1 and P2, T is the average anticipated temperature in the period P1–P2 and TB a base temperature of mainly 5°C. For forecasts, it is necessary to make assumptions about the course of temperatures in the forecast period, because temperatures higher than normal lead to advanced plant development and temperatures below normal can delay the developmental process. The best way is to use the data of medium-range weather forecasts. If they are not available it is also possible to estimate the timing of plant development with climatic data. In this case the error is only relatively small if the average temperature in the forecast period is nearly the same as the long-term average. If not, deviations of several days between estimated and observed timing are possible. There are also ways to estimate phenological events without any information about the temperature values in future days, since phenological events are usually well correlated with each other. There is on the one hand a good correlation between phenophases of different species in the same

514

Phenology: An Integrative Environmental Science

time of the year (Table 3), and on the other hand good correlations in the successive plant development of one species (see later Table 4). For instance, a good indicator for the beginning of fieldwork in mid-latitudes (the first cutting time of meadows) is the anthesis of winter rye. About one week after anthesis the meadows can usually be cut. The beginning of growing season in Germany is well correlated with the blossom of cherry (r=0.92, p<0.01) and the blossom of apple trees (r=0.95, p<0.05). So, it is possible to predict the average blossom of fruit trees with this event. The stem elongation of winter rye and the blossom of cherry trees start nearly at the same time in the year (Figure 2). This phenophase Table 7.4-3. Statistical measures (x: mean, s: standard deviation) and correlation coefficients (r) between different phenophases (BG: beginning of growing season, BC: beginning of cherry trees blossom, BA: beginning of apple trees blossom, B31: beginning of stem elongation of winter rye (BBCH 31) in Germany, 1961-2000. Correlation coefficient r ≥ 0.31 are significant with p<0.05. x s BG B31 BC BA BG B31 BC BA

109.8 116.6 116.9 125.3

7.3 6.1 7.5 7.4

1.00

0.88 1.00

0.92 0.91 1.00

0.95 0.88 0.96 1.00

can be used to investigate the probability for killing frosts after the beginning of blossom, and to evaluate the planting risk of fruit trees in a region. Phenological observations of fruit trees are not always necessary to do this. The good correlation between the phenological stages of one species provides another way to predict phenological events (Table 4). For this method very detailed phenological observations should be obtained. The timing of a phenological phase (P2) can be estimated using the previous phase (P1): P2 = a + b P1

(4)

Larger deviations in the plant development from the long-term average are reduced gradually, so that successive phenological stages are usually positively correlated. There is always a minimum period between two successive phases. Table 4 shows that for winter rye it is possible to estimate the beginning of heading using the beginning of shooting (r=0.93). Also the timing of emergence and full flowering can be estimated by the previous phase. With increasing time periods, the correlation between the phenological events decreases, so that it is not possible to estimate a

Chapter 7.4: Phenology and Agriculture

515

phenological event by a very early one. Relatively good results (shown in bold in Table 4) can always be achieved if the previous phase is taken.

4.5

Crop-Weather Relationships

Crop yields are influenced by the variability of weather in many ways. For cereals the yield formation is very complex. Phenological observations can help to divide the growing time of crops into different periods that are important for the yield formation (Table 5). So, it is possible to find out

Figure 7.4-2. Beginning of growing season (BG) and the beginning of blossom of cherry trees (C60), apple trees (A60), and beginning of stem elongation of winter rye (R31) in Germany, 1961-2000. Data source: German Weather Service (DWD).

Table 7.4-4. Correlation coefficients between different phenophases of winter rye in Germany, 1961-2000, (BBCH code is used, 00: seeding, 09: emergence, 31: beginning of stem elongation, 51: beginning of heading, 65: full flowering). Correlation coefficient r ≥ 0.31 are significant with p<0.05. BBCH 00 09 31 51 65 harvest 00 09 31 51 65 harvest

1.00

0.97 1.00

0.52 0.53 1.00

0.45 0.46 0.93 1.00

0.31 0.31 0.69 0.85 1.00

0.20 0.18 0.53 0.64 0.66 1.00

516

Phenology: An Integrative Environmental Science

relationships between weather and crop yield for different periods in which the individual yield component is affected. Investigations by Chmielewski and Köhn (1999) showed that for spring barley in the first period mainly the number of kernels per ear, in the second period the crop density, and in the last period the kernel weight are influenced by weather. For winter rye, it was also possible to find out the relevant phenological periods with regard to the yield formation (Chmielewski and Köhn 2000). Thus, phenological observations are absolutely necessary to describe the yield formation of crops more in detail. Crop models use phenological information as well, to steer physiological processes in the model. Therefore, phenological models are always subroutines of mechanistic crop models (see Chapter 4.1). Table 7.4-5. Relevant developmental periods date ranges are given for spring barley at the 1996. Period 2-leaf-stage (BBCH 12) – Beginning of stem elongation (BBCH 31) 27 April – 21 May

Beginning of stem elongation (BBCH 31) – General flowering (BBCH 65) 21 May – 16 June General flowering (BBCH 65) – Yellow ripeness (BBCH 87) 16 June – 20 July

4.6

for the yield formation of spring cereals. The Experimental Station in Berlin-Dahlem, 1962Yield Forming Processes - formation of side-shoots - tillering Beginning of development at the growth apex: - formation of spikelets (D*: 26.04 – 03.05.) - differentiation of spikes (A*: 06.05 – 10.05.) - formation of florets (B*: 12.05. – 17.05.) - reduction of tillers and spikelets - differentiation and reduction of florets - flowering - formation of kernels - growth and ripeness of kernels

Microclimate of Crop Stands

Crop stands have their own climate that can differ tremendously from the climatic conditions at a meteorological station. The microclimate of crop stands (phytoclimate) depend on various meteorological (solar radiation, air temperature, precipitation, wind speed) and plant-morphological factors (structure of plant cover, plant height, density of stand, etc.) and thus also on the plant development. Therefore, phenological observations are essential to analyze and to understand the microclimate of crop stands. During the day the air within a crop stand is warmer and wetter than the air above a bare soil (Wittchen 2002). The phytoclimate varies during the daytime and with the growth and development of the crop stand (Figure 3). The largest air temperatures differences occur in the last developmental period (P4) at noon. Then the winter rye stand in 0.20 m height is

Chapter 7.4: Phenology and Agriculture

517

significantly warmer, compared with the air above a bare soil at the same level. During the flowering phase (Phase 3) the differences in water vapor are the largest. They are already reduced in the grain filling and ripeness period, because then the stand becomes drier and drier. These specific phytoclimatic conditions have effects on the phytopathological situation within the crop stand, because the development of fungi and insects is determined by given temperatures and humidity levels, typical of each species. In the case of insects, the air temperature in the breeding places of insects is important for the eggs to hatch and for the larvae to be able to perform their various transformations (see Chapter 6.2). For fungi, beside air temperature mainly air moisture or water on plant organs play important roles. Since the phytoclimate depends on the plant development, the infection risk also changes with plant development. For example for winter rye the infestation frequency is very high in the last two phases (P3, P4), because of the higher air humidity in these two periods compared with the climatic conditions outside the stand. As a measure for the infection risk in phytopathological models, often the relative humidity is used. Here mainly the time periods with values of at least 75% are relevant.

Figure 7.4-3. Average anomalies in air temperature (ΔT) and in water vapor pressure (Δe) between a winter rye stand and a bare surface in 0.2 m height during different developmental periods. P1: leaf development and tillering period (BBCH 10-29), P2: stem elongation and heading period (BBCH 30-49), P3: flowering period (BBCH 50-69), P4: grain filling and ripeness period (BBCH 70-89), at the Experimental station in Berlin-Dahlem, 1981-1999 (positive anomalies mean the stand is warmer and more wet).

Investigations by Wittchen (2002) showed that for time intervals of more than 12 hours the relative humidity in Phase 3 and Phase 4 is indeed

518

Phenology: An Integrative Environmental Science

significantly higher compared with the conditions in a weather shelter (see bold values in Table 6). On the other side in Phase 4 the air temperatures mostly exceed the optimal range for infections, so that the infection risk in the winter rye stand is the highest during the flowering period. Table 7.4-6. Relative frequency of relative humidity HR ≥ 75% within a winter rye stand, compared with HR in a weather shelter (2m height). Duration of HR ≥ 75 % for different time intervals in % Time interval 1…6 7 … 12 13 … 18 19 … 24 25 … 36 > 36 (h): Flowering period (P3) Stand 15.3 32.3 33.5 8.2 1.6 9.1 Shelter 37.6 38.1 14.6 4.9 1.3 3.5 Grain filling and ripeness period (P4) Stand 17.7 34.8 33.4 8.8 1.0 4.3 Shelter 19.8 3.8 0.9 2.5 31.6 41.4

5.

CLIMATE CHANGE AND PHENOPHASES

The strong relationships between air temperature and plant development in mid- and high-latitudes make phenological observations sensitive indicators to evaluate possible biological impacts of climate change. Since the end of the 1980s, clear changes in air temperature have been observed in Germany and many other parts of Europe. Mainly the temperatures in winter and in the early spring—which are decisive for the plant development in spring—changed distinctly. Most recent years were warmer than the long-term average (Chmielewski and Rötzer 2002). These observed changes in temperature correspond well to changes in the circulation pattern over Europe. The increased frequency of positive phases in the North Atlantic Oscillation index (NAO) since 1989 led to milder temperatures in winter and the early spring, because of the prevailing westerly winds from the Atlantic Ocean during this time of the year (Chmielewski and Rötzer 2001). These climate changes led to distinct reactions in the flora. Between 1969 and 2000, the average beginning of growing season in Europe has advanced by nine days. This corresponds to a significant trend of -2.8 days/decade (p<0.05, Figure 4). In accordance with the climatic changes, mainly since the end of the 1980s, early dates prevail. Between 1989 and 2000, 10 out of 12 years had an advanced onset of spring. Compared to the beginning of growing season, the average end of growing season shows smaller annual variations. The trend to a later end of about one day/decade is relatively small. Mainly influenced by the earlier beginning, the average

Chapter 7.4: Phenology and Agriculture

519

length of growing season had advanced by 11 days, corresponding to a significant trend of 3.5 days/decade (p<0.01). Because of the very early onset in 1990, this year had the longest growing season (200 days). Changes in the length of growing season can influence crop management in agriculture and horticulture such as cultivar selection, catch cropping, and crop rotation. Small changes in the length of growing season can already influence the choice of varieties. Fast ripening varieties can be replaced by slower ripening ones, if the growing season extends. This measure could have positive effects on the yield variability and on the yield level. In fruit growing changes of the varieties are also possible, because the differences in the growing time of fruit varieties are mostly only several days. Catch cropping g depends on the time available after harvest in the late summer or early autumn, before another crop is cultivated. Thus, catch cropping is only possible in regions where the growing season in autumn is long enough and the air temperatures and precipitation are still favorable for plant growing. For example an advanced harvest date of spring cereals would improve the conditions for catch cropping, so that it is possible to grow legumes, before winter cereal is sown. More distinct changes in growing season length by several weeks can influence the possibilities for catch cropping and for crop rotation.

Figure 7.4-4. Trends in the beginning (B), end (E) and length (L) of growing season in Europe, 1969-2000, based of observations from the International Phenological Gardens.

Crop rotation is a system in which the crops on a certain plot are followed by other crops according to a predefined plan. Normally the crops are changed annually, but in some areas where the growing season is sufficiently long, multiple cropping is possible. An extension in the length of growing season can improve the scope for multiple cropping and crop rotation. For example, if the end of growing season is extended the sowing time of winter cereals can be shifted to the end of the year, so that in the

520

Phenology: An Integrative Environmental Science

space before crops with a relatively long growing time (such as sugar beets) can grow. In the field, vegetable multiple cropping is common. Here also an extension of growing season can improve the crop rotation and number of harvests within a year. But not only the length of growing season has changed in recent years. Events, such as the first bloom of fruit trees in Germany (apple, cherry, etc.), were also influenced by higher temperatures during the end of winter and in early spring (Chmielewski et al. 2003). An increase of average air temperature between February and April of 1.6°C between 1961 and 2000 led to an advanced tree blossom in Germany of eight days (Figure 5). Here also the most distinct changes occur since 1989. An earlier blossom of fruit trees holds at the same time the danger of damage by late frosts. Thus it is very important to monitor the changes in plant development, to be prepared for the impacts of climatic changes on agriculture and horticulture. Phenological research can improve the selection of fruit varieties for specific regions to reduce the risk of potential frost damage. It can result in better projections and assessments of flowering times and occurrence of pests and diseases. This will improve and substantially reduce the use of pesticides and increase the agricultural production potential in Europe. Information on the potential effects of climate change on agricultural and horticultural crops by changes in phenology are still limited at this time.

Figure 7.4-5. Anomalies in the first bloom of cherry trees (BC) and in the average air temperature from February to April (T24) in Germany, 1961-2000. Data: German Weather Service (DWD).

