Capacity Building for the Planning, Assessment, and Systematic

Environmental Science

Karan Deo Singh Capacity Building for the Planning, Assessment, and Systematic Observations of Forests With Special Reference to Tropical Countries Environmental Science and Engineering

Environmental Science

Series Editors

Rod Allan Ulrich Förstner Wim Salomons

For further volumes: http://www.springer.com/series/3234 Karan Deo Singh

Capacity Building for the Planning, Assessment, and Systematic Observations of Forests

With Special Reference to Tropical Countries

123 Karan Deo Singh Academy of Forest and Environmental Sciences New Delhi India

ISSN 1431-6250 ISBN 978-3-642-32291-4 ISBN 978-3-642-32292-1 (eBook) DOI 10.1007/978-3-642-32292-1 Springer Heidelberg New York Dordrecht London

Library of Congress Control Number: 2012944395

Ó Springer-Verlag Berlin Heidelberg 2013 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this publication or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s location, in its current version, and permission for use must always be obtained from Springer. Permissions for use may be obtained through RightsLink at the Copyright Clearance Center. Violations are liable to prosecution under the respective Copyright Law. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. While the advice and information in this book are believed to be true and accurate at the date of publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com) to Indira Participants of the ﬁrst FAO/SIDA Training Course on Forest Inventory held in Sweden, 1974 Foreword

It is a great pleasure to write a few words on the book and on the author, with whom I have been associated for over 50 years. When reading the text now I put some questions to myself. Is the book only a survey into the history of international forestry cooperation that is of relevance only for persons like me, who have been involved in the capacity development since the early 1960s? Does the book contain knowledge that is worth preserving for future generations? My answer is both. My first experience of international forestry cooperation started with the Joint Meeting of the European Forestry Commission of the FAO and the Timber Committee of the UNECE in October 1963, in connection with the presentation of new European Timber Trends and Studies. The primus motor of the discussions in this joint meeting was Jack Westoby of FAO. His name is worth mentioning in this foreword since both the author of this book and many others including myself have been greatly stimulated by him in their professional work within or outside the FAO. I visited the GOI/FAO Pre-Investment Survey of Forest Resources Project for the first time in 1966 as a forest inventory consultant. K. D. Singh had just taken over as the Senior Statistical Officer in charge of the data processing unit after completing doctoral studies in Germany. In those days, as there were very few electronic data processing facilities available in Delhi, we initially undertook data processing in Sweden. Soon, when the IBM 1620 Computer of the Planning Commission became available to the Project, the Indian Team shuttling between Dehradun and Delhi and working late in the night developed the entire EDP system. At least four of the junior forest officers got opportunities for studies in the USA and they became all important actors within international forestry. I visited PIS finally in 1978, when it hosted a Regional Workshop on Forestry and Land- Use planning and soon after it become the Forest Survey of India. I consider PIS a good example of national/international capacity building initiative. Regional Training Courses and Fellowships have been used as other forms for technology transfer. FAO/SIDA sponsored several training courses during 1972–1978 hosted by the Department of Forest Survey of the SLU where I was in charge since 1958. For one of these courses, we made a sabbatical arrangement for

ix x Foreword

K. D. Singh for one year, partly to carry out the preparatory work for a training course and partly for him to undertake independent research on tropical forest assessments. Such twinning arrangements and training courses seem to offer a very useful approach for individual capacity building measures while the South–South Cooperation seems to make a more lasting impact and provide problem-oriented solutions, as demonstrated by the FAO/SIDA CCB Project, during 1995–1998, with FAO Rome as HQ and working with a number of Lead Centers in the different tropical regions. During 1988–1998, I visited the FAO HQ several times for evaluating the donors’ supported components of Global Forest Resources Assessment (FRA) and CCB Project. I observed that most of the FRA staff had been one-time associate professional ofﬁcers, deputed by donors to work and learn in the FAO Field Projects in the developing countries. This past investment in capacity building was now paying back. During my tenure as head of the Department of Forest Survey in Sweden, two of my colleagues Klaus Janz and Reidar Persson also worked as Associate Professional Ofﬁcers in the FAO World Forest Inventory Unit and made important contributions to the Global FRA process later. Capacity development, retaining expertise, and continuity in forest assessments seem as important a need in the countries as at the FAO HQ, which is facing high expectations of the international community for reliable global level data on a continuing basis. The UNCED Agenda highlights the need for capacity building of countries as well as international institutions and donors. Global assessments and country capacity buildings are interrelated challenges for which appropriate strategy and international cooperation has to be evolved. I believe that this book will make an important contribution toward achieving these objectives.

Avesta, Sweden, February 2013 Nils-Erik Nilsson Preface

The purpose of the book is to contribute to establishing/strengthening capacities in the planning, assessment, and systematic observations of forests in the framework of UNCED Agenda 21, Programme Area D of Chap. 11 Combating Deforestation. Following the Rio Conference, a number of countries, donors, and international organizations have implemented capacity building projects with varying degrees of success. A main reason for the varying success rate seems to lie in the design of technical assistance programmes, which have been formulated on the traditional lines mainly to generate forest resources information or transfer of technology; whereas Agenda 21 requires fundamental changes in a country’s institutions and approach to plan and implement the conservation and sustainable development of forests through a process of continuing research and analysis. States, according to the legally non-binding Forestry Principles, have the sovereign and inalienable right to utilize, manage, and develop their forests. Accordingly, they have to take the ﬁrst charge of establishing institutions and strengthening of their capabilities; international and regional cooperation can only build on the national initiatives. With these considerations in mind, the presentation in the book has been divided into two parts. Chapters 1–10 in the ﬁrst part cover topics related to country led initiatives in planning and implementation of forest inventories; and Chaps. 11–20 in the second part describe areas for regional and international cooperation to advance the country capacity building process. This division of contents is expected to delineate areas of national and international action for capacity building; and basic and more advanced areas of forest inventories. The book places emphasis on South–South Cooperation as a means to rapid strengthening of country capacity. The tropical forest formations are homologous across the continents, with remarkable similarities in ecological conditions, structure, and physiognomy of the vegetation, as social and economic conditions. Sharing of knowledge and data among countries of the region has great value for modeling and survey studies to meet the research and development needs of countries. Such a cooperation has proved most valuable for the temperate and boreal zone countries, where UNECE/FAO has been playing an important role

xi xii Preface since 1945 in sharing of knowledge, harmonization of assessment techniques, and consensus building on important forestry issues in the industrialized regions. The international organizations have a catalytic role to play in realizing the UNCED Agenda 21 Objectives. Since its foundation in 1945, FAO has been contributing to development and dissemination of knowledge in the ﬁeld of forest inventories and providing technical assistance to member countries (on request). FAO HQ, unquestionably, is the most important repository of knowledge on tropical forests. This fact coupled with its presence in most tropical countries, lends the Organization a unique advantage to contribute to country capacity building on a continuing basis, using mechanism such as South–South Coopera- tion. This will also signiﬁcantly improve the quality of global forest assessments, which is an important mandate of the Organization. The preamble of Agenda 21 states the role of international cooperation clearly: ‘‘No nation can achieve this on its own. Together we can, in a global partnership for sustainable development’’.

New Delhi, February 2013 Karan Deo Singh Acknowledgments

This book is an outcome of the continuing engagement of the author with the subject for over 50 years starting with his appointment in 1960 as Assistant Sil- viculturist, Uttar Pradesh Forest Department, India, which had traditions for systematic observations of forests using permanent sample plots dating back to 1925. The posting as Senior Statistical Officer in the Pre-investment Survey of Forest Resources, Dehradun, initially a FAO/UNDP/GOI Project and later transformed into Forest Survey of India, gave the opportunity to work on theoretical and practical aspects of forest inventory in different parts of the country with many national and international experts and, in particular, Nils-Erik Nilsson, FAO Data Processing Consultant to the Project, which grew into a long lasting association. Short-term FAO assignments as Data Processing Consultant to other countries helped to widen the knowledge base of the author and also promote South–South Cooperation. The continuing assignment at Rome, in 1979–1998, provided an excellent opportunity to progressively improve FAO Regular Programme activities in forest inventory, technical backstopping of country projects, organize several national international training courses, and undertake the challenge of Global Forest Resources Assessment 1990. The teamwork of the FRA staff, concerned FAO officers Mr. J. P. Lanly and Mr. Klaus Janz, and a network of international and national experts made possible to complete the FRA1990 report on the state of world forests and ongoing changes; and a paper on the state of tropical forests and country capacity in planning and forest assessments presented at UNCED 1992. A concrete outcome of the latter was international support to Programme Area D of Chap. 11 of Agenda 21 and the Interregional Project on Country Capacity Building in Forest Resources Assessment, Planning and Evaluation, 1995–1998, funded by the Government of Sweden. After a research fellowship on biological diversity at the Harvard Center for International Development, USA, it is a great pleasure to work again at home on local forestry and livelihood issues. The book reflects the long innings of work on forest inventory problems of India, other countries of the tropics, and FAO HQ Rome. I take the opportunity to express my sincere thanks to countries, donors, and all colleagues and in particular the staff of the Pre-Investment Survey of Forest Resources, GOI, Dehradun, and

xiii xiv Acknowledgments

Global Forest Resources Assessment, FAO, Rome, for the pleasure of working together to enrich the knowledge base for planning and assessments and country capacity building in the tropical regions for sustainable forest management, a common obligation set forth in Chap. 11 of UNCED Agenda 21. Abbreviations

APM Area Production Model ARD Afforestation, Reforestation, and Deforestation AVHRR Advanced Very High Resolution Radiometer CBD Convention on Biological Diversity CCB Country Capacity Building CGIAR Consultative Group on International Agricultural Research CIFOR The Center for International Forestry Research CTFS Center for Tropical Forest Science CTFT Centre Technique Forestier Tropical DBH Diameter at Breast Height ECE Economic Commission for Europe EFZ Ecoﬂoristic Zone FAO Food and Agriculture Organization of the United Nations FDP Forest Dynamics Plots FINN1DA Finnish International Development Authority FORIS 1990 Forest Resources Information System 1990 FRA 1990 Forest Resources Assessment 1990 Project FSI Forest Survey of India FORIS FAO Forest Resources Information System GIS Geographic Information System GOFC GOLD Global Observations for Forest Cover and Land Dynamics GOI Government of India IIASA The International Institute for Applied Systems Analysis ICIV Institut de la Carte Internationale de la Végétation IPI Indian Photo-Interpretation Intitute IRS Indian Remote Sensing Satellite ITTO International Tropical Timber Organization IUCN International Union for the Conservation of Nature

xv xvi Abbreviations

IUFRO International Union of Forest Research Organizations LANDSAT MSS/TM LANDSAT Satellite Multi-spectral Scanner/Thematic Mapper LCCS /GLCN Land Cover Classiﬁcation System / Global Land Cover Network LIL Low Intensity Logging RIL Reduced Impact Logging NFMA National Forest Monitoring and Assessment NTFP Non-Timber Forest Produce NASA National Aeronautics and Space Administration (United States of America) OECD The Organisation for Economic Co-operation and Development PIS /FR Preinvestment Survey of Forest Resources, India PSP Permanent Sample Plots REDD Reduction of Emission from Deforestation and Forest degradation RPF Relative Production Function RS Remote Sensing SFM Sustainable Forest Management SIDA Swedish International Development Agency TCDC Technical Cooperation among Developing Countries TFAP Tropical Forests Action Programme TOF Trees Outside Forests UNCED United Nations Conference on Environment and Development UNDP United Nations Development Programme UNESCO The United Nations Educational, Scientiﬁc and Cultural Organization UNEP United Nations Environment Programme UNFCCC UN Framework Convention on Climate Change UNFF United Nations Forest Forum UNIDO United Nations Industrial Development Organization USGS EDC US Geological Survey / EROS Data Center WCMC World Conservation Monitoring Centre WFI Word Forest Inventory WWF World Wide Fund for Nature Contents

Part I Building Forest Inventory Institutions

1 The Growing Mandate of Forest Inventories ...... 3 1.1 Emerging Environmental Problems ...... 3 1.2 The Road from Stockholm to Rio...... 4 1.3 Global Forest Resources Assessments During 2000–2010 . . . . 5 1.4 The Existing Capacity in the Tropical Regions...... 6 1.5 The Purpose and Organization of the Book ...... 7 1.5.1 Purpose of the Book ...... 7 1.5.2 Organization of the Book ...... 8 1.5.3 The Information Sources ...... 9 Recommended Further Reading ...... 10

2 Forest Inventory Problem Formulation ...... 13 2.1 Linking Forest Inventory with the Problem ...... 13 2.2 The Changing Demand for Forest Inventory Information. . . . . 14 2.3 Problem-Oriented Classification of Forest Inventories ...... 15 2.4 Identification of Information Needs ...... 16 2.5 Identification and Assessment of Environmental Functions of Forests ...... 18 Recommended Further Reading ...... 20

3 Organizing Existing Information...... 23 3.1 The Role of Existing Information ...... 23 3.2 The Existing Forest Inventories Data and Reports ...... 24 3.3 The Existing Forest Research Data ...... 24 3.4 National/International Libraries and Journals ...... 25 3.5 Forest Dynamics Plots (FDP) ...... 26 3.6 FAO FORIS: An Example of Organizing Country Data . . . . . 27 Recommended Further Reading ...... 29

xvii xviii Contents

4 Technology Transfer and Applications ...... 31 4.1 The Role of Technology in Forest Inventory ...... 31 4.2 A Classification of Emerging Technologies ...... 31 4.3 Strategy for Adopting New Technologies ...... 32 4.3.1 Strengthening Core Competence...... 32 4.4 Special Considerations in Technology Applications ...... 34 4.5 FAO Remote Sensing Surveys of Tropical Forests ...... 35 4.5.1 Background ...... 35 4.5.2 Methodology ...... 36 4.5.3 Main Findings ...... 38 Recommended Further Reading ...... 40

5 Capacity Building in Planning and Forest Assessments ...... 41 5.1 The Problem Formulation ...... 41 5.2 Areas for Capacity Development in Forest Assessments . . . . . 42 5.3 Integration of Planning with Forest Inventory: An Important Issue ...... 43 5.3.1 Long-Term Forestry Planning (Strategic Forestry Planning) ...... 43 5.3.2 Medium and Short-Term Forestry Planning (or Forest Management Planning) ...... 44 5.4 Sustainable Non-Timber Forest Management: An Emerging Area ...... 45 5.5 European Experience with Capacity Development ...... 46 5.6 The Role of International/Regional Cooperation ...... 47 Recommended Further Reading ...... 48