Chapter 7.4: Phenology and Agriculture

521

REFERENCES CITED Broekhuizen, S. and J.C. Zadoks, Proposal for a decimal code of growth stages in cereals, Stichting Netherlands Graan-Centrum, Wageningen, 1967. Brown, J. A., Shortening of growing season in the U.S. corn belt, Nature, 260, 420-421, 1976. Chmielewski, F.-M., and W. Köhn, Impact of weather on yield components of spring cereals over 30 years, Agricult. Forest Meteorol., 96, 49-58, 1999. Chmielewski, F.-M., and W. Köhn, Impact of weather on yield and yield components of winter rye, Agricult. Forest Meteorol., 102, 253-261, 2000. Chmielewski, F.-M., and T. Rötzer, Response of tree phenology to climate change across Europe, Agricult. Forest Meteorol., 108, 101-112, 2001. Chmielewski, F.-M., and T. Rötzer, Annual and spatial variability of the beginning of growing season in Europe in relation to air temperature changes, Clim. Res., 19(1), 257264, 2002. Chmielewski, F.-M., A. Müller, and E. Bruns, Climate changes and trends in phenology of fruit trees and field crops in Germany, 1961-2000, in Beiträge zur Klima- und Meeresforschung, edited by F.–M. Chmielewski, T. Foken, pp. 125-134, Eigenverlag Berlin und Bayreuth, 2003. Clive-James, W., Growth stages key for cereals, Can. Pl. Dis. Surv, 51, 42-43, 1971. Critchfield, H. J., General Climatology, Prentice-Hall, Englewood Cliffs, New Jersey, 420 pp., 1966. Feekes, W., De tarwe en haar milieu, Versl. techn. Techn. Tarwe Commissie, 12, 523–888, and 17, 560–561, 1941. Goodrich, S., Checklist of vascular plants of the Canyon and Church Mountain (Utah, USA), Great Basin Nat., 44, 277-295, 1984. Hörmann, G., and F.-M. Chmielewski, Consequences for agriculture and forestry, Chapter 3.32, in The climate of the 21st century, edited by J. L. Lozan, P. Hupfer, and H. Grassl, pp. 322-330, Wissenschaftliche Auswertungen, Hamburg, 2001. Keller, C., and M. Baggiolini, Les stades repéres dans la vegetation du blé, Revue Romande d’ Agriculture 10(3), 17-30,1954. Large, E. C., Growth stages in cereals, Illustration of the Feekes Scale, Plant Pathol., 3, 128129, 1954. Mitchell, T. D., and M. Hulme, Length of the growing season, Weather, 5, 57, 196-198, 2002. Nerson, H., M. Sibony, and M. J. Pinthus, A scale for the assessment of the development stages of the wheat (Triticum aestivum L) spike, Ann. Bot., 45, 203-204, 1980. Petr, J., A precise phenological scale for grain cereals, Rostl. Vyroba, 12, 207-212, 1966 Reiner, L., A. Mangstl, F. Strass, W. Teuteberg, E. Panse, P. W. Kürten, B. Meier, W. Grosskopf, U. Deecke, P. Kühne, and J. G. Schwerdtle, Winterroggen aktuell, DLG Verlag, Frankfurt a.M., 13pp., 1979. Rötzer, T., and F.-M. Chmielewski, Phenological maps of Europe, Climate Research, 18(3), 249-257, 2001. Rötzer, T., W. Würländer, and H. Häckel, Agrar- und Umweltklimatologischer Atlas von Bayern, Selbstverlag Deutscher Wetterdienst, 1997. Schnelle, F., Pflanzenphänologie, Geest and Portig, Leipzig, 299 pp., 1955. Schnelle, F., Agro-phenological annual course of the German and European agricultural regions, German Geographic Meeting, Wiesbaden, 1961. Strauss, R., H. Bleiholder, T. van den Bomm, L. Buhr, H. Hack, M. Heb, R. Klose, U. Meier, and E. Weber, Einheitliche Codierung der phänologischen Entwicklungsstadien monound dikotyler Pflanzen, Erweiterte BBCH-Skala, Basel, 27 pp., 1994.

522

Phenology: An Integrative Environmental Science

Wittchen, U., Beschreibung und Modellierung des Mikroklimas in Wintergetreide-Beständen unter besonderer Berücksichtigung langjähriger Messungen in Berlin-Dahlem und Dahnsdorf, f PhD-Thesis, HU-Berlin, Landwirtschaftlich-Gärtnerische Fakultät, 2002. Zadoks, J. C., T. T. Chang, and C. F. Konzak, A Decimal code for the growth stages of cereals, Weed Res., 14, 415-421, 1974.

Chapter 7.5 WINEGRAPE PHENOLOGY Gregory V. Jones Department of Geography, Southern Oregon University, Ashland, OR, USA

Key words: Vitis vinifera, Winegrapes, Grapevines, Viticulture, Bioclimatic indices

1.

INTRODUCTION

Grapevines are a woody, herbaceous tree-climbing plant/shrub of the Vitaceae family with a largely uncertain origin. Fossilized remains have been found in Paleocene and Eocene deposits indicating that vines have been around for at least 37 million years (Galet 1979). While approximately 24,000 varieties of vines have been named, it is thought that one fifth or less are probably genuinely distinct varieties and less than 150 are cultivated to any degree (Coombe and Dry 1988). All grapevines belong to the genus Vitis, including the Euvitis (true grapes with both European and North American species) and Muscadinia (whose fruit is more properly called muscadine) subgenera, with most found mainly in the temperate zones in the Northern Hemisphere. Of the main species, Vitis vinifera, which is a Eurasian native, is responsible for most of the table grapes, raisin grapes, grape juice, wine, and vinegar produced today. Most grapevines evolved with adaptive features such as tendrils for climbing (there are a few shrub-like varieties) and a very high hydraulic conductivity that enabled them to survive in a forest habitat. As part of the forest adaptation, the light-driven formation of flowers at the forest canopy replaced tendrils (homologous structures) at the terminal nodes. Today, grape growers attempt to maximize this characteristic by managing vines such that vigor is constrained and a perennial structure that promotes Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 523-539 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

524

Phenology: An Integrative Environmental Science

initiation and differentiation of fruitful buds is produced. In addition, it is thought that the early grapevines were dioecious but that over time hermaphroditic forms appeared. All species in the Vitis genus hybridize easily (highly heterozygous) and will mutate under different environmental conditions. This trait, while producing a large diversity within the genus, has lead to difficulties in identifying the parentage of some of today’s most prized varieties. The cultivation of grapevines predates written history. Archeological findings in the Caucasian Mountains, near the town of Shiraz in ancient Persia, indicate that viticulture (the cultivation of grapes) existed as early as 3500 B.C. (Penning-Roswell 1989). Vitis vinifera (“wine-bearing vine”) was first domesticated in this region and soon spread to Assyria, Babylon and to the shores of the Black Sea. The Assyrians, Phoenicians, Greeks, and Romans furthered viticulture and viniculture (the process of fermenting grapes into wine) and spread its knowledge to Palestine, Egypt, North Africa, the Iberian Peninsula, and throughout Europe to as far north as the British Isles (Unwin 1991). During the Dark Ages grape growing declined throughout most of Europe and would have died out had it not been for the Christian monks who preserved the methods of viticulture and made vast improvements in cellaring techniques (Loubere 1990). With new world explorations, Europeans carried the vine with them and helped establish the industry in regions that were well suited for cultivation. Today viticulture is an ever-growing agribusiness, even more so as more is learned about the health properties of drinking wine (Mansson 2001). Better understanding of the biology of the vine and climatic constraints on the individual varietals of V. vinifera has opened up new regions to viticulture. In general, mechanization of viticulture and viniculture has furthered the grower’s ability to produce at a greater profit and new technologies in genetics (clones and rootstocks) have reduced the susceptibility of V. vinifera to many parasites and diseases.

2.

PHENOLOGICAL CYCLE AND CHARACTERISTICS

Like all agricultural and natural plant systems, grape producers and viticulture scientists need crop developmental scales that are easy to use, universally accepted, and accurate. Grapevine growth stage identification has been carried out to better provide standards by which growers worldwide can communicate information (see Coombe 1995 for a review). In addition, the study of grapevine phenology has allowed growers to better understand whether or not a given variety can produce a crop within a given climate

Chapter 7.5: Winegrape Phenology

525

regime (e.g., Tesic et al. 2002a, b); has facilitated husbandry by providing the structure by which cultural and chemical practices can be applied at optimum periods during a plant’s growth (Matthews and Anderson 1987); and has been useful in estimating crop yields (Clingeleffer et al. 1997). Four generally accepted grapevine growth descriptive systems, each based upon principal growth stages with varying levels of secondary growth stages, have been established. 1) Baggiolini (1952), later amended by Baillod and Baggiolini (1993), was the first system proposed for grapevines and consisted of 16 stages from bud break to leaf fall that was widely accepted to aid pesticide applications. 2) Eichhorn and Lorenz (1977) followed with a more comprehensive system that entails 22 stages from “winter bud” to “end of leaf fall” that has been accepted by many in the industry. 3) The BBCH system was developed by the European Union to standardize the many scales that were in use for single plant species or group of related species and has been proposed as a prototype of a universal scale– the Extended BBCH Scale for all monocot and dicot crops (Lorenz et al. 1995; Meier 1997). The adaptation of the BBCH system to grapevines resulted in seven macro-stages of growth with numerous micro-stages in between. 4) In a thorough assessmentt of descriptive schemes for the grapevine, Coombe (1995) found that the BBCH system included terms that were not universally understood, had a zigzag rather than continuous developmental sequence, and had minor errors in the descriptions. From this assessment and detailed data analysis from phenology trials in Australia, the author suggests the Modified Eichhorn and Lorenz system (Modified E-L) as meeting universal requirements for grapevines. The Modified E-L provides a 47-stage system, which includes a simple listing of eight major growth stages along with detailed intermediate stages (Figure 1). The growth cycle of grapevines starts in the early spring with the shoot and inflorescence development stage. This stage commences with the breaking of the dormant winter stage and a rise in sap or bleeding of the vine. Four to six weeks after the sap starts to rise, the vine starts to produce foliage (tendrils and leaves) from the latent primary buds formed at the end of the previous year. This stage, termed budburst or bud break (debourrementt by the French), is the budding out of the vine before the floral parts develop later in the spring. Once shoot growth reaches roughly 10 cm (Modified E-L stage 12, Figure 1), numerous leaves are separated and the inflorescence becomes clearly visible. The flowering ((floraison) stage starts in late spring or early summer when the young shoots differentiate their meristems and put forth flower clusters at nodes along the vine. Each flower in the cluster is covered by a cap that, through further growth, breaks away (anthesis) revealing the flower parts of the vine. Full bloom is considered when 50% of the caps have fallen off. After flowering, the green

526

Phenology: An Integrative Environmental Science

berries set (nouaison) on the clusters (only 20-60% of the potential berries set) and develop at right angles to the stem. The berries grow in bulk over the next two months and when they reach pea size the clusters start to hang downward. During this stage, there is very little chemical differentiation

Figure 7.5-1. Modified E-L system for grapevine growth stages, reprinted from Coombe (1995), with permission from the Australian Society of Viticulture and Oenology.

Chapter 7.5: Winegrape Phenology

527

except for a slight change in acidity of the fruit. Nearing the end of berry development, however, the berries go through a period of rapid physical and chemical change called véraison, which initiates the ripening stage. During this stage, the grapes soften, enlarge, start accumulating sugar, and change color to translucent greenish-white (white varietals) or to a red-purple hue (red varietals). In managed systems, the berries are considered harvest-ripe when the sugar/acid ratio and flavor profile is optimized for a particular variety or desired wine style. After harvest the vines begin their senescence stage with the canes maturing (turning hard and woody) and the leaves falling. While the onset and duration of each of the main phenological stages of V. vinifera grapevines varies spatially and for individual varieties, they are very consistent for the physiology of the main varietals in a given region and can be approximated by: – Stage I: Shoot and Inflorescence Development – commencing around the end of March or first week of April (end of September to early October in the Southern Hemisphere). – Stage II: Flowering – generally occurring in the first few weeks of June (late November to early December in the Southern Hemisphere). – Stage III: Berry Developmentt – from the end of flowering in mid-June (mid-December in the Southern Hemisphere) to the ripening stage. – Stage IV: Ripening – starts with véraison near the end of July or the first week of August (late January to early February in the Southern Hemisphere). – Stage V: Senescence – from harvest, late September through early November (late March through late April in the Southern Hemisphere) and leaf fall, over the winter months leading back to bud break.

3.