Part II Practice of Forest Inventory

6 Statistical Planning...... 51 6.1 The Purpose of Statistical Planning...... 51 6.2 Role of Forest Statistician ...... 52 6.3 Main Steps in the Sample Survey Design ...... 52 6.4 Some Commonly Used Designs for Forest Assessments . . . . . 54 6.4.1 A Brief Description of Designs ...... 55 6.5 Survey of Trees Outside Forests...... 56 6.5.1 Introduction ...... 56 6.5.2 The Formulation of Survey Objectives ...... 57 6.5.3 Defining Survey Universe ...... 57 6.5.4 Survey Methodology ...... 57 6.5.5 Bangladesh National Inventory of Village Forests . . . 58 Contents xix

6.5.6 Survey of Trees Outside Forests in India...... 60 6.5.7 Distance Method for Study of Discontinuous Vegetation of Andhra Pradesh, India...... 60 6.6 A Forest Inventory Planning Checklist ...... 61 Recommended Further Reading ...... 62

7 Special Studies ...... 63 7.1 The Scope of Special Studies...... 63 7.2 The Planning of Special Studies ...... 63 7.3 Development of Volume Equations...... 64 7.4 Biomass Functions ...... 67 7.5 Non-Wood Forest Products ...... 69 7.5.1 Fruit/Seed/Pulp Yield ...... 70 Recommended Further Reading ...... 72

8 Data Collection...... 73 8.1 Classification of Data Sources ...... 73 8.2 Field Plan and Logistics ...... 74 8.3 Field Manual and Field Forms ...... 74 8.4 Special Studies...... 76 8.5 Check-Crew Work ...... 77 8.6 Computer-Assisted Editing and Data Archival Routines . . . . . 77 Recommended Further Reading ...... 78

9 Data Processing ...... 79 9.1 Roles of Data Processing...... 79 9.2 Data Processing Operations in a Forest Inventory...... 80 9.2.1 Phase I: Manual and Computer-Assisted Editing of Field Forms ...... 80 9.2.2 Phase II: Development of Volume Functions ...... 82 9.2.3 Phase III: Tree Volume Estimation and Plot Level Summaries for Error Calculation ...... 82 9.2.4 Phase IV: Estimation of Means and Standard Errors ...... 83 9.2.5 Phase V: Final Tabulations and Database Storage and Archival Routines...... 84 9.3 Some Strategic Data Processing Questions? ...... 84 9.4 Generalized Versus Tailor-Made EDP Systems ...... 85 9.5 Case Study of FAO Forest Inventory Data Processing System (FIDAPS) ...... 86 9.5.1 Output and Input Specifications ...... 86 9.5.2 PC-FIDAPS Documentation...... 87 9.5.3 Concluding Remarks...... 87 Recommended Further Reading ...... 88 xx Contents

10 The Report Writing ...... 89 10.1 General Comments on Reporting ...... 89 10.2 Forest Inventory Problem Formulation ...... 89 10.3 The Statistical Planning...... 90 10.3.1 The Sampling Design ...... 91 10.3.2 Special Studies...... 93 10.4 Main Findings of the Survey ...... 94 10.4.1 The Land Cover and Forest Changes ...... 94 10.4.2 The Condition of the Forest Floor ...... 95 10.4.3 Trees Outside Forests ...... 95 10.4.4 Comparison with Other Forests in the District ...... 96 10.4.5 Livelihood and Resource-Use Pattern ...... 97 10.4.6 Fuelwood Gathering and Sal Leaf Plucking ...... 98 10.5 Survey Evaluation and Recommendations ...... 100 Recommended Further Reading ...... 101

Part III South–South Cooperation

11 Common Patterns of Spatial Variations in the Tropics ...... 105 11.1 Similarities in Forest Formations Across the Continents . . . . . 105 11.2 Macro-Variation Patterns and Their Significance for Stratification ...... 106 11.3 Meso-Variation Patterns and Their Significance for the Sampling Design ...... 107 11.4 Micro-Variation Patterns and Their Significance for Plot Size and Shape...... 109 11.5 Rain forest Loss and Change ...... 111 Recommended Further Reading ...... 113

12 Remote Sensing Applications in Forest Inventory ...... 115 12.1 On Rapid Developments in Remote Sensing Technology . . . . 115 12.2 Lidar Potentials in Forest Inventory ...... 118 12.3 Applications of Aerial Photographs in Forest Inventory...... 119 12.3.1 Complete Photo Interpretation ...... 120 12.3.2 Point Photo Interpretation ...... 123 12.4 Estimating Cost-Effectiveness of Remote Sensing ...... 126 12.5 A Case Study of FSI State of Forest Report...... 127 12.5.1 Assessment Method ...... 127 12.5.2 Accuracy Assessment ...... 128 Recommended Further Reading ...... 129

13 Growth and Yield Studies...... 131 13.1 Special Growth and Yield Conditions in the Tropics...... 131 13.2 Growth and Yield of Tropical Plantations ...... 132 Contents xxi

13.2.1 Methods of Study ...... 132 13.2.2 The Current State of Knowledge ...... 133 13.3 Growth and Yield Research in the Temperate Zone ...... 134 13.3.1 The Current Status ...... 134 13.3.2 Forest Plantation’s Development Modeling ...... 135 13.4 Growth and Yield of Mixed Tropical Forests ...... 138 13.5 Growth and Yield of Individual Trees...... 138 13.5.1 Methods of Research...... 138 13.5.2 Stump and Stem Analysis ...... 139 13.5.3 Increment Borings ...... 140 13.6 Applications of G&Y Research in Forest Management Planning ...... 142 13.6.1 Hill Dipterocarp Forests of Malaysia...... 143 13.6.2 Mixed Tropical Forests, Indonesia ...... 143 Recommended Further Reading ...... 144

14 Estimating Potential Productivity of Forests ...... 147 14.1 The Need for Potential Productivity Estimation ...... 147 14.2 Description of Climatic Indices ...... 148 14.2.1 Paterson’s Climate–Vegetation–Productivity Index . . . 148 14.2.2 Validation of Paterson Index for India ...... 149 14.2.3 Weck Productivity Index (WPI) ...... 149 14.2.4 Validation of Weck Productivity Indices ...... 150 14.3 Recent Availability of Climatic Data for the Tropics ...... 151 14.3.1 Temperature...... 152 14.3.2 Growing Season ...... 152 14.3.3 Relative Humidity ...... 152 14.3.4 Day Length ...... 153 14.3.5 Precipitation...... 153 14.4 The Areas of Further Research on Forest–Climate Relation . . . 153 14.4.1 Productivity Indices ...... 153 14.4.2 Annual Growth Indices ...... 154 14.4.3 Hardness Index...... 155 14.4.4 Suggestions for a Climatic Index ...... 156 14.5 South–South Cooperation in Climate–Change Research...... 157 Recommended Further Reading ...... 158