FACTORS AFFECTING PHENOLOGY

It has been said, “viticulture is perhaps the most geographically expressive of all agricultural industries” (de Blij 1983, p. 112). Along with this geographical signature comes the concomitant climate of a region, which influences and controls the characteristics and quality of the grapes and wine produced there. The Mediterranean and marine west-coast climate regimes are synonymous with the majority of the viticulture regions of the world (however, both table and wine grape growing has become more viable in other areas as knowledge and varietal development for those regions has grown) as the mild wet-winter, dry-summer climate found there is ideal for the cultivation of V. vinifera grapes. The main regions are found along the Mediterranean basin of Europe, the fynbos of South Africa, the mallee of

528

Phenology: An Integrative Environmental Science

southern Australia, the mattorall of Chile and Argentina, and the chaparral and coastal valleys of western North America. Each of the major phenological stages of V. vinifera grapevines are governed by critical climatic influences. Temperature effects are evident in the spring where vegetative growth is initiated by prolonged average daytime temperatures above 10°C (Amerine and Winkler 1944). During this stage, temperatures below -2.5°C can adversely affect the growth of the herbaceous parts of the plant and hard freezes can reduce the yield significantly (Riou 1994). During floraison and throughout the growth of the berries, extremes of heat can be detrimental to the vines. While a few days of temperatures greater than 30°C can be beneficial to ripening potential, prolonged periods can induce heat stress in the plant and lead to premature véraison, the elimination of the berries through abscission, permanent enzyme inactivation, and partial or total failure of flavor ripening. During the maturation stage, a pronounced diurnal temperature range effectively synthesizes the tannins and sugars in the grapes. Early frost or freezes during maturation can lead to the rupture of the grapes along with significant loss of volume. During the dormant stage, a temperature minimum or effective chilling unit (hours below a certain temperature) is generally needed to effectively set the latent primary buds (Galet 1979). Throughout the phenological stages off the grapevine, the amount of solar irradiance is critical in maintaining the proper levels of photosynthetic activity for the production of assimilate. The most critical stages come during the development of the berries starting at floraison and continuing through the harvest. During floraison, high amounts of solar irradiance result in effective meristem differentiation into flowers. During this stage, prolonged periods of cloudiness, cold temperatures, and excessive rain results in coulure or the failure to fully flower and set the berries. During maturation, solar irradiance mainly acts to control the amount of sugar accumulation in the grapes, and therefore, the potential wine alcohol content. The role that the wind plays in the growth of the grapevine and the production of fruit is mainly through the effects on vine health and yield. This is manifested in both a physical nature through direct contact with the vines and through physiological effects of stomata closure and reduced disease infestations. During the early stages of vegetative growth, high winds can break off the new shoots, delaying and even reducing the amount of flowering. As the berries proceed through véraison and into the maturation stage, high winds can desiccate the fruit resulting in lower volume and quality. However, drying winds that occur at night and early morning can help reduce the occurrence of fungus born diseases by limiting the formation of dew on the leaves and berries. Nighttime winds can also be beneficial in that they can help limit the occurrence of radiation frosts.

Chapter 7.5: Winegrape Phenology

529

During the growth cycle, the weather conditions that can most severely afflict the vines and berries are associated with moisture. Atmospheric moisture, in the form of humidity and rainfall, hastens the occurrence of fungal diseases (i.e., powdery mildew, downy mildew, botrytis bunch rot). In extreme cases, water stress resulting from high evaporative demand can manifest itself in leaf loss, severe reductions in vine metabolism, and fruit damage or loss. Even mild periods of moisture stress can substantially reduce the relative level of photosynthesis, resulting in lower fruit yields and quality. Fungal diseases can be problematic in the morning when high temperatures and condensation combine to hasten the disease. The formation of fungal diseases can cause defoliation, reduced sugar accumulation, and a reduction in winter hardiness (Amerine et al. 1980). The occurrence of rain during critical growth stages can lead to devastating effects. While ample precipitation during the vegetative stage is beneficial, during flowering rainfall can reduce or retard inflorescence differentiation and during berry growth rainfall can enhance the likelihood of fungal diseases. As the berries mature rainfall can further fungus maladies, yellow and dilute the berries, reducing the sugar and flavor levels, and severely limit the yield and quality. Examination of the world’s viticulture regions suggests that there is no upper limit on the amount of precipitation needed for optimum grapevine growth and production (Gladstones 1992), but grapevine viability seems to be limited in some hot climates by rainfall amounts less than 600-750 mm, although this can be overcome by regular irrigation. Extreme meteorological events, such as thunderstorms and hail, while rare in most viticultural regions, can be extremely detrimental to the crop by damaging the leaves, tendrils, and berries and if they occur during maturation can split the grapes, causing oxidation, premature fermentation, and a severe reduction in volume and quality of the yield (Winkler et al. 1974). While grapevines are planted to a wide diversity of landscapes, the sites that produce the best quality winegrapes are generally planted on moderate slopes with good sunlight exposure (aspects). South-facing slopes provide for more irradiance, and therefore photosynthesis, with a south-facing slope of 8° providing a 77° noon sun angle at 44°N latitude (a 12% irradiance increase over a flat site). Depending on numerous other factors such as obstructions (e.g., trees, buildings, other hills, etc.), a properly situated slope can enhance growth and maturation or limit disease problems (Jackson and Schuster 1987). The slope’s aspect influences phenological development and canopy characteristics through the amount and timing of solar irradiance received. In general, northwest, north, and northeast aspects will experience delayed grape growth stages, less sunlight, and lower evapotranspiration rates from the soil and canopy (Northern Hemisphere). Southeast, south, southwest, and west aspects, on the other hand, will tend to exhibit earlier

530

Phenology: An Integrative Environmental Science

grape growth stages and show varying increases in irradiance and evapotranspiration (Northern Hemisphere). In addition, cold air drainage problems from the pooling of cold air in low spots or via obstructions on sloping land can delay a plant’s phenology or even reduce its viability (Wolf 1997). Soil effects on grapevine phenology are mostly through water retention and plant water relationships. Gravelly to rocky soils provide good drainage and higher heat storage, which accelerates phenological development, while heavy clay soils, which retain moisture, can slow down growth and inhibit productivity. The temperature of the soil can also have a strong influence on vine growth and fruitfulness with warmer spring soils hastening early season growth (Robinson 1994). In addition to the relative amount of sunlight, the composition and color of the soil, the local topography, and drainage capabilities are all factors affecting canopy temperatures, especially at night.

4.

PHENOLOGICAL OBSERVATIONS AND RESEARCH

4.1

Bioclimatic Indices

Some of the first viticulture-climate studies were conducted by A. P. de Candolle in France during the mid 19th century. It was de Candolle’s observation that vine growth started when the mean daily temperature reached 10°C that led to the notion of a heat summation above a base temperature that could define vine growth stages and grape maturation. Amerine and Winkler (1944) furthered research of the amount of available heat that governs the growth period needed to produce high quality premium winegrapes. The authors developed a heat summation index for California that is widely used as the definitive criterion for determining a given area’s suitability to producing quality winegrapes. The heat summation index is calculated for the period of April 1 through October 31 (Northern Hemisphere) by taking the mean monthly temperature and subtracting a base of 10°C (the minimum at which vine growth occurs) and multiplying by the number of days in the month (or summed with daily data). Using the index, five climatic regions based on vine growth and ripening potential are defined, with region I having less than 1400 degree-days and region V having more than 2250 degree-days. The index gives a minimum and maximum threshold of 900 and 2700 degree-days, respectively, for the cultivation of winegrapes. The best quality wine is found to coincide with regions that experience between 1400-1950 degree-days.

Chapter 7.5: Winegrape Phenology

531

Other bioclimatic indices have been used to characterize a region’s potential for viticulture and are mostly developed on the basis of heat units (degree-days). Various forms of a heliothermal index have been used (Branas 1974; Huglin 1978) along with a latitude-temperature index (Jackson and Cherry 1988) to help define the suitability of a region to the planting of certain varietals. Smart and Dry (1980) developed a simple classification of viticultural climates that uses five dimensions of mean temperatures, continentality, sunlight hours, aridity, and relative humidity. Gladstones (1992) developed a classification similar to Amerine and Winkler (1944) but refines it by imposing an upper limit on mean temperatures (19°C), a correction factor for latitude, and a correction for each month’s temperature range. In general, the growing season of winegrapes varies from region to region but averages approximately 170-190 days (Mullins et al. 1992) and the length of the frost-free season is important to the onset of bud break, flowering, and the timing of harvest. Research has also shown that there is a minimum winter temperature that grapevines can withstand. This minimum ranges from -5°C to -20°C and is chiefly controlled by micro-variations in location and topography (Amerine et al. 1980; Winkler et al. 1974). Temperatures below these thresholds will damage plant tissue by the rupturing of cells, the denaturing of enzymes by dehydration, and the disruption of membrane function (Mullins et al. 1992). Prescott (1965) also notes that an area is suitable for grape production if the mean temperature of the warmest month is more than 18.9°C (66°F) and that of the coldest month exceeds -1.1°C (30°F).

4.2

Phenology, Yield, and Quality Interactions

Phenological observations by individual growers, through collective networks, or research-focused studies have been used to assess regional differences in plant maturity potential, to determine the timing of cultural practices, and provide some measure of a given vintage’s development. While no worldwide single method of observing phenological events exists, it is common to indicate that an event (e.g., bud break, flowering, etc.) has occurred when 40-60% of the plants in a given area (an entire vineyard or within a vineyard block containing a specific variety) are exhibiting the event (Coombe and Dry 1988). In general, bud break, flowering, véraison, and harvest dates are the most observed events with very few growers noting any of the more detailed micro-stages in the Modified E-L system. Of these, McIntyre et al. (1982) and Jimenez and Ruiz (1995), show that the onset of bud break and flowering is very consistent from variety to variety while

532

Phenology: An Integrative Environmental Science

véraison and maturity (date of harvest) is less predictable due to greater vine management variability between growers. Numerous studies based on climate parameters (mostly temperature) have been used to try to predict the dates of the individual phenological events. Swanepoel et al. (1990) developed a bud break model using cultivar specific constants that explains over 80% of the variability in the events in the warm climate region of Stellenbosch, South Africa. Due et al. (1993) reveal that average temperatures, summed from specific dates and not one event to another, are good predictors of events. Calò et al. (1994) found that the length of the interval from bud break to flowering and the flowering date were best modeled using daily average maximum temperatures rather than averages or summations of temperature. Tesic et al. (2002b) reveals an interesting relationship between the relative coldness of the winter on bud break with warm winters inducing an early, slow, and heterogeneous event and cold winters a later, more rapid, and homogeneous bud break. While some success has been achieved, all of the authors agree that more detail is needed to increase the understanding of the processes and the accuracy of the models. Much of the research mentioned above was done using climate indices or summation methods. Giomo et al. (1996) discussing the usefulness of these climate indices in grapevine growth analyses, indicates that most were developed to study global to regional climates and are not useful at the subregional level. McIntyre et al. (1987) showed that a simple summation of the number of days in an interphase, averaged over a long period, is a better predictor of phenological events than the single sine degree-day method (Zalom et al. 1983) used in California. Jones and Davis (2000b) also found that degree-days do not readily predict phenological timing, yield, or quality in Bordeaux, France. Moncur et al. (1989), examining bud break and leaf appearance for 10 cultivars, suggest that 4°C and 7°C are better base temperature values to use when modeling phenology than the more commonly used 10°C. Jimenez and Ruiz (1995) also note that using accumulated degree-days above 0°C or the number of days between events is a better predictor of phenology on average than degree-days above a 10°C threshold. The rate of development between the growing season phenological stages varies with variety, climate, and topography. In regions with cool climates, early ripening varieties produce well and in hot climates, late ripening varieties are better suited. In addition, differences in vine management affect harvest dates, yield, and quality. Growers who want higher yields generally will harvest lower quality fruit and produce moderate to low quality wine, whereas higher quality grapes are associated with lower yields. The timing of these developmental stages is also related to the ability of the vine to

Chapter 7.5: Winegrape Phenology

533

produce, with early and fully expressed phenological events usually resulting in larger yields (Mullins et al. 1992; Jones 1997). Additionally, the pace by which vines go through their stages has been related to vintage quality with shorter intervals and early harvests generally resulting in higher quality (Jones and Davis, 2000b). During flowering, the weather is crucial and can ultimately determine the vine yield. A serious failure to flower properly may mean that the vines will develop clusters with few or no berries. During maturation, the crop can be greatly affected by rainfall and high humidity, which can induce fungal diseases. These diseases rob moisture from the berries, which reduces the yield and can render much of the harvest useless. Jackson and Lombard (1993) note that excess precipitation has a negative effect on fruit quality by increasing vigor, delaying phenological events, reducing berry set, and increasing disease pressure. Barbeau et al. (1998a, b) established a connection between soil texture, soil temperature, and vine phenology in Cabernet Franc vines in the Loire Valley of France. The authors found that sites with good drainage have earlier phenological events, whereas heavy clay soils and soils with perched water tables have later phenological events. In addition, the sites that produced earlier phenological events showed increases in accumulated sugar, achieved better anthocyanin levels (color pigments), and retained optimum acidity. In examining soil fertility, Costantini et al. (1996) conducted complex soil analyses in Montepulciano, Italy and found that fertile soils increased yield and berry weights, while infertile soils were detrimental to yield and berry weights. Both situations produced lower quality wine while intermediate soil types provided optimum yields and much better quality. Tesic et al. (2002a, b) also found that increased vegetative growth, mostly attributed to fertile soils, was associated with late phenology in Cabernet Sauvignon grapevines in New Zealand. Long-term trend analyses of multiple phenological events for winegrapes are scarce. In one of the few analyses, Braslavska (2000), studying grapevine phenology of the Müller-Thurgau variety in Slovakia from 1971-2000, found no trend in the dates of bud break and first leaf, however flowering, véraison, and harvest dates were earlier by eight, 11, and 15 days, respectively during the time period. In addition, the length of time from bud break to harvest in Dolné Plachtince, Slovakia declined by 15 days over the period and was related to an increase in degree-day accumulation and the relative number days with Tmax >25°C. Jones and Davis (2000b), analyzing bud break, flowering, véraison, and harvest dates from 1952-1997 for Merlot and Cabernet Sauvignon in Bordeaux, France, found that a high positive correlation exists between each successive event. The results show that predicting maturity dates is very possible as the season progresses from one phenological event to the next.