15 Land Evaluation Techniques for Forestry Planning ...... 159 15.1 The Purpose of Land Evaluation ...... 159 15.2 Description of Land Evaluation Techniques ...... 159 15.3 Applications of Land Evaluation Techniques in Forestry . . . . . 163 15.4 Land Evaluation for Forestry Planning at the National Level . . . 164 15.5 Land Evaluation for Forestry Planning at the District Level . . . 166 Recommended Further Reading ...... 167 xxii Contents

Part IV International Dimensions of Forest Resources Assessments

16 Identiﬁcation and Evaluation of Environmental Functions of Forests ...... 171 16.1 The Problem Formulation ...... 171 16.2 Components of Cultural and Natural Ecosystems ...... 173 16.3 The Ecosystem Dynamics ...... 175 16.4 The Ecosystem Variables and Change Model...... 176 16.5 Example of a Study Using Ecosystem Approach ...... 178 Recommended Further Reading ...... 180

17 Ecological Zoning and Assessments of Biological Diversity in the Tropics...... 181 17.1 The Need for Ecological Zoning ...... 181 17.2 The Approach for Ecological Zoning ...... 182 17.2.1 The Choice of Parameters ...... 182 17.2.2 The Classification and Mapping of EFZ ...... 182 17.2.3 The Validation Phase ...... 184 17.3 The EFZ Map and the Database ...... 184 17.4 The Tropical Forest Ecosystems Report 1992...... 184 17.5 Biodiversity Loss Associated with Tropical Deforestation . . . . 186 17.5.1 Problem Formulation...... 186 17.5.2 Modeling of Biological Diversity Richness Loss: FRA1990 Approach ...... 187 17.5.3 Species Area Relation by Ecological Zone ...... 188 17.5.4 Rate of Deforestation by Ecological Zone ...... 188 17.5.5 Risk of Species Richness Loss ...... 189 Recommended Further Reading ...... 190

18 Forest Assessments for Climate Change Reporting ...... 191 18.1 Reporting Requirements Under Kyoto Protocol ...... 191 18.2 The Approaches for Reporting Used by Annex I Parties . . . . . 192 18.3 Potential Approaches for Non-Annex I Parties ...... 195 18.3.1 Estimating Forest Area Changes...... 196 18.3.2 Estimation of Growing Stock and Changes ...... 196 18.3.3 Estimation of Above-Ground Biomass and Carbon . . . 197 18.3.4 Estimation of Belowground Carbon, Litter, and Woody Debris ...... 197 18.4 Uncertainties in ARD Reporting...... 199 18.4.1 Uncertainty Arising from Definitions of Forest . . . . . 199 18.4.2 Uncertainties Related to ‘‘Direct Human Induced’’ Activities...... 199 18.5 Validation and Verification of Reports ...... 200 Recommended Further Reading ...... 201 Contents xxiii

19 Global Forest Resources Assessments ...... 203 19.1 The Rising Importance of Global Forests ...... 203 19.2 The Early Assessments: 1948–1968 ...... 203 19.3 Recent Assessments: 1980–2010...... 204 19.3.1 FRA1980...... 204 19.3.2 FRA1990...... 204 19.3.3 FRA2000...... 205 19.3.4 FRA2005 and FRA2010 ...... 206 19.3.5 Some Comments on Global FRA Figures ...... 207 19.3.6 Questionnaire Approach ...... 208 19.3.7 Remote Sensing Survey Method...... 208 19.4 Future Global Forest Assessments ...... 208 19.4.1 Problem Formulation...... 208 19.4.2 The Changing Objectives and Information Needs . . . . 209 19.4.3 Method of Data Collection and Analysis ...... 209 19.4.4 Institutional Strategy ...... 210 Recommended Further Reading ...... 211

20 International Support to Country Capacity Building ...... 213 20.1 Agenda 21 Recommendations ...... 213 20.2 Development and Dissemination of Knowledge for Tropical Forest Assessments ...... 214 20.3 Review of International Assistance to Country Capacity Building...... 215 20.3.1 Technical Assistance During 1980–1994 ...... 215 20.3.2 Technical Assistance During 2000–2010 ...... 216 20.4 Lessons of Interregional Country Capacity Building Project, 1995–1998 ...... 218 20.4.1 National Forest Resources Assessments...... 219 20.4.2 Regional/International Network and Activities ...... 220 20.4.3 The Global Forest Resources Assessments...... 221 20.5 Concluding Observations ...... 221 Recommended Further Reading ...... 222

Index ...... 223 Part I Building Forest Inventory Institutions

Abstract This part presents cornerstones of the country capacity building in forest assessments and planning to meet the growing demand of information about multiple functions of forests at the national and global levels. The measures include: (1) Enhancing ability to understand and analyze the decision problems and formulate them in forest inventory terms; (2) Building capacity to organize and use the existing information for solving decision problems and in planning of new surveys; (3) Acquisition of appropriate technology like GPS, Remote Sensing, and GIS in solving inventory problems; and (4) integrating inventory information with the medium and long-term forestry planning. Chapter 1 The Growing Mandate of Forest Inventories

1.1 Emerging Environmental Problems

This introductory chapter will briefly describe the rapid emergence of a new genera of problems, like climate change, biodiversity loss, land degradation, etc., not much heard of say 40 years back. These problems are not only conceptually complex, but international in scope, calling for consistent information at national, regional, and global levels in the form of a time series. This poses a problem to tropical countries as many of them lack institutional capacity as well as financial resources to collect and provide information. Techniques and technology are advancing fast to ameliorate the situation, but in turn, they also create problems for continuity and compatibility of national/global assessments. The twin issues, viz., strategy for national/global forest assessments and strategy for country capacity development for the purpose, are the main concerns of the book. The Stockholm Conference on Human Environment 1972 sowed the seeds of change by bringing environment issues in the global focus. In the background was the report:‘‘Limits to Growth’’, published a year before by the Club of Rome, which questioned the sustainability of exponentially rising trends of consumption of the planet’s limited resources, arising from the unprecedented population growth, rising per capita income, fuelled by scientific and technological advances (Meadow et al. 1972). Forests and forestry dominated the debate. The Conference recommended countries to: • Strengthen basic and applied research for improved forest planning and management with emphasis on environmental functions of forests; • Modernize forest management concepts by including multiple functions and reflecting the cost and benefits of amenities which forests provide; and • Introduce a minimum of management plans where none currently exist and where governments already committed, should increase their efforts.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 3 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_1, Springer-Verlag Berlin Heidelberg 2013 4 1 The Growing Mandate of Forest Inventories

The Conference recommended the UN Secretary General to take steps to ensure that: • UN Bodies cooperate to meet the needs for new knowledge to incorporate environmental values in the national land use and forest management; and • Support continuing surveillance of the world’s forest cover through establishment (in countries) of appropriate monitoring systems. Though many of the global problems, like climate change and biological diversity, were not explicitly mentioned, the basic direction for the national action and global thinking was well laid at Stockholm. The ensuing account will present the rising importance of environment in the global development debate.

1.2 The Road from Stockholm to Rio

The two important outcomes of the Stockholm Conference were: (i) Establishment of United Nations Environmental Programme (UNEP) in 1972 at Nairobi, Kenya; and (ii) Decision to undertake FAO/UNEP Tropical Forest Resources Assessment 1980 Project with the following objectives: (a) Assess, at regional and global levels, the present state of closed tropical forests and woodlands and the rate and pattern of their depletion and degradation, as a prerequisite for the definition and implementation of the appropriate measures; (b) Determine the methodology and the means needed for the continuous updating of this first assessment. The FAO/UNEP Project made an in-depth assessment and produced the most quoted statistics those days on the annual rate of tropical deforestation, viz., 11.3 million ha. The findings sent an alarm signal worldwide about the state of tropical forests and contributed to two developments : (i) Initiation in 1985 of the Tropical Forestry Action Plan (TFAP) to promote sustainable management of tropical forests; and (ii) Starting again the Global Forest Resources Assessments (FRA), discontinued in 1968, with 1990 as the reference date. FRA1990 was implemented in three phases, viz., the updating of the 1980 assessment for tropical part by FAO Rome and the temperate and boreal parts by UNECE/FAO Geneva; and making of a global synthesis by FAO Rome. FRA1990 for the tropical part, while maintaining continuity of concepts and definitions used in the earlier assessment, made use of two complementary approaches for estimating the rate of tropical deforestation: viz., (i) a model-based assessment using reliable country data; and (ii) a statistical assessment using a stratified random sample of 117 high resolution satellite data spread over the entire tropics for the reference years 1980 and 1990. FAO Rome completed the tropical assessment and released a report on the state of tropical forests coinciding with the United Nations Conference on Environment and Development (UNCED) 1992 at Rio. The annual forest loss during the decade 1.2 The Road from Stockholm to Rio 5

1980–1990 was estimated at 15.3 million ha. Another important ﬁnding was weak country capacity to make reliable assessment of forest resources (FAO 1993). These ﬁndings were very discomforting, but most timely for drafting of the UNCED Agenda 21 recommendations for country capacity building in forest resources assessments.

1.3 Global Forest Resources Assessments During 2000–2010

In the new millennium, FAO has implemented three rounds of global forest assessments with the reference years 2000, 2005, and 2010, making the global assessments ﬁve-yearly as was the practice during 1948–1968. The number of variables for which data has been compiled has grown steeply in successive rounds, from 10 in 2000 to 40 in 2005 to 90 in 2010! Table 1.1 presents thematic elements and main information categories in FRA2010, as an example.

Table 1.1 Parameters of global forest resources assessment 2010 Thematic element FRA 2010 Information Categories 1. Extent of forest resources Area of forest Growing stock of forests Forest carbon stock in living biomass 2. Forest biological diversity Area of primary forest Area of forest designated primarily for conservation of biodiversity Area of forest within protected areas 3. Forest health and vitality Area of forest affected by ﬁre Area of forest affected by insects 4. Productive functions of forest Area of forest designated primarily for production resources Area of planted forest Total wood removals 5. Protective functions of forest Area of forest designated primarily for protection of soil resources and water 6. Socio-economic functions of Area of forest under private ownership forests Value of total wood removals Employment in primary production of goods 7. Legal, policy, and institutional Forest area with management plan framework Human resources in public forest institutions Number of students graduating in forestry 6 1 The Growing Mandate of Forest Inventories

The provision of reliable and timely information on forests and forestry for international purposes is a major challenge to developing countries, not only on account of the complexity of deﬁnitions and associated assessments, but also due to the technical and ﬁnancial means required to plan and implement inventories, analyze data, and report information for international use.

1.4 The Existing Capacity in the Tropical Regions

It may be noted that forest inventory capacity in the industrialized countries is generally well developed, while that in the developing world it is far from adequate. Table 1.2, compiled by GOFC–GOLD in 2009 for reporting to UNFFF, summarized the existing capacities of 99 tropical countries (non-Annex I) to implement an accurate forest area change and carbon stock monitoring. The statistics show that capacity is much less for volume assessment (based on ﬁeld inventories) than for forest area change assessment (based on satellite remote sensing). About 70 % of countries seem to have limitations in their ability to provide a complete and accurate estimation of greenhouse gas (GHG) emissions and forest loss; less than 20 % of the countries had submitted a complete GHG inventory; and only 3 out of the 99 countries had capacities considered to be very good for both forest area change monitoring and volume assessments. The need for country capacity building to improve the reliability of global forests assessments is, therefore, obvious. In addition, cross-country data consistency is also an important factor in determining the quality of global estimates. A main reason for this is the Law of Propagation of Errors: country observations with low reliability make proportionately larger contribution to the error in estimating the global total. To achieve the latter objective, South–South Cooperation could play an important role very much like UNECE/FAO Geneva, which serves as a platform in evolving a common standard for the boreal and temperate zone assessments. FRA1990 noted considerable variation among countries with respect to com- pleteness and quality of forest resources information, with Asia faring better than

Table 1.2 The current country capacities for undertaking reliable forest inventories and forest area change monitoring based on GOFC–GOLD (2009) Forest inventory capacities Forest area change monitoring capacities Total Very low Limited Some Good Very good Very low 19 3 6 8 3 39 Limited 13 3 6 7 2 31 Some 1 0 2 0 0 3 Good 4 7 2 5 3 21 Very good 1 0 0 1 3 5 SUM 38 13 16 21 11 99 1.4 The Existing Capacity in the Tropical Regions 7 tropical America and the latter better than tropical Africa and concluded that ‘‘it is unlikely that the state and change information on forest cover area, biomass and biodiversity could be produced on a statistically reliable basis within the next 10 or 20 years unless a concerted effort is made to enhance the country capacity in forest inventory and monitoring.’’ It may be interesting to note the FRA1948 conclusion that ‘‘lack of reliable forest inventory information’’ was among the ‘‘most important’’ and ‘‘fundamental difﬁculty’’ in implementing global assessment. The FRA2010 states that ‘‘the conclusions regarding the poor information availability in earlier FRA reports are still valid: many developing countries have difﬁculty in reporting because their national monitoring systems are inadequate both for international reporting and domestic needs’’. A well-articulated strategy for the regional as well as country capacity building is obvious.

1.5 The Purpose and Organization of the Book

1.5.1 Purpose of the Book

The book is intended to contribute to UNCED Agenda 21 Objectives of strengthening capacities for the planning, assessment, and systematic observations of forests with special reference to tropical regions. It may be mentioned at the very outset that capacity building is a concept broader than transfer of forest inventory skills; its purpose is to develop competence ‘‘for a better understanding of the role and importance of forests and to realistically plan for their effective conservation, management, regeneration, and sustainable development (The Basis of Action in Agenda 21)’’. The OECD defines ‘‘capacity development as the process by which individuals, groups, institutions and societies increase their abilities to: perform core functions, solve problems, define and achieve objectives, understand and deal with their development needs in a broad system context and in a sustainable manner.’’ Obviously, the solution is country specific. There seems to be no cut and paste model for planning CCB. The problem is known since long, but realization of the need for concerted efforts is very recent, viz., after the UNCED Report on the current country capacity for planning and assessment of forests. Following the Rio Conference, the FAO Forestry Department commissioned an Auto-Review of its activities in the area of forest resources assessment in 1994. The findings of the review led to establishment of an Interregional Project in 1995 at FAO HQ for developing and demonstrating an approach for country capacity building. The project, led by the present author during 1995–1998, made a systematic study of the problem to find solutions. The experience gained has been the motivating factor to undertake the writing of this book and the choice of the topics. 8 1 The Growing Mandate of Forest Inventories

1.5.2 Organization of the Book

The book is divided into four Parts as follows: I Building a Forest Inventory Institution (Chaps. 1–5) II Practice of Forest Inventory (Chaps. 6–10) III South–South Cooperation (Chaps. 11–15) IV International Dimensions of FRA (Chaps. 16–20) The ﬁrst Part is concerned with establishment/strengthening of forest inventory institutions for designing problem-oriented inventories, identifying information needs, technology choices, and taking appropriate action for further capacity building. The second Part presents the practice of forest inventory including statistical methods, special studies, data collection, data processing, and report writing. It is important to visualize these steps as a connected whole in order to achieve the best result (see Fig. 1.1).