534

Phenology: An Integrative Environmental Science

The growing season averaged 193 days (from bud break to harvest) and declined by 13 days over the period (each of the interphases declined in length from 4-10 days). Climate parameters influencing the timing of the events indicate that precipitation and general cloudiness tend to delay events, while increased insolation and the relative number of days over 25 and 30°C hasten the events. In an integrated synoptic air mass and circulation analysis, Jones and Davis (2000a) related the mean climate characteristics, phenology, yield, and wine quality in Bordeaux to variations in air mass and circulation frequencies. The results indicate that increased frequencies of cold- and moisture-producing events (e.g., low pressure and frontal passages) during bud break, flowering, and berry set both delay vine phenology and reduce yield and quality. The delayed phenology in Bordeaux also showed strong correlations with sugar and acid levels (which largely determine quality) as delayed events resulted in higher acid levels, while early events resulted in greater sugar levels. Likewise, nearly half of the variation in wine quality (as told by vintage ratings) was related to earlier phenological timing and shorter stages between phenological events.

4.3

Climate, Climate Change, and Phenology

Long-term historical records of European viticulture have been maintained for nearly a thousand years with harvest dates and yield being observed and recorded initially by the monks during the Middle Ages and later by the merchants and the prominent producers. The records indicate that the region has experienced wide fluctuations of climate and viticultural productivity (Ladurie 1971; Penning-Roswell 1989). Gladstones (1992) depicts climatic variability and viticulture t in Europe from historical records and shows that, during the medieval “Little Optimum,” temperatures were about 1°C above the present day in Europe and vineyards were planted over most of southern England and along the coasts of the North and Baltic Seas. During the “Little Ice Age” vineyards throughout the British Isles and northern Europe died out and harvests did not occur for many years in southern Europe and the Mediterranean. Pfister (1988), using the recorded dates of harvests and other vine developmental stages to study the direct effects of climate variability on viticulture in Europe from the Middle Ages to 1860, inferred that temperatures during the growing season in the High Middle Ages must have averaged 1.7°C warmer than today, and that harvest dates began around the first of September compared to early to mid October today. These grape harvest records correlated well with other long-term records of climate change as evidenced in glacial advancements and retreats, ice core analyses, palynological studies, varve chronologies, and dendrochronologies.

Chapter 7.5: Winegrape Phenology

535

In light of the growing interest in climate change, Lough et al. (1983) looked at scenarios of a warmer world in Europe using the noted 20th century warming as an analog. The authors show that the length of the growing season should expand over all of Europe, with precipitation increasing in the north and decreasing in the south. Using mean climate values and Broadbent’s (1981) compilation of vintage ratings for Bordeaux and Champagne, the authors note that the climate variables explain 58 percent and 63 percent of the vintage ratings, respectively and suggest that warmer conditions would lead to improvements in wine quality. Examining climate change adaptation in phenology and yield between varieties and viticultural regions in Europe, Kenny and Harrison (1992) find that northern regions might once again become grape growing areas, while the southern regions may become too hot to produce high quality wines. In addition, Bindi et al. (1996) compared different models of future climate change for Italy and found a composite 23 day reduction in the interval from bud break to harvest for Cabernet Sauvignon and Sangiovese grapes and attributed the changes to elevated CO2 levels and temperature increases. In California’s premier wine producing areas of Napa and Sonoma, Nemani et al. (2001) found that that higher yields and quality were related to an asymmetric warming (greatest warming at night and in the spring), a reduction in frost occurrence, an advanced start to the growing season, and an increase in the growing season length.

5.

FUTURE MONITORING AND APPLICATIONS

Although it has become more refined, the major drawback to grapevine phenological research (many other systems as well) has always been the temporal resolution of the observed events. On an annual basis, phenological research has been hampered by incomplete records caused by the death or translocation of the observer, lack of funding, and/or lack of interest. On a seasonal basis, 1-3 macro-stages are generally the most observed events leaving the vast majority of the plant system’s physiology unaccounted for. Results from the few long-term studies t of observed grapevine phenology have shown a tendency toward earlier phenological events and a shortening of the interphases between events. In a novel use of grapevine phenology, Tesic et al. (2002a, b) used indices of vine precocity (flowering) for Cabernet Sauvignon to characterize viticultural environments and developed a site index in New Zealand, which appears to have potential use in vineyard zoning, site assessment, and site selection. In addition, Souriau and Yiou (2001), using grape harvest dates from northeastern France and Switzerland, showed significant correlations between harvest dates and the North Atlantic

536

Phenology: An Integrative Environmental Science

Oscillation (NAO) and suggested using the record “as an interesting proxy” to reconstruct the NAO back in time. To facilitate future use of grapevine phenology, this overview suggests that viticulturalists adopt and use the Modified E-L system. The system is readily accepted by the industry in Australia and the United States and refines the BBCH scheme used in Europe by reducing errors, removing terms that are not universally known, and develops a continuous sequence of stages1. In addition, it is suggested that users adopt observational practices that allow for inter- and intra-regional analyses of grapevine phenology. These include, 1) identifying uniform blocks of the same varieties to observe; 2) select for long term monitoring a collection of representative vines; 3) observe only typical shoot development; 4) consistently record macro-stages when 50% of the marked vines/shoots exhibit the event (compare with Figure 1 for consistency); 5) log into a record book or computer spreadsheet as the calendar date and/or the numerical day of the year. By applying these practices, growers and researchers can begin to develop databases and phenological calendars that can be incorporated into a vineyard management plan or larger research agenda. Over time, the shared results will facilitate greater understanding of the regional variations in phenology and provide the baseline from which better plant growth modeling can be developed.

NOTES 1

Because the BBCH system is gaining use in Europe by pesticide manufacturers it may be necessary to convert between the BBCH and Modified E-L system. Conversions between each of the four systems that have been developed can be found in Coombe (1995).

REFERENCES CITED Amerine, M. A., and A. J. Winkler, Composition and quality of musts and wines of California grapes, Hilgardia, 15, 493-675, 1944. Amerine, M. A., H. W. Berg, R. E. Kunkee, C. S. Ough, V. L. Singleton, and A. D. Webb, The Technology of Wine Making, (4th ed.), AVI Publishing Company, Inc., Westport, Connecticut, 795 pp., 1980. Baillod, M. and M. Baggiolini, Les stades repères de la vigne, Revue Suisse de Viticulture, Arboiculture et Horticulture, 25, 7-9, 1993. Baggiolini, M., Les stades repères dans le développement annuel de la vigne el leur utilisation practique, Revue Romande d’Agriculture de Viticulture et d’Arboriculture, 8, 4-6, 1952. Barbeau, G., C. Asselin, and R. Morlat, Estimate of the viticultural potential of the Loire Valley “terroirs” according to a vine’s cycle precocity index, Bulletin de L’O.I.V., 805806, 247-262, 1998a.

Chapter 7.5: Winegrape Phenology

537

Barbeau, G., R. Morlat, C. Asselin, A. Jacquet, C. and Pinard, Behaviour of the cabernet franc grapevine variety in various terroirs of the Loire Valley. Influence of the precocity on the composition of the harvested grapes for a normal climatic year (example of the year 1988), Journal International des Sciences de la Vigne and du Vin, 32(2), 69-81, 1998b. Bindi, M., L. Fibbi, B. Gozzini, S. Orlandini, and F. Miglietta, Modeling the impact of future climate scenarios on yield and yield variability of grapevine, Clim. Res., 7, 213-224, 1996. Branas, J., Viticulture, Dehan, Montpellier, 990 pp., 1974. Braslavska, O., Tendencies and trends in the Grapevine Growing Season at the Locality Dolne Plachtince in 1971-2000, Národný Klimatický Program SR, 8, 69-78, 2000. Broadbent, M., The Great Vintage Wine Book, Sotheby's, London, 432 pp., 1981. Calò, A., D. Tomasi, A. Costacurta, S. Boscaro, and R. Aldighieri, The effect of temperature thresholds on grapevine (Vitis sp.) bloom: an interpretive model, Rivesta Viticultura Enologia, 1, 3-14, 1994. Clingeleffer, P. R., K. J. Sommer, M. Krstic, G. Small, and M. Welsh, Winegrape crop prediction and management, Australian and New Zealand Wine Industry Journal, 12(4), 354-359, 1997. Coombe, B.G., Adoption of a system for identifying grapevine growth stages, Australian J. Grape and Wine Research, 1, 104-110, 1995. Coombe, B. G., and P. R. Dry, Viticulture, Volume 1 – Resources (211 pp.) and Volume 2 – Practices (376 pp.), Winetitles, Adelaide, Australia, 1988. Costantini, E. A. C., P. G. Arcara, P. Cherubini, F. Campostrini, P. Storchi, and M. Pierucci, Soil and climate functional characters for grape ripening and wine quality of "vino nobile di montepulciano,” Acta Horticulturae, 427, 45-56, 1996. de Blij, H. J., Geography of viticulture: rationale and resource, J. Geography, 112-121, 1983. Due, G., M. Morris, S. Pattison, and B. G. Coombe, Modeling grapevine phenology against weather: considerations based a large data set, Agricult. Forest Meteorol., 56, 91-106, 1993. Eichhorn, K. W., and D. H. Lorenz, Phönologische entwicklungsstadien der rebe, Nachrichtenblatt des Deutschen Pflanzenschutzdienstes (Braunschweig), 29, 119-120, 1997. Galet, P., A Practical Ampelograph, Comstock Publishing, London, 248 pp., 1979. Giomo, A., P. Borsetta, and R. Zironi, Grape quality: research on the relationships between grape composition and climatic variables, in Proceedings of the Workshop Strategies to Optimize Wine Grape Quality, edited by S. Poni, E. Peterlunger, F. Iacono, and C. Intrieri, 427 (pp. 277-285), Acta Horticulturae, 1996. Gladstones, J., Viticulture and Environment, Winetitles, Adelaide, 310 pp., 1992. Huglin, P., Nouveau mode d'evaluation des possibilities heliothermiques d'un milieu viticole, C. R. Academy of Agriculture in France (1117-1126), 1978. Jackson, D. I., and N. J. Cherry, Prediction of a district's grape-ripening capacity using a latitude-temperature index, Amer. J. Enology and Viticulture, 39(1), 19-28, 1988. Jackson, D. I., and P. B. Lombard, Environmental and management practices affecting grape composition and wine quality – a review, Amer. J. Enology and Viticulture, 44(4), 409430, 1993. Jackson, D., and D. Schuster, The Production of Grapes and Wine in Cool Climates, Buttersworths Horticultural Books, New Zealand, 208 pp., 1987. Jimenez, J., and V. Ruiz, Phenological development of Vitis vinifera L. in Castilla – La Mancha (Spain), A study of 21 cultivars (10 red and 11 white cultivars), Acta Horticulturae, 388, 105-110, 1995.