Fig. 1.1 Forest inventory as an information process Problem Formulation

Statistical Design

Survey Data Collection Evaluation

Data Processing

Report Writing 1.5 The Purpose and Organization of the Book 9

The third Part emphasizes the role of South–South cooperation to accelerate the pace of country capacity building and presents: common patterns of spatial variation in the rainforests of the world, advancements in remote sensing technology and its applications, assessment of growth and yield, estimation of Potential Productivity of Tropical Forests based on climatic parameters, and Land Evaluation Techniques for forestry planning. Finally, Part IV is devoted to international dimensions of FRA, including Identiﬁcation and evaluation of environmental functions of forests, Ecological zoning of tropical forests and biological diversity assessments, Estimation of forest carbon, FAO Global/Regional Forest Assessments; and, ﬁnally, the need for an International and Donors’ support to country capacity building and promoting South-South Cooperation among countries of the tropical regions.

1.5.3 The Information Sources

The presentation in the book is richly illustrated with data and results from past forest inventories and direct observations of the author, some of which are unpublished. The purpose throughout is to stimulate thinking to find innovative solutions, rather than guide thinking or suggest technology-driven choices. Regional cooperation is emphasized in the fast growing technology field such as digital remote sensing techniques, geographic information systems, communication net-working, etc. The South–South Cooperation could provide access to very useful information and ensure that appropriate technologies are adopted by countries. North–South Cooperation could also facilitate technology transfer and R&D studies on global forestry issues. However, each country needs to develop a long-term strategy for application of fast growing technology and make necessary investments on human resources and institutional development. The main purpose of the book is to contribute to country capacity building in the tropical regions to solve emerging national and global forests and forestry problems with focus on planning and forest assessments. Not to digress from the main goal, the book has purposely avoided going into details of techniques and technology associated with forest inventories, which is literally vast and growing fast. The interested readers may wish to consult standard books on related thematic areas listed in the Reference Section. Elementary Forest Sampling by FRANK FREESE is a very useful handbook, which may be freely downloaded from the site: www.fs.fed.us/fmsc/ftp/measure/cruising/other/docs/AgHbk232.pdf. Keeping this in view no formula for estimating means, variances, or developing regression functions have been provided. Reports on forest inventories conducted in different countries, in the framework of FAO and donor-assisted projects, are a very good source of information. A copy of many of these reports is available in the Country Boxes of the FAO Forest Resources Assessment (FRA) Library. Finally, proceedings of forest inventory meetings and workshops, often available on the Web could serve a most useful 10 1 The Growing Mandate of Forest Inventories purpose for getting information and strengthening of country capacity in forest inventory. FAO and the Swedish International Development Agency (SIDA) had jointly organized a number of training courses in forest inventory with focus on capacity building. The Finnish Development Agency (FINNIDA) has also supported some training courses. The course proceedings are available at FAO HQ and FAO FRA Library, which contain a wealth of theoretical and practical information relevant to forest inventories with special reference to forests and forestry in the tropical regions. Finally, a question can be asked: what does it involve to make a forester a forest inventory expert? Loetsch and Haller (1964), reviewing the growth of forest inventory techniques since the beginning of the century, concluded that ‘‘mastering of the new inventory techniques requires a level of knowledge which is normally not expected of the academically trained forester for the fulfillment of his/her normal professional duties. Consequently and similar to the situation in many other professions, the tendency toward increased specialization is spreading. It is certainly unnecessary for each academically trained forester to possess all the technical knowledge required for the planning and execution of a forest inventory. In so far as tradition and the corresponding instructions prescribe such inventories, it is usually easy for any trained forester to fulfill this task simply by following the prescribed procedures. But a thorough knowledge of the key technical developments is essential if the forester is faced with the task of revising existing instructions or of devising a first set of instructions which will almost invariably be the case in forest inventories in tropical developing countries’’. If the same question were posed in 2012, an obvious reply will be that countries and states need to establish forest inventory institutions and strengthen capability in theory as well as practice of forest inventory in order to meet growing demands for forest resources information at national and global levels. International agencies like FAO need to contribute to the process in an effective manner as there is strong synergy between the national and global assessments, as described in this chapter.

Recommended Further Reading

FAO (1948) World forest inventory. FAO, Rome FAO (1982) Tropical forest resources: FAO forestry paper 30. FAO Rome, Italy FAO (1993) Forest resources assessment 1990: tropical countries, FAO forestry paper 112. FAO Rome, Italy FAO (2001) Global forest resources assessment 2000: main report, forestry paper 140. FAO Rome, Italy FAO (2006) Global forest resources assessment 2005: progress towards sustainable forest management, forestry paper 147. FAO Rome, Italy FAO (2010) Global forest resources assessment 2010, main report, forestry paper 163. FAO Rome, Italy Loetsch F, Haller KE (1964) Forest inventory, vol 1, BLV Verlagsgesellschaft Munchen, Germany Recommended Further Reading 11

UNECE/FAO (2000) Forest resources of Europe, CIS, North America Japan and New Zealand, United Nations, New York

On Web

Stockholm conference on human environment 1972 The Limits to Growth 1972 commissioned by the Club of Rome and authored by Donella H. Meadows, Dennis L. Meadows, Jørgen Randers and William W. Behrens III UNCED Agenda 21 UNEP: Home page Elementary forest sampling by Frank Freese: www.fs.fed.us/fmsc/ftp/measure/cruising/other/ docs/AgHbk232.pdf Chapter 2 Forest Inventory Problem Formulation

2.1 Linking Forest Inventory with the Problem

An Inventory is seldom carried out as a stand-alone activity, it forms almost always a part of a larger decision problem. Developing the ability to understand the decision problems, to assess information needs, and to formulate matching inventory designs are the primary capacity building tasks for any forest inventory institution. A special effort is required on the part of a forest inventory specialist to fully understand the problem, in as much detail as possible, and even anticipate information likely to be asked later, as the decision makers have often a very vague idea of the problem and may not be in a position to state precisely the problem, still less to specify the information required. The users of information are statisticians, economic analysts, and staff associated with policy makers. Very often, they too have hazy ideas, which become clearer as inventory planning proceeds and as more information comes to light. In view of these considerations, the decision process needs to be viewed in a holistic manner including the problem and linking forest inventory with the problem. (refer Fig. 2.1). The inventory professional must try to fully grasp the end result and make a backward analysis of the information needs and then only begin with the inventory planning. The communication gap in the problem deﬁnition phase is likely to affect adversely and even defeat the very purpose for which the inventory is intended. This has been termed as an error of relevance or ‘‘right solution to the wrong problem’’. Janz et al. (2002), in a review of information collected for policy making in the forestry sectors of developing countries (and often also in developed countries), found that data had been gathered not because it was needed but because donors were willing to fund inventories that were vaguely thought to be potentially useful. Information was usually inadequate on topics such as actual removals of wood and other products or the usefulness of the forests especially to the local people. Not enough provision was made for continuous inventories that are needed to measure changes. They suggest, for example, that the resources could be better used on the

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 13 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_2, Springer-Verlag Berlin Heidelberg 2013 14 2 Forest Inventory Problem Formulation

Fig. 2.1 Linking Forest Inventory inventories with the problem

Information

Decision base Monitoring

Decision Problem assessment of changes in land use and forest cover, rather than on elaborate remote sensing or cartographic tasks. A key recommendation by them is establishment of ‘Analysis Units’ to collate and interpret available information, help users to deﬁne their needs, and make the link with information suppliers so that information gathering can be more driven by demand.