538

Phenology: An Integrative Environmental Science

Jones, G. V., A Synoptic Climatological Assessment of Viticultural Phenology, Ph.D. dissertation, Department of Environmental Sciences, University of Virginia, 1997. Jones, G. V., and R. E. Davis, Using a synoptic climatological approach to understand climate/viticulture relationships, Int. J. Climatology, 20, 813-837, 2000a. Jones, G. V., and R. E. Davis, Climate influences on grapevine phenology, grape composition, and wine production and quality for Bordeaux, France, Amer. J. Enology and Viticulture, 51(3), 249-261, 2000b. Kenny, G. J., and P. A. Harrison, The effects of climate variability and change on grape suitability in Europe, J. Wine Research, 3, 163-183, 1992. Ladurie, E. L. R., Times of Feast, Times of Famine: A History of Climate Since the Year 1000, Doubleday and Company, Inc., New York, 426 pp., 1971. Lorenz, D. H., K. W. Eichhorn, H. Bleiholder, R. Klose, U. Meier, and E. Weber, Phenological growth stages of the grapevine (Vitis vinifera L. ssp. vinifera)--Codes and descriptions according to the extended BBCH scale, Australian J. Grape and Wine Research, 1, 91-103, 1995. Loubere, L. A., The Wine Revolution in France, Princeton University Press, New Jersey, 288pp., 1990. Lough, J. M., T. M. L. Wigley, and J. P. Palutikof, Climate and climate impact scenarios for Europe in a warmer world, J. Climate and Appl. Meteorol., 22,1673-1684, 1983. Mansson, P.-H., “Wine and Health”, “The Case for Red Wine”, and “Wine and Health Trailblazers”, The Wine Spectator, pp. 32-65, Dec. 15, 2001. Matthews, M. A., M. M. Anderson, and H. R. Schultz, Phenologic and growth responses to early and late season water deficits in Cabernet Franc, Vitis, 26, 147-160, 1987. McIntyre, G. N., L. A. Lider, and N. L. Ferrari, The chronological classification of grapevine phenology, Amer. J. Enology and Viticulture, 33(2), 80-85, 1982. McIntyre, G. N., W. M. Kliewer, and L. A. Lider, Some limitations of the degree day system as used in viticulture in California, Amer. J. Enology and Viticulture, 38(2), 128-132, 1987. Meier, U., Growth stages of mono- and dicotyledonous plants, BBCH-Monograph, Blackwell Wissenschafts-Verlag, Berlin, 622 pp., 1997. Moncur, M. W., K. Rattigan, D. H. MacKenzie, and G. N. McIntyre, Base temperatures for budbreak and leaf appearance of grapevines, Amer. J. Enology and Viticulture, 40(1), 2126, 1989. Mullins, M. G., A. Bouquet, and L. E. Williams, Biology of the Grapevine, Cambridge University Press, Great Britain, 239 pp., 1992. Nemani, R. R., M. A. White, D. R. Cayan, G. V. Jones, S. W. Running, and J. C. Coughlan, Asymmetric climatic warming improves California vintages, Clim. Res., 19(1), 25-34, 2001. Penning-Roswell, E., Wines of Bordeaux, (6th ed.), Penguin Books, London, 610 pp., 1989. Pfister, C., Variations in the spring-summer climate of central Europe from the High Middle Ages to 1850, in Long and Short Term Variability of Climate, edited by H. Wanner and U. Siegenthaler, pp. 57-82, Springer-Verlag, Berlin, 1988. Prescott, J. A., The climatology of the vine (Vitis vinifera L.) the cool limits of cultivation, Transcriptions of the Royal Society of Southern Australia, 89, 5-23, 1965. Riou, C., The Effect of Climate on Grape Ripening: Application to the Zoning of Sugar Content in the European Community, Commission Europeenne, Luxembourg, 319 pp., 1994. Robinson, J., The Oxford Companion to Wine, (1st ed.), Oxford University Press, New York, 1088 pp., 1994.

Chapter 7.5: Winegrape Phenology

539

Smart, R. E., and P. R. Dry, A climatic classification for Australian viticultural regions, Australian Grapegrower and Winemaker, 196, 8-16, 1980. Souriau, A., and P. Yiou, Grape harvest dates for checking NAO paleoreconstructions, Geophysical Research Letters, 28(20), 3895-3896, 2001. Swanepoel, J. J., F. S. de Villiers, and R. Pouget, Predicting the date of bud burst in grapevines, South African J. Enology and Viticulture, 11(1), 46-49, 1990. Tesic, D., D. J. Woolley, E. W. Hewett, and D. J. Martin, Environmental effect on cv Cabernet Sauvignon (Vitis Vinifera L.) grown in Hawkes Bay, New Zealand, 1. Phenology and characterization of viticultural environments, Australian J. Grape and Wine Research, 8, 15-26, 2002a. Tesic, D., D. J. Woolley, E. W. Hewett., and D. J. Martin, Environmental effect on cv Cabernet Sauvignon (Vitis Vinifera L.) grown in Hawkes Bay, New Zealand, 2. Development of a site index, Australian J. Grape and Wine Research, 8, 27-35, 2002b. Unwin, T., Wine and the Vine: A historical Geography of Viticulture and the Wine Trade, Routledge, London, 409 pp., 1991. Winkler, A. J., J. A. Cook, W. M. Kliewere, and L. A. Lider, General Viticulture, (4th Ed.), University of California Press, Berkley, 740 pp., 1974. Wolf, T. K., Site Selection for Commercial Vineyards, Virginia Agricultural Experiment Station, Winchester, Virginia, Publication Number 463-016, 9 pp., 1997. Zalom, F. G., P. B. Goodell, W. W. Wilson, and W. J. Bentley, Degree-Days: the calculation and the use of heat units in pest management, Leaflet no. 21373, Division of Agriculture and Natural Resources, University of California, Davis, 12 pp., 1993.

Chapter 7.6 LONG-TERM URBAN-RURAL COMPARISONS Claudio Defila and Bernard Clot Biometeorology, MeteoSwiss, Zürich and Payerne, Switzerland

Key words:

1.

Urban, Rural, Climate change, Trends, Heat island effect

PHYTOPHENOLOGICAL OBSERVATIONS IN SWITZERLAND

Since earliest history, humans have applied themselves to phenology. Hunters and gatherers looking for food had every interest to know where, for example, they could find ripe berries. At a later date, farmers who cultivated the land planned their work such as the sowing, planting or harvesting of crops according to the recurrent cycles of growth and development of plants. Numerous accounts dating from the Middle Ages document extraordinary phenological occurrences. They were observed and recorded by scholars, mainly residents of monasteries. Such records offer valuable insight into climatic conditions at a time when measuring instruments were not yet in use (Pfister 1984). The first phenological observation network in Switzerland dates back to the year 1760. Unfortunately, these observations were discontinued after a short while. Between 1869 and 1882, forest phenological observations were carried out in the canton of Berne. Since 1808 the leaf bud burst of horse chestnut trees has been observed and recorded every year in the city of Geneva and the full flowering of cherry trees in Liestal since 1894. In this publication, we will present and discuss these two unique and valuable phenological series at length (Figure 1).

Schwartz (ed.), PHENOLOGY: AN INTEGRATIVE ENVIRONMENTAL SCIENCE, 541-554 © 2003 Kluwer Academic Publishers. Printed in the Netherlands.

542

Phenology: An Integrative Environmental Science

Figure 7.6-1. Map of Switzerland and locations mentioned in the text.

In 1951 the Swiss national phenological observation network was founded. It is operated by MeteoSwiss, the national weather service. Presently, 26 different plant species and 69 phenophases are observed and recorded at approximately 160 observation posts in all regions and altitudes of Switzerland (200 to 1800 m above sea level). Forty selected observation posts report 17 phenophases immediately on their appearance. Based on this current information, MeteoSwiss issues weekly phenological bulletins during the vegetation period that are published on the Internet. In 1958, the International Phenological Gardens (IPG) were founded (Schnelle 1958). Various stations throughout Europe observe genetically identical plant material (shrubs and trees): one of them is located in Birmensdorf, near Zurich (Swiss Federal Institute for Forest, Snow and Landscape Research). The Geographical Institute of the University of Berne has maintained a phenological observation network since 1970 (Jeanneret 1996): a few phenological phases are observed in an area extending from the Jura over the Mittelland (midlands) to the Berner Oberland (Bernese Oberland). Since 1994, phenological observations have also been carried out in the Swiss National Park. Included in this phenological program are some Alpine plants, which grow at altitudes where there are no human settlements and little if any human interference. In the year 2001, 15 forest phenological stations were added to the national phenological network. So phenology has a long tradition in Switzerland, where phenological research is made particularly interesting by the orographic structure with its diverse climatic regions and altitudes.

Chapter 7.6: Long-Term Urban-Rural Comparisons

1.1

543

Historical Phenological Time Series

The historical forest phenological observations in the canton of Berne were carried out by its forestry service from 1869 to 1882; the collected data are still available. They were evaluated and compared to current phenological observations (Vassella 1997). The results showed that in the Bernese series comparable phenophases occurred later in the season than in the recent MeteoSwiss observations. However, such a comparison is not unproblematic since we do not know the methods of observation employed at that time. In addition, the historical observations were made in the forest, whereas today the observer views the forest from the outside. Since 1808, civil servants in Geneva have been observing the leaf bud burst of horse-chestnut trees in their city. This date is of great importance since it marks the official beginning of spring in the city of Geneva. Accordingly, the local media spread the news every year. In the canton of Baselland (Liestal), the full flowering of cherry trees has been recorded since 1894. Baselland is one of the main areas for the cultivation of cherries in Switzerland. The date is important for a forecast of the harvest date which takes into account the flowering date and temperature data. The phenological series of the national network operated by MeteoSwiss have an almost historical dimension as they cover a period of more than 50 years.

1.2

Phytophenological Trends in Switzerland 1951-1998

896 phenological time series, carried out at 68 observation stations in different regions and altitudes of Switzerland and covering 19 different phenophases were examined (Defila and Clot 2001). Only time series of a minimum duration of 20 years were evaluated. From the 896 tested phenological time series, 269 (30%) show a significant trend. Out of these, 98 (36%) show a positive trend (towards later appearances) and 171 (64%) a negative trend (towards earlier appearances). In almost all the phenophases positive and negative trends occur. Thus there are almost no phenophases that show only a tendency to earlier or later appearance dates at all the stations. The distribution of the positive and negative trends however is uneven. In the case of wild species, the biggest share of the negative trends (early appearance date) is found for the flowering of the hazel. This phenophase occurs at an early stage of the year (partly already in January). As the occurrence of the phenophases depends greatly on the air temperature (Defila 1991), this result is an indication of the mild winter temperatures of recent years. The biggest share of the positive trends (late appearance date) is found for the coloring of leaves of the horse chestnut.

544

Phenology: An Integrative Environmental Science

Considerable seasonal and regional differences can be observed. Figure 2 presents the phenological spring (430 phenophases), summer (119 phenophases) and autumn (230 phenophases) phases from all the regions North of the Alps. In spring and summer a distinct predominance of the negative trends (early appearance date) can be identified whereas the share of positive and negative trends is similar in autumn. On the contrary, in the Rhone Valley and the South side of the Alps the positive trends (late appearance date) predominate in summer and autumn whereas the share of positive and negative trends is equal in spring. Figure 3 presents the phenological spring (80 phenophases), summer (15 phenophases) and autumn (22 phenophases) phases from these two regions. From both Figures 2 and 3, it is also clear that the overall share of significant trends is higher in the Rhone Valley and the South side of the Alps than North of the Alps. Taking into consideration all stations and phenophases showing a significant trend in Switzerland, the early appearance dates in spring (- 11.6 days) and late appearance dates in autumn (+ 1.7 days) result in a prolongation of the vegetation period of 13.3 days within 50 years (1951-2000) or 0.27 days a year. These results are comparable with those from the IPG (Menzel and Fabian 1999).

Figure 7.6-2. Share of significant negative (towards earlier appearance, white bars) and positive (towards later appearance, black bars) trends in regions North of the Alps.

Chapter 7.6: Long-Term Urban-Rural Comparisons

2.

545

INFLUENCE OF THE CITY EFFECT (HEAT ISLAND) ON PHENOLOGY

The city effect is well known in climatology. As a result of the emissions from heating systems, traffic and industry, and of the differences in the absorption and radiative properties of the surfaces within the urban areas, compared to the rural (Oke 1982), cities of a certain size are two to three degrees warmer than their surroundings. This leads to so-called “heatislands” in middle-sized or large cities. Since phenological spring phases are predominantly influenced by temperature (Defila 1991; D’Odorico et al. 2002), an earlier development of vegetation can be expected in urban areas

Figure 7.6-3. Share of significant negative (towards earlier appearance, white bars) and positive (towards later appearance, black bars) trends in the Rhone Valley and South side of the Alps.

as opposed to rural areas. The city effect is especially felt in the winter months (heating period). The very early phenophases like the full flowering of the hazel are particularly influenced by winter temperatures (January, February). This city effect can easily be observed every year with the full flowering of the forsythia. The brightly colored yellow petals of the forsythia are very striking at a time when vegetation is still scarce. Depending on weather conditions, the flowering of the forsythia takes place some days earlier in the city than in the countryside. This influence of the city effect on phenology has been studied for different regions in Central Europe (Roetzer et al. 2000).

546

Phenology: An Integrative Environmental Science

In this work, an earlier development of four days was established for the very early spring phases in cities as opposed to rural areas. For later spring phases, the earlier development in cities still amounted to two days. On the other hand, for the period of 1980 to 1995 a stronger trend towards earlier appearance was found for rural areas than for cities. Similar studies carried out in the agglomeration of Zurich, Switzerland, show that, depending on plant species and phenophases, the city effect produces different results (Defila 1999). In addition, the influence of microclimate has to be considered. In general terms, the influence of the heat island in cities on spring and summer phases has been substantiated. In the case of phenological autumn phases, the city effect could not be established, since autumn phases do not exclusively depend on temperature.