2.2 The Changing Demand for Forest Inventory Information

The capacity building in forest inventory is a continuing process. During the last 40 years, there has been major rethinking on the roles of forests and trees. Forestry for food and forestry for people started receiving attention in the 1970s in the context of forestry for local community development (FLCD). Soon after, non- timber forest produce (NTFP) became important in the context of local livelihood. Since UNCED, demand is growing steadily for information about the roles of forests and trees in relation to human environment like: climate change, biological diversity, and desertiﬁcation. Such inventories, informing policy makers and management authorities on the environmental functions of forests, are very few. The environment is unquestionably an important emerging concern of forestry planning. Table 2.1 shows, as an example, changes in forest inventory information needs of India in the course of the last 30 years. The forestry principles (1992) in the Preamble states that: ‘‘The subject of forests is related to the entire range of environmental and development issues and opportunities, including the right to socio-economic development on a sustainable basis’’….. ‘‘Forestry issues and opportunities should be examined in a holistic and balanced manner within the overall context of environment and development, taking into consideration the multiple functions and uses of forests, including traditional uses, and the likely economic and social stress when these uses are 2.2 The Changing Demand for Forest Inventory Information 15

Table 2.1 The changing information needs for forestry planning in India Serial number Forest goods/services Changing demand for forestry information 1980 2010 1. Industrial forestry Very high High 2. Fuel wood Very high Very high 3. Fodder and grazing Very high Very high 4. Non timber forest products Medium Very high 5. Agro-forestry Limited Very high 6. Climate change Almost unknown Very high 7. Watershed services Medium Very high 8. Protected areas Medium High 9. Biological diversity Limited Very high 10. Eco-tourism Limited Medium Source India-IIASA Report (2008) constrained or restricted, as well as the potential for development that sustainable forest management can offer’’. During the last 30 years, the users of forest-related information have grown in a phenomenal manner, in terms of number of users, quality, and variety of information demanded. The methods of assessments have not grown in a corresponding manner. Country capacity has the tendency to stagnate. The conventional inventory approaches cannot be used to collect some of the most needed information. Chapter 17 covers some of the questions related to assessment of biological diversity; and Chap. 18 the issues related to forest assessments in the context of climate change.

2.3 Problem-Oriented Classiﬁcation of Forest Inventories

Nilsson presents a broad classiﬁcation of forest inventory techniques and relates them with the following three distinct planning situations (FAO 1978): (a) Strategic forestry planning; (b) Tactical forestry planning; and (c) Operational forestry planning. National forest inventories, belonging to the strategic group, are most comprehensive in scope and have the aim to meet the information needs of planners who are interested to develop an appropriate strategy for satisfying the present and future forest-related needs of the society. The main survey elements included are the following: • Assessment of existing forest and tree resources; • Estimates of availability of wood or non-wood products from the existing resources; and 16 2 Forest Inventory Problem Formulation

• Study of past trends and future prospects based on assumptions regarding rehabilitation of waste, degraded, or marginal lands or farm forestry and forest management options. A province or regions within a country are usually the lowest unit of reporting in an NFI. The objective generally is to estimate the balance (or imbalance!) between the current or projected forest product needs of the nation compared to what is potentially harvested. This information is of great interest for planning of forest sector development. Pre-investment surveys of forest resources, supported by donors during the 1960s are further examples of strategic forest inventories. The decision problem is defined by the question: ‘‘How much industrial wood, specified by species, dimensions and grades, which can be made available at tentative mill sites, within alternative cost limits per volume unit?’’ As a step towards formulating more detailed objectives, a problem-oriented list of parameters and appropriate classification has to be developed. For example, in constructing a cutting model or a management model, one would need to identify key variables and classify them suitably to enable hierarchical breakdown of the land area and forest area into fairly homogeneous treatment classes. The following groups of variables were included in a pre-investment study in India (Nilsson in FAO 1978): (a) Productivity (site class) (b) Utility (forest type) (c) Maturity (age class of cutting class) (d) Density (volume class) (e) Accessibility (cost class) Management plan inventories are examples included under tactical forestry planning, which aim to provide information needed for estimating and locating the average annual cut during the management plan period in the form of clear cutting or thinning; and restrictions placed on cutting/logging operations. Logging plan inventories are typical examples of the third group of forest inventories aiming to guide: where, when, how much, and in what form logging operations should take place?

2.4 Identiﬁcation of Information Needs

The process of identifying the information associated with different types of forest inventories is illustrated for national and management plan inventories respectively in Tables 2.2 and 2.3 (Husch 1971, FAO 1978, Singh 2000). The National Forest Monitoring and Assessment (NFMA) Programme was initiated by FAO in 2001 with the original aim of supporting countries to enhance forest data collection and analysis, and assisting them in setting up and organizing national forest monitoring and assessment systems. Since 2009, the stated 2.4 Identiﬁcation of Information Needs 17

Table 2.2 National forest inventory and its main data requirements General objectives Specific objectives Information needs 1. Land-use planning and Land characteristics, soil Topography, slope, aspect, soil policy decisions information, economic impact depth, texture and structure, of land use, land-use organic matter, erosion/ planning, keeping in view deposition. Land-use economic, social, and planning. Land-use intensity, environmental impacts erosion hazard 2. Management and Forest area classification from Forest data: forest type, stocking, utilization of existing different planning top height, origin of stand, forest resource for perspectives, growing stock etc., Management data: multiple use broken into major utilization volume by species and classes, diameter class. Year of management classification, logging, Forest condition and accessibility classes, annual regeneration, Growth and potential cut by regions yield data Accessibility data: distance from road, erosion risk 3. Creation of man-made Present consumption and Market demand study, demand forests projection of future projection (rural, urban, requirements industrial) Industrial forestry programmes, Industrial and agro-forestry agro-forestry programmes, plantation, species choice i.e., integrated rural development planning 4. Environmental issues The environmental impact of land Land-use pattern in watersheds, use, watershed condition (soil/ intensity of erosion, water conservation); national environmental impact of land parks (flora, fauna); recreation use. Choice of area/habitat for parks national/recreation parks

Table 2.3 Example of detailed objectives of a management plan inventory and information needs General objectives Detailed objectives Data elements/methods Where Problem-oriented classification of Mapping of stands using aerial stands and compartments from photographs and fieldwork with the point of view of treatment regard to terrain, stand, and tree needs parameters When and how should Treatment priorities within the 10- Accessibility and market studies; cutting/logging year period. Cutting/logging road building plans; fieldwork operations take systems to be applied for integration of data from place? strategic inventories 18 2 Forest Inventory Problem Formulation objective of NFMA is ‘‘to contribute to the sustainable management of forests and trees outside forests by providing national decision makers and stakeholders with the means of acquiring relevant cost-effective information on the state, uses, management of the forest resources and land use changes’’. Management Plan Inventories (also called Working Plan Surveys in many countries) serve as the main instruments of forest planning and control. Some of the countries (India is one example) have a history of forest management dating back to 1856. They are prepared at the forest division level and updated on a 10–15 year cycle following a strict Working Plan Code. Table 2.3 provides an example of problem statement and data categories included in a typical management inventory.