2.1

Urban Environment: Leaf Bud Burst of the HorseChestnut in Geneva, 1808-2002

The appearance dates of horse-chestnut bud burst in Geneva from 1808 to 2002 are shown in Figure 4. This graph shows a clear trend towards earlier appearance dates (0.24 days per year, P<10-10) and this becomes much more pronounced from about the year 1900. The latest bud burst date

Figure 7.6-4. Dates of the leaf bud burst of the horse chestnut in Geneva 1808-2002. Smoothing: Gauss low-pass filter with a period of 20 years (Extended from Defila and Clot 2001).

Chapter 7.6: Long-Term Urban-Rural Comparisons

547

was in 1816 on 23rd April, the earliest was in 1991 on 3rd January. If the linear trends are divided into periods of 50 years each, the beginning of a marked earlier appearance can be distinguished more easily (Figure 5). During the periods from 1808 to 1850 (this period consists only of 43 years) and 1851 to 1900 no significant trend appears. However, for the periods from 1901 to 1950 and from 1951 to 2000, the trends are evident and highly significant (P<0.005). This coincides with the growth and the industrial development of Geneva. The advent of cars also took place during the 20th century. Over the period 1951-2000, the earlier appearance amounts to 0.45 days a year. Overall (1808 to 2002), the earlier occurrence amounts to 47 days, a result which can be considered extreme.

Figure 7.6-5. Trends on 50 years periods of the leaf bud burst of the horse chestnut in Geneva.

2.2

Rural Environment: Flowering of Cherry Trees in Liestal, 1894-2002

The flowering of the cherry trees in rural Liestal from 1894 to 2002 (Figure 6) does not show such a strong trend (0.08 days per year, P=0.013). The latest appearance date was observed on 4th May 1917, the earliest on 16th March 1990. Considering the two periods of 50 years, 1901 to 1950 and 1951 to 2000, the analysis of the linear trends show substantial differences (Figure 7). In the first period no trend is observed. In the following period (1951 to 2000) a clear significant trend towards earlier appearance is

548

Phenology: An Integrative Environmental Science

revealed. For this period (1951-2000) an earlier occurrence of 0.24 days/year can be calculated. The influence of the years after 1985 on this result is evident from Figure 6. This is still considerably less than the result shown for the leaf bud burst of horse-chestnut trees in Geneva (0.45 days per year). The overall earlier appearance for the period from 1894 to 2002 amounts to nine days.

Figure 7.6-6. Dates of the full flowering of cherry trees in Liestal 1894-2002. Smoothing: Gauss low-pass filter with a period of 20 years (Extended from Defila and Clot 2001).

This evaluation suggests clear evidence that in rural areas the tendency for earlier appearance is less marked than in urban environments. It seems very likely that the city climate contributes to a great extent to the earlier appearance of the leaf bud burst of horse-chestnut trees in urban Geneva. Accordingly, the influence of global warming on the above mentioned phenophase in Geneva amounts to less than 0.45 days per year. It has to be noted, however, that these two phenological time series are not entirely comparable. The leaf bud burst of horse chestnut trees in Geneva happens considerably earlier in the year than the full flowering of cherry trees in Liestal. Therefore, the first phenophase depends much more on the winter climate than the second one, at a time when domestic heating will lead to higher winter temperatures in the city.

Chapter 7.6: Long-Term Urban-Rural Comparisons

549

Figure 7.6-7. Trends on 50 years periods of the full flowering of cherry trees in Liestal.

2.3

Comparison of Temperature Time Series with Phenological Time Series in an Urban and Rural Region

In this chapter we would like to discuss the comparison of temperature conditions in urban and rural environments and their relevant influence on phenology. It is well known that, over the year, temperatures in larger cities are on average two or three degrees higher than those in the rural areas surrounding them. This city effect orr heat island is produced by heat emission from traffic, industry and heating and by the particular microclimate in cities (densely populated areas). The city climate has some influence on phenological appearance dates, since they depend strongly on temperature (Defila 1991). It has already been shown that vegetation development in cities occurs earlier than in rural areas (Defila 1999; Roetzer et al. 2000). In our study, a marked increase in temperature since 1808 was expected for the city of Geneva. Climatic conditions there are not entirely equivalent to city conditions; however, it is not a rural environment either. From 1864 to 1961 air temperature was measured in the city of Geneva (GenèveObservatoire). Since 1952 the climate station has been located at Geneva airport (Genève-Cointrin). In order to evaluate the climatological time series

550

Phenology: An Integrative Environmental Science

of the periods from 1864 to 2000, air temperatures have been homogenized for the city station on the one hand and for the airport station on the other (Begert et al. 1998). In Figure 8, both time series of mean annual values

Figure 7.6-8. Mean annual air temperature at the Geneva-Observatory (upper line) and Geneva-Cointrin (lower line) 1864-2000 (20-year Gaussian filter).

are displayed using the Gaussian 20-year filter. Both show the same kind of development, but throughout the period, the temperature curves are slightly higher in the city than at the airport (about 0.7°C). This small difference suggests that the station of Geneva-Cointrin (airport) is probably partly included in the heat island of the city. A very similar curve can also be displayed for the somewhat rural and slightly colder station BaselBinningen, over the same period (Figure 9). Basel-Binningen (the climate station situated closest to Liestal) is part of the agglomeration of Basel. However, the trends of annual temperature are almost the same in the two cities, and fail to explain the observed differences in the trends between the leaf bud burst of the horse-chestnut in Geneva and the full flowering of cherry trees in Liestal. As the temperatures of the month preceding the spring phenophases are decisive, the mean monthly temperatures are displayed in Figures 10 and 11. For the leaf bud burst of the horse-chestnut trees in Geneva, the mean February temperature of the city station is shown and for the full flowering of the cherry trees in Liestal the mean March-April temperatures of the climate station Basel-Binningen. Both temperature series show an overall significant trend towards higher temperatures in the period from 1864 to 2000 (F-test, P=0.03 in Geneva and P=0.00004 in Basel), which are responsible for the tendency to the earlier appearance of phenophases.

Chapter 7.6: Long-Term Urban-Rural Comparisons

551

Figure 7.6-9. Mean annual air temperature of Basel-Binningen 1864-2000. Smoothing: 20year Gaussian filter.

Figure 7.6-10. February mean temperature of Geneva-Observatory, 1864-2000. Smoothing: 20-year Gaussian filter.

552

Phenology: An Integrative Environmental Science

Figure 7.6-11. March/April mean temperature of Basel-Binningen 1864-2000. Smoothing: 20-year Gaussian filter.

During the years 1901 to 1950, the increase of March-April temperatures in Basel amounts to 0.033°C/year, and the corresponding increase of February temperatures in Geneva are not significant. However, over 19512000, when in Basel the March-April temperature increase reaches only 0.02°C/year, the increase of February temperatures in Geneva amounts to 0.045°C/year. These values match quite well with the strong trend observed towards earlier appearance dates of the leaf bud burst of the horse-chestnut trees in Geneva, and the slighter trend of the full flowering of the cherry trees in Liestal. However, the same reservation about the conclusions on the causes of the differences observed between Basel and Geneva in the previous section have to be restated. The available phenological and meteorological data do not allow us to definitively conclude. A part of the causes of these differences is likely to be found in the contribution of the city effect to the stronger reaction of the horse-chestnut trees in Geneva, but one cannot forget that the climate change has also been less important in March and April than in January and February.

3.

CONCLUSIONS

The process of climatic warming during recent decades has had an effect on phytophenological commencement dates in Switzerland. In the period

Chapter 7.6: Long-Term Urban-Rural Comparisons

553

from 1951 to 1998 an earlier start of vegetation growth of roughly 11 days was recorded. In autumn, however, a slightly later commencement (approx. two days) of phenological phases has been noted. Depending on region and season, plants react in different ways. North of the Alps, a more marked trend towards earlier commencement in spring has been observed, whereas in the Rhone valley and South of the Alps, delayed phases clearly predominate in summer and autumn. Since phenological spring phases are to a great extent influenced by air temperature, it is to be expected that the two historical phenological time series in Switzerland should show a significant trend. These time series are the leaf bud burst of horse chestnut in Geneva (1808-2002), and the full flowering of cherry trees in Liestal (1894-2002). Regarding the Geneva observation post, the city influence (heat island) has to be taken into consideration as well. Homogenized temperature series for the city of Geneva and Geneva airport (rural environment) show that temperatures in the city are approx. 0.7°C higher than in the countryside. The city of Geneva has grown considerably since 1808 and heat emission by traffic, industry, and domestic heating has increased accordingly. Both temperature series show a clear tendency to higher values. For the rural phenological station Liestal (full flowering of cherry trees) the temperature series of the climate station Basel-Binningen has been used. Here, too, a trend towards higher values can be observed. Both phenological time series show a tendency to earlier commencement, whereby this tendency is much more marked with the leaf bud burst of horse chestnut. From this it can be concluded that in both phenological time series general warming has led to a trend towards earlier commencement dates. In the case of the leaf bud burst of horse-chestnut in Geneva, this effect has been reinforced by the city climate, whereas the influence of other factors such as differential rates of warming between the winter and spring months could also have played a role in the observed differences between this phase and the cherry trees flowering in Basel.

ACKNOWLEDGEMENTS Thanks to S. Bader and M. Begert for their help in preparing the figures.

554

Phenology: An Integrative Environmental Science

REFERENCES CITED Begert, M., M. Giroud, R. Kegel, G. Seiz, V. Koehli, C. Haeberli, O. Bochnicek, M. Fukasz, E. Nieplova, and L. Sramo, Operational homogenization of long term climate data series at SMI and SHMI, Proceedings 2nd European Conference on Applied Climatology, pp. 1923 October 1998, Vienna, Austria, 1998. Defila, C., Pflanzenphänologie der Schweiz, Dissertation, University of Zürich, Veröff. Schweiz. Meteorologischen Anstalt, 50, 1-235, 1991. Defila, C., Der Einfluss des Stadtklimas auf die phänologischen Eintrittstermine, Schweiz. Z. Forstwes., 150, 151-153, 1999. Defila, C., and B. Clot, Phytophenological trends in Switzerland, d Int. J. Biometeorol., 45, 203-207, 2001. D’Odorico, P., J. Yoo, and S. Jäger, Changing seasons: An effect of the North Atlantic Oscillation?, J. Climate, 15, 435-445, 2002. Jeanneret, F., Phänologie in einem Querschnitt durch Jura, Mittelland und Alpen, Jb. Gg. Ges. BE, 59, 159-203, 1996. Menzel, A., and P. Fabian, Growing season extended in Europe, Nature, 397, 659, 1999. Oke, T.R., The energetic basis of the urban heat island, Quart. J. Royal Met. Soc., 108, 1-24, 1982. Pfister, C., Klimageschichte der Schweiz 1525 bis 1860, Das Klima der Schweiz 1525 bis 1860 und seine Bedeutung in der Geschichte von Bevölkerung und Landwirtschaft, t Band 1, Haupt, Bern, 184 pp., 1984. Roetzer, T., M. Wittenzeller, H. Haeckel, and J. Nekovar, Phenology in Central Europe differences and trends of spring phenophases in urban and rural areas, Int. J. Biometeorol., 44, 60-66, 2000. Schnelle, F., Ergebnisse aus den Internationalen Phänologischen Gärten in Europa - Mittel 1973-82, Wetter und Leben, 38, 5-17, 1958. Vassella, A., Phänologie von Waldbäumen, BUWAL Umwelt-Materialien, 73, 9-75, 1997.

Acknowledgments

I am grateful for the help I received from many individuals as this book was constructed. I would especially like to thank my wife, Ann Lessner Schwartz, for her support and patience during this long project, and for taking the time to proofread the final manuscript. My graduate student assistant, Jeffrey Kueny, provided invaluable support, including many long hours of tedious work reformatting and checking references. The following individuals generously took the time to review one or more of the chapter manuscripts: Elisabeth Beaubien, Gregory Carbone, Xiaoqiu Chen, Frank Chmielewski, Peter Curtis, Peter Dunn, Kevin Gallo, Donn Haglund, Heikki Hänninen, Gregory Jones, Marie Keatley, Annette Menzel, Patrícia Morellato, Eric Post, Bradley Reed, Jacques Régnière, James Reinartz, Donatella Spano, Larry Tieszen, Arnold vanVliet, and Michael White. I also appreciated the assistance of editors and staff at Kluwer, including: Helen Buitenkamp, Jacco Flipsen, Deborah Doherty, Noeline Gibson, Sandra Oomkes, Lanette Setkoski, and Claire vanHeukelom. Lastly I would like to acknowledge the contributions that my high school science teacher, Manuel Thies, my master’s advisor, Jay Harman, and my doctoral advisor, Glen Marotz, made toward advancing my career as a scientist.