2.5 Identiﬁcation and Assessment of Environmental Functions of Forests

It is obvious from the Forestry Principles that forest assessments need to take into account the social and ecological roles of forests. By limiting the valuation to timber and other material production alone a considerable underestimation of value of forests occurs. The consequences of a wrong land-use choice, especially in environmentally sensitive areas such as the tropic humid belt, arid and semiarid zones, hilly and mountainous regions, are rather evident. Removal of the protective cover usually results in adverse influences on the soil, water, and climatic regimes, resulting in a gradual deterioration of the productive capacity of the land and an eventual formation of barren and denuded sites, wastelands, sand dunes, and deserts. An extended valuation system would also help in deciding alternative and conflicting forest uses, e.g., timber production versus protective and recrea- tional functions. This latter problem is perhaps the one, which at present appears to be most relevant for the European countries, but would soon arise in non-industrialized countries too. For ecosystem services a generalized assessment procedure is required, which should enable identifying, describing and, eventually, quantifying all goods and services under different land uses, including both intended as well as unintended outputs (or disservices). The procedure should satisfy the following requirements (Singh et al. 1974). • It should identify all kinds of services and commodities produced by each option. • It should describe the actual production and if possible estimate the volume, value, or importance of each kind of production. • It should indicate in which way the actual production is influenced by natural production factors or by cultural practices and it should consider the stability of the actual production. • It should indicate potential production alternatives under given management conditions. 2.5 Identification and Assessment of Environmental Functions of Forests 19

Specification of the ‘‘natural’’ as well as ‘‘cultural’’ environment of the production process is essential as they are determining factors of the intended and unintended outputs. On the basis of a fairly loosely defined above criteria, Singh et al. (1974) proposed an ‘‘area production approach’’ as the basis of assessment. Area production was defined as ‘‘a process within an area unit that creates services and commodities, which may be beneficial for man and could be utilized by him’’. Forest production and agriculture production are parts of what could be defined as ‘‘area production’’. Seven production classes (or main outputs), arranged as rows, are given in Table 2.4. The first level distinction is made in three classes: (i) production of environmental services; (ii) production of goods; and (iii) joint (undistinguishable) output including goods and services. The (environmental) services include functions, effects, and influences which may occur (be active) within the area unit itself or outside the area unit. Three classes of environmental services are further distinguished in Column 2, viz., environmental services which occur inside the area unit; those which derive from the area unit but are active outside it; and a third class provisionally termed as ‘‘recipiental services’’. The last class refers to environmental services provided, when the area unit receives, absorbs, or carries influences from outside the area such as: recreation/tourism, absorption of air pollution, water pollution, etc. The production of ‘‘goods’’ is distinguished into three classes, appearing in rows 4–6, and refer to mineral, flora, and fauna commodities. Finally, a seventh product class is intended for joint output (‘‘services and commodities’’), which covers the use of forest land for rural, urban, and industrial sites and infrastructure purposes. The factors influencing the area production are: (i) natural (climate, soil flora, and fauna); (ii) cultural (like harvesting, cultivation, irrigation and fertilization, urbanization and devastation); and (iii) influences from outside the area unit (like policy, laws, or other undertaking

Table 2.4 Identification of primary area production as a part of forestry land-use planning Production Production class Examples of production Category Services 1. Environmental services active Erosion control, soil amelioration, wind break, inside the area unit shelter belts, Conservation of biodiversity 2. Environmental services active Influences on general climate Groundwater, outside the area unit floodwater 3. Recipiental services (carrying Absorption of air and water pollution, Tourism capacity) Goods 4. Mineral commodities Water and other minerals 5. Flora commodities Timber, non-timber 6. Fauna commodities Fish, fowl, and game Goods + 7. Urban/rural services and Infrastructure, industrial use, urban and rural services commodities sites 20 2 Forest Inventory Problem Formulation influencing the environment, etc.). These factors determine both intended as well as unintended outcomes. Table 2.5, based on Singh and Nilsson (2008), provides a list of variables of a comprehensive national forest assessment, including both goods and services. As was mentioned earlier, methodological issues related to inventories of environmental functions of forests are still in an early stage of development. Authors also found some unwillingness to accept environmental services as an ‘‘area production’’, a parallel approach based on ‘‘ecosystem concept’’ was found more acceptable, presented in detail in Chap. 16: Identification and Evaluation of Environmental Functions of Forests.

Table 2.5 Identification of variables for a comprehensive national forest inventory Production class Data to be collected or problem to be analysed Environmental services active Total land area (per civil district or smaller unit) classified by inside the area unit vegetation cover, erosion status, soil depth, and type (such as lateritic soil, sand dunes, ‘waste land’, etc). Location, longitude, latitude, altitude, climate, topography, and soils. Relevance of present land use Implemented and planned projects Areas of specific value from the point of view of landscape, flora, and fauna Environmental services active Wind break, shelter belts, climatic abnormalities and apparent outside the area unit changes of climate, famine years Groundwater level, flood damage Recipiental services (carrying Population pressure, grazing pressure, scenic spots capacity) Mineral commodities Supply of water for households and potential irrigation Oil and mineral resources influencing on the land-use pattern Flora commodities Trends in land use Growing stock, Growth estimates, bamboo, reeds and rattan, NTFP, Accessibility and Market Fauna commodities General estimates, Trends in land use Urban services and commodities General description

Recommended Further Reading

FAO (1974) First FAO/SIDA training course on forest inventory, FAO, Rome, Italy, it includes the article by Nilsson NE with the title: Forest Inventory Problem Formulation FAO (1978) FAO/SIDA/GOI seminar on forest resources appraisal in forestry and land-use planning (asia and the paciﬁc region) held in New Delhi and Dehra Dun, India, 27 November– 15 December 1978 Husch B (1971) Planning a forest inventory. FAO, Rome Janz K, Persson R (2002) How to get to know more about forests? Supply and Use of Information for Forest Policy, CIFOR Occasional Paper, Bogor, Indonesia (in press) Recommended Further Reading 21

Singh KD, Nilsson N-E (1974) On the problem of identiﬁcation and evaluation of environmental functions of forests, FAO/ECE Ad hoc meeting of forest inventory experts, Geneva, Switzerland Singh KD (2000) Guidelines on national inventory of village forests. CIFOR, Bogor Singh KD, Nilsson N-E (2008) Institutionalizing