555

Index Aboriginal 27, 35 Advanced Very High Resolution Radiometer 367, 456, 492 Agricultural Experiment Station 57, 58, 60, 61, 63 agriculture 58, 108, 112, 113, 115, 122, 123, 144, 145, 181, 505, 506, 509-513, 529, 520 albedo 196, 339, 340, 356, 455, 456, 487-492, 494-499, 501 algorithm 129, 223, 225, 241, 243, 244, 368, 370, 373, 459, 498, 499, 501 alpine 70, 179, 185, 195, 196, 198, 199, 201-203, 208, 209, 508, 509, 542 animal 3, 5, 24, 49, 51, 52, 76, 93, 98, 107-109, 123, 126, 128, 145, 177, 179, 206-209, 237, 239, 244, 245, 320, 325, 326, 385, 386, 388, 391, 394, 395, 399, 405, 421, 454 animal ecology 206, 242, 244, 251 anthesis 162, 169, 514, 525 application 179, 226, 228, 240, 241, 244, 245, 247, 271, 316, 322, 323, 345, 351, 366, 375378, 407-409, 411, 412, 454, 463, 468, 483, 505, 511, 512, 525, 535 aquatic 5, 81, 107, 208, 387, 389, 392-394, 397 Australia 4, 27-31, 33-35, 37-40, 139, 142, 143, 145, 146, 202, 525, 528, 536 autumn 15, 16, 19, 23, 52, 100, 101, 147, 176, 182, 183, 187, 189, 228, 257-259, 261-264, 278, 279, 286, 288-293, 296,

297, 306, 312-316, 320, 322, 325-327, 337, 338, 425-427, 432, 494, 506-508, 519, 544, 546, 553 BBCH scale 99, 103, 270, 272, 273, 277, 280, 510, 525 beech 257, 258, 260-264, 266 biocalendar 178, 179 biome 158, 285, 286, 298, 299, 461 bird 3, 31, 38, 51, 63, 66, 68, 109, 113, 321, 421-433 breeding 38, 325, 422, 425, 426, 428, 429, 431-433, 437, 438, 442, 517 budburst 217, 218, 229, 321, 371, 373, 525 calendar 15, 16, 27, 36, 40, 45, 51, 52, 54, 61, 65, 71, 111, 195, 249, 303-309, 315, 323, 324, 339, 352, 406, 407, 411413, 472, 536 Canada 61-71, 94, 102, 178, 179, 181, 182, 189, 243, 245, 248, 338, 407, 409, 422, 472, 481, 488, 492 canopy 83, 124, 131, 132, 148, 159, 160, 167, 169, 228, 286, 288, 290, 291, 354, 360, 366, 375, 376, 394, 414, 457, 461463, 469-475, 479-481, 483, 484, 487-491, 494, 501, 523, 529, 530 carbon flux 341, 342, 470 caribou 437, 439, 440-442, 444447 chill date 338

557

558 chilling 184, 185, 199, 200, 220223, 225, 229, 335, 336, 513, 528 China 3, 11, 12, 14-19, 101-103, 286, 293, 338, 373 climate 3-6, 17, 20, 30, 31, 46, 47, 51, 57, 66, 70, 71, 75, 79, 83, 84, 87, 93, 94, 101, 106, 113, 121, 122, 124, 125, 139142, 144-146, 148, 149, 152, 175, 179, 180, 182, 183, 187, 195, 196, 198, 199, 209, 228, 242-245, 247, 248, 250, 265, 266, 285, 286, 295, 297-299, 301, 303, 304, 315, 316, 319, 325, 331, 332, 338, 339, 342, 345, 351, 356, 370, 388, 406, 411, 426, 429, 430, 433, 443, 444, 447, 453-456, 458, 460, 463, 468, 474, 489, 487, 496, 497, 506, 516, 524, 527, 529532, 534, 548-550, 553 climate change 5, 14, 16, 17, 20, 23, 24, 39, 52, 67, 70, 71, 93, 106, 114, 151, 152, 178, 185, 187, 189, 195, 204, 206-209, 217, 227, 228, 244, 245, 250, 251, 255, 266, 294, 299, 301, 308, 317, 319, 322-324, 326, 331, 345, 430, 431, 433, 447, 454, 468, 474, 483, 505, 518, 534, 535, 552 climate variability 151, 534 communication 94, 103, 105-107, 109-116, 270 comparison 54, 60, 68, 125, 129, 179, 187, 223, 246, 251, 272, 301, 312, 314, 316, 317, 324, 325, 333, 338, 341, 356, 393, 407, 408, 432, 441, 442, 455, 457, 463, 468, 541, 543, 549 conservation 51, 75, 106, 123, 126, 152 continental scale 60, 320 cooperation 88, 93, 94, 103, 105107, 109, 110, 113-116, 270, 342

Index crop management 505, 512, 519 curve derivative 370 daily weather generator 244 data smoothing 366 definition 5, 54, 69, 99, 100, 102, 105, 114, 159, 176, 189, 195, 272, 273, 295, 326, 346, 347, 392, 455, 461, 499, 507 degree-day 204, 219, 298 degree-hour 353 density 123, 127, 128, 139, 166169, 241, 351, 354, 356, 367, 388, 422, 437, 442-447, 470, 516 development rate 238-241, 262, 263, 353, 406 development stages 256-262, 269-271, 273, 525 drought 100, 125, 140-144, 146149, 151, 152, 158, 159, 164, 167, 168, 201, 228-230, 245, 259, 303, 325, 391, 484, 497 dry forest 75, 80, 83, 84, 87, 121132 ecology 32, 39, 47, 115, 152, 169, 206, 242, 244, 247, 251, 319, 397, 421, 442 ecosystem 47, 67, 70, 87, 107, 121-123, 129, 131, 132, 139, 140 142, 144-147, 151, 152, 158, 159, 161, 165, 169, 195197, 203, 206, 209, 242, 255, 264, 265, 298, 301, 304, 326, 341, 386, 393, 396-398, 456, 461, 463, 469-471, 474 eddy covariance 469, 470, 473 energy balance 160, 167, 340, 345, 349, 351, 354-357, 469 entomology 39, 114, 246, 406 environnment 3, 27, 51, 70, 111, 112, 114, 147, 167, 220, 267, 349, 357, 360, 376, 389, 392, 407, 468, 546, 547, 549, 553

Index Environment Canada 64, 70 environmental link 324 Europe 4, 11, 45-47, 52, 54, 97, 101-103, 105, 106, 115, 145, 151, 156, 176-180, 186, 189, 200, 245, 302, 309, 311, 316, 317, 320-322, 325, 337, 338, 342, 346, 422, 508, 509, 518520, 524, 527, 534-536, 452, 545 evolution 208, 219, 251, 439 experimental phenology 183 fetch 357 fire 14, 144, 145, 152, 158, 159, 162, 165, 166, 168, 454 first bloom 12, 16, 18, 63, 65, 66, 68, 161, 189, 286, 289, 331, 336-338, 341, 513, 520 first leaf 12, 16, 65, 99, 274, 278, 280, 286, 287, 289, 322, 331, 334-340, 378, 523 First Nations 61, 62, 70 flowering 3, 17, 18, 21-24, 2830, 32, 33, 45, 62, 67-69, 76, 77, 79-83, 85-87, 95, 96, 98100, 102, 109, 111, 1126-128, 147-151, 162-168, 176, 178182, 185, 187-189, 197, 199204, 206-209, 217-219, 224, 228, 229, 272, 275-279, 281, 285, 290, 291, 320, 321, 323325, 407, 467, 508, 509, 511518, 525, 527-529, 531-535, 541, 543, 545, 547-550, 552, 553 forcing 97, 158-160, 165 forecast 48, 111, 169, 197, 356, 505, 513, 543 forest 14, 29, 31, 33, 34, 47, 48, 50, 5,3, 54, 57, 64, 75-77, 7988, 102, 121-132, 142-146, 151, 152, 182, 224, 228, 249, 255, 257, 264, 265, 294, 338, 340, 342, 373, 375, 377, 386, 393, 394, 412, 456, 459, 461,

559 470-474, 480-482, 523, 537, 542, 543 fresh water 3, 387, 391, 394 frost damage 112, 183-185, 204, 226, 227, 266, 337, 408, 505, 513, 520 frost date 303, 337, 338 frost hardiness 226-229 fruit trees 47-50, 53, 60, 78, 96, 226, 413, 505, 511, 514, 520 fruiting 27, 28, 32, 67, 76, 77, 79, 81, 83, 85-87, 100, 128, 147, 152, 199, 201, 229, 322, 391 global 4, 6, 20, 23, 24, 37, 93, 94, 96, 100, 103, 130, 131, 151, 185, 187, 197, 204, 207, 217, 225, 226, 230, 245, 251, 299, 323, 326, 331-333, 342, 370, 377, 385, 393, 395, 397, 399, 432, 454-456, 468-471, 474, 483, 487, 532 global change 5, 6, 47, 60, 94, 106, 169, 177, 195, 319, 322, 325, 365, 378, 453, 454, 456, 460, 461, 463 gradient 158, 163, 164, 184, 198201, 209, 242-245, 249, 312, 321, 356, 360 grapevine 50, 321, 322, 523-533, 535, 536 graphic design 305 grassland 5, 80, 84, 145, 157161, 164, 165, 167, 168, 196, 342, 377, 455, 456, 459, 488, 492-496, 510 Great Plains 5, 157, 158, 164, 165, 169, 456 greenness 366, 367, 370, 372, 376, 483, 492, 494 growing season 19, 46, 61, 63, 96, 98, 108, 159, 160, 163, 164, 168, 175, 182, 183, 187, 189, 195, 196, 198-204, 220, 228, 229, 233, 265, 285, 286, 293-299, 303, 315, 316, 322,

560 325-327, 337, 369-371, 373, 375, 378, 411, 442, 461, 472, 474, 475, 477, 479-481, 483, 484, 492, 496-498, 506-509, 512, 514, 515, 518, 519, 531, 532, 534, 535 growing season length 210, 213, 368, 369, 375, 455, 457, 462, 463, 474, 480, 481, 505, 507509, 519, 520, 535 growth 3, 5, 60, 98, 99, 107, 108, 112, 122, 146, 147, 151, 152, 157, 160, 162, 165-168, 175, 177, 179, 183, 185, 187, 199201, 204, 206, 213, 223, 224, 226, 228, 255, 256, 262-275, 269-282, 290, 303, 3315, 332, 338, 353, 370, 372, 376, 393, 405, 406, 408, 414, 428, 441, 443, 456, 475, 480, 483, 494, 496, 506, 507, 509-512, 516, 524-526, 528-530, 532, 533, 536, 541, 547, 553 heat island 324, 541, 549, 550, 553 heat unit 151, 152, 239, 352, 353, 408, 531 herbivory 76, 97, 98, 124, 168 high altitude 5, 18, 195-202, 206, 208, 209, 508, 509 high latitude 4, 95, 96, 175-183, 185, 187, 189, 195, 198, 206, 209, 375 history 4, 28, 38, 46, 48, 57, 61, 62, 69, 76, 105, 114, 169, 177, 201, 217, 269, 345, 346, 394, 395, 397, 407, 444, 457, 524, 541 horticulture 39, 505, 506, 509, 513, 519, 520 hydrosphere 385-387, 390-393, 399 ICP Forests 45, 47, 54

Index impacts 23, 40, 47, 60, 93, 94, 96, 97, 106, 109, 111, 114, 116, 132, 142, 145, 148, 151, 165, 183, 185, 217, 220, 225, 226, 228, 230, 245, 249, 255, 266, 301, 306, 314, 319, 322326, 331, 332, 339, 408, 454456, 483, 505, 506, 518, 520 inconsistencies 323 indicator 6, 14-16, 62, 64, 68, 93, 163, 285, 286, 295, 319, 322, 323, 336-339, 395, 407-409, 411-414, 418, 422, 468, 474, 479, 480, 487, 499, 507, 514, 518 indices 23, 60, 63, 132, 187, 266, 286, 325, 331, 334, 336-338, 341, 366-368, 372, 469, 472, 474, 480, 481, 483, 484, 523, 530-532, 535 inflection point 369-373 insect 3, 5, 14, 35, 37, 50, 57, 64, 108, 109, 115, 127, 163, 209, 237-240, 244, 248-251, 285, 321, 331, 387, 394, 405-412, 414, 418, 426, 484, 517 International Phenological Garden 45, 46, 101, 176, 320, 508, 519, 542 Japan 11, 20-24, 45, 94, 200, 429 land cover change 121, 123, 454 landscape 15, 61, 83, 126, 142, 157-160, 169, 170, 203, 242244, 246, 247, 251, 285, 290292, 298, 411, 454, 463, 529 life cycle 5, 93, 96, 106-109, 111, 127, 154, 155, 200, 237-240, 248-250, 255, 331, 405, 411 lilac 58-61, 64, 67, 98, 101, 144, 178, 206, 311, 320, 331, 333, 334, 336-341, 407, 416 linear change 308

Index mammal 5, 37, 69, 128, 331, 437, 440, 442 marine 107, 140, 386, 387, 389, 391-395, 397, 399, 527 measurement 5, 38, 60, 61, 96, 132, 176, 203, 204, 224, 235, 266, 303, 323, 327, 336, 340, 342, 345-347, 349, 351, 356, 357, 359, 360, 372, 392, 397, 422, 468, 469, 487, 490, 492, 494, 498, 500, 501 mechanistic model 219, 220, 224, 227 Mediterranean 5, 75, 80, 84, 87, 139-148, 151, 152, 321, 325, 508, 527, 534 microclimate 142, 167, 168, 177, 197, 200, 203, 320, 359, 360, 408, 511, 516, 549 migration 3, 66, 107, 109, 113, 179, 391, 395, 406, 422-424, 426, 427, 432, 433 model 5, 6, 17-22, 58, 111, 115, 131, 151, 157, 159, 185, 187, 194, 200, 204, 207, 217-221, 223-229, 237-242, 244-251, 294, 297-299, 310, 325, 331338, 345, 351, 353, 354, 356, 357, 359, 360, 370, 371, 373, 375-377, 405, 408, 411, 414, 443, 444, 453, 455-457, 460, 487, 488, 490, 497, 500, 505, 513, 516, 517, 532, 535 monitoring 4, 6, 13, 20, 29, 3537, 39, 40, 45, 47, 48, 52-54, 65, 70, 76, 93, 94, 96, 97, 100, 103, 105, 106, 110, 111, 115, 116, 130, 132, 151, 169, 170, 177, 265, 266, 317, 333, 342, 375, 376, 399, 407, 408, 463, 467, 468, 497, 535, 536 montane 75, 79-81, 84, 85, 87, 195-197, 203 National Weather Service 6, 47, 49, 54, 351, 542

561 nesting 3, 428, 429, 431, 438 network 4, 6, 9-15, 20, 30, 35, 35, 38, 40, 45-54, 57-65, 67, 69-71, 78, 88, 93, 94, 96, 1001-03, 105, 106, 111, 113116, 169, 176-178, 189, 218, 265, 301, 317, 320, 333, 334, 342, 351, 470, 488, 510, 531, 541-543 North America 4, 5, 57-59, 62, 64, 97, 101, 158, 174, 176, 178, 185, 244-246, 248, 249, 320, 322, 327, 332-334, 337, 338, 341, 342, 413, 422, 429, 442, 528 North Atlantic Oscillation 46, 186, 314, 325, 390, 395, 437, 446, 447, 518, 535, 536 observation 3, 4, 6, 11, 12, 14-16, 19-21, 28-30, 32, 34, 37, 38, 45-54, 57-61, 63-69, 77-79, 88, 93-99, 101, 102, 106, 110, 11, 114, 115, 147, 148, 157, 159, 161-164, 169, 176-180, 183, 187, 202, 203, 218, 219, 224, 226, 226, 239, 251, 255-257, 261, 266, 270, 285, 286, 294, 301, 302, 304-309, 311, 313, 317, 319, 322-324, 334, 339, 365, 366, 371, 373, 376, 378, 392, 393, 397, 399, 408, 421, 422, 425, 433, 446, 457, 461, 463, 467-469, 471, 472, 483, 488, 489, 491-493, 495, 501, 505, 508-516, 518, 519, 530, 531, 541-543, 553 organization 4, 29, 30, 35, 37, 39, 50, 51, 102, 110, 112, 114, 116, 271, 505 oscillations 307, 309, 310, 314, 441, 442, 445-447, 457, 536 parturition 438-442, 446, 447 patterns 23, 33, 35, 58, 76, 78, 79, 81-84, 87, 88, 123, 124,

562 126-128, 131, 146, 151, 157, 159, 160, 162-165, 169, 198, 200-203, 206, 242, 250, 253, 259, 262, 263, 286-288, 295, 298, 322, 337-339, 341, 376, 385, 389, 391, 406, 423, 433, 438, 440, 441, 453-456, 459, 460, 463, 469, 471-475, 479, 480, 484, 518 pest management 247, 251, 351, 406, 411, 412, 414 phenological change 45, 78, 106, 123, 124, 127, 130, 132, 152, 210, 289, 319, 320, 326, 333, 492, 494, 496 Phenological Emergence Index 256, 257 phenological interception 182, 304, 305 phenophase 11, 12, 14-16, 18-21, 23, 50, 52-54, 65, 69, 102, 152, 180-182, 187, 189, 198, 217, 218, 256, 257, 261, 266, 285292, 294, 298, 303, 304, 309, 312, 313, 320322, 325, 326, 505, 510, 511, 513-515, 518, 542-546, 548, 550 photosynthesis 5, 146, 228, 299, 355, 372, 456, 461, 467, 469, 470, 475, 477-484, 529 plankton 391, 392, 394, 395, 397 plant 3-6, 11-16, 18-21, 23, 24, 28, 30, 46, 47, 49-52, 57, 58, 60-71, 75-80, 83, 88, 93, 96102, 107, 109, 121-128, 131, 139, 142, 143, 145-147, 151, 152, 160-168, 175-185, 187, 188, 195, 197-204, 206, 208, 209, 217, 219-221, 226, 228, 237, 239, 242, 251, 255, 256, 269-275, 277, 285, 286, 288295, 298, 303, 315, 319-322, 325, 326, 331-333, 335-340, 354, 355, 360, 365-367, 372, 376, 378, 385-387, 392, 393,

Index 399, 405, 407, 408, 410-414, 421, 432, 441-443, 454-457, 467-471, 474, 475, 483, 484, 489, 492, 497, 498, 505-507, 510, 512-514, 516-520, 523525, 528, 530, 531, 535, 536, 541, 542, 546, 553 plant community 77, 122, 142, 143, 146, 198, 242, 286-288, 291-296, 298, 299, 322, 468471, 474, 475, 483, 484 plant growth stages 269, 496 plant-climate interaction 339, 342 Plantwatch 61, 64-71, 102 pome fruit 277-280 prediction 14, 17, 20, 58, 106, 151, 182, 225, 226, 239, 240, 251, 335, 336, 345, 353, 360, 376, 407-409, 414, 467 radiation 16, 131, 140, 196, 244, 254, 303, 314, 341, 349-352, 354-356, 359, 376, 387, 391, 414, 455-458, 460, 461, 471, 487-491, 494, 496, 497, 499, 500, 516, 528 red deer 206, 437, 442-447 regression 17-19, 23, 220, 243247, 294, 304, 321, 323, 334336, 359, 428, 458, 471, 472 remote sensing 5, 6, 14, 20, 60, 93, 113, 123, 124, 129, 131, 132, 285, 365, 366, 369, 373, 375, 376, 378, 460, 461, 463, 498, 501 Rocky Mountains 196, 197, 203, 206, 251, 375, 456 Royal Society 28-30, 40, 61-63, 178 rural 15, 145, 323, 545-550, 553 satellite 45, 46, 79, 93, 123, 177, 179, 187, 299, 311, 323, 326, 327, 331-334, 337, 365-369, 371-376, 455, 457, 463, 488, 491, 492, 498, 501

Index season 14-16, 19, 27, 30, 35, 36, 47, 58, 61, 63, 69-87, 96, 98, 108, 121, 123-126, 129, 130, 159, 160, 162-168, 175, 182, 183, 185-187, 189, 195, 196, 198-204, 208, 220, 228, 229, 237, 249, 265, 285-294, 296299, 303, 304, 306, 307, 311316, 319, 320, 322, 325-327, 335, 338, 340, 357, 368-375, 378, 396, 397, 405, 409, 411, 421, 429, 431, 438, 440-442, 446, 455, 457, 461-463, 472, 494, 496-498, 505-509, 512, 514, 515, 518-520, 530-535, 543, 553 seasonality 3, 14, 20, 27, 83, 86, 89, 124, 126, 151, 160, 166, 237, 240, 249-251, 285, 286, 292, 298, 301, 303, 312, 314, 373-375, 378, 390, 395, 405, 406, 438, 439, 455, 456 sensor 129, 132, 169, 299, 331, 332, 345, 347, 349-351, 356, 359, 365-369, 371, 461, 476, 494 shielding 345, 349, 359 siting effect 354 snowpack 195, 197, 198, 203, 205, 206, 208, 209 soil 14, 16, 47, 51, 54, 95, 122, 129, 141-146, 151, 158, 160, 161, 163, 167, 181, 228, 251, 273, 274, 363, 339, 354-357, 366, 369, 371, 372, 375, 408, 409, 414, 454, 455, 458-460, 470, 471, 487, 494, 496, 497, 499, 516, 517, 529, 530, 533 South America 4, 75-80, 83, 85, 87, 122 spatial interpolation 245-247 species range 228 species-specific response 166, 189, 389, 390 spring 3, 15-17, 19, 23, 36, 46, 58, 60, 62, 64, 65, 67, 69, 71,

563 83, 85, 87, 96, 108, 140, 146, 148, 149, 152, 157, 162, 165169, 173, 179-184, 186, 187, 189, 206, 208, 220, 247, 249, 256-258, 261, 262, 264, 266, 278, 285-293, 297-299, 304, 306, 309, 312-316, 320-326, 331-342, 351, 365, 372, 375, 378, 388, 389, 395-397, 408, 412-414, 421, 422, 425, 427, 430-433, 443, 472-475, 477, 481-483, 492-494, 496, 497, 501, 506, 507, 510, 516, 518520, 525, 528, 530, 535, 543, 544-546, 550, 553 Spring Indices 60, 219, 334, 336338, 341 standard surface 357 statistical model 17, 18, 219, 224, 225 subalpine 196, 198-200, 206 synoptic 220, 332, 334-336, 354 tallgrass prairie 157-163, 165167, 472-474, 480, 488, 492 teaching 19, 20, 251 temperature 3, 5, 15-17, 19, 2224, 27, 46, 51, 67, 85, 93, 9598, 123, 140, 141, 144, 146, 148, 149, 151, 152, 158, 163, 175, 176, 180-182, 185-187, 195-198, 204, 207, 217, 219229, 237-245, 248-251, 254, 259, 260, 262-264, 277, 292294, 296-299, 302, 303, 315, 316, 321-326, 331-333, 335341, 345-360, 365, 371, 373, 375, 376, 385, 386, 388, 389, 391, 393-397, 405-411, 414, 426-428, 430-433, 441, 442, 449, 454, 455, 458, 460, 461, 463, 469-471, 496, 506, 507, 512, 513, 516-520, 528-535, 543, 545, 548-553 treshold 16, 167, 180, 200, 220, 221, 224, 227, 239, 266, 286,

564 294, 305, 307, 315, 326, 332, 334, 335, 353, 356, 369, 371, 373, 389, 392, 406-414, 457, 461, 462, 507, 530-532 Transient Maxima Hypothesis 158 trend 22, 23, 52, 66, 78, 94, 144, 151, 166, 179, 197, 206-208, 288, 295, 304, 305, 307-310, 316, 320-324, 326, 337, 338, 340, 367, 369, 371, 375, 422426, 428-430, 432, 433, 454, 459, 461, 475, 493, 494, 496, 518, 519, 533, 543-547, 549, 550, 552, 553 tropical 4, 75, 77, 79, 80, 83, 8688, 121-124, 126-128, 131, 132, 140, 202, 209, 219, 393 United States 11, 57, 58, 61, 220, 367, 411-414, 455, 456, 459, 536 urban 15, 20, 22, 23, 97, 145, 323, 324, 454, 459, 541, 545, 546, 548, 549 users 106, 109-113, 115, 116, 316, 317, 375, 536 variation 61, 107-109, 125, 127, 131, 147, 149, 151, 152, 158, 161, 162, 164-166, 169, 175, 176, 180, 182, 196, 198, 201, 203, 204, 208, 209, 219, 241, 245, 246, 255, 261, 263, 313, 320, 377, 386, 388, 391, 405, 439-441, 446, 454, 456, 459, 460, 462, 463, 488-494, 496, 497, 501, 534 vegetation 47, 76, 78-81, 83-87, 129, 132, 139, 142-147, 152, 158, 159, 167, 228, 265, 266, 294, 295, 298, 315, 339, 356, 357, 365-368, 370, 372, 375, 414, 453-457, 461-463, 467, 469-471, 474, 475, 478-483, 487, 488, 491, 492, 497, 501,

Index 507, 510, 542, 544, 545, 549, 553 vegetation indices 299, 327, 366368, 371, 372, 378, 456, 494, 498 vegetation type 75, 76, 78, 79, 81, 83, 85, 87, 88, 142, 369, 370, 378, 473, 480, 482, 483 viticulture 524, 527, 529-531, 534 wavelets 453-460, 463 winegrape 5, 145, 523, 529-531, 533 winter 15-17, 23, 36, 83, 87, 89, 140, 144, 147, 148, 152, 179, 183-186, 196, 197, 200, 202204, 206, 207, 228, 237, 242, 246-249, 255, 264, 273, 278, 286, 288-290, 292, 293, 298, 299, 302, 303, 312, 314-316, 325, 327, 338, 375, 396, 397, 406, 426, 430, 432, 439-446, 483, 489, 493-498, 506, 507, 510, 514-519, 525, 527, 529, 531, 532, 543, 545, 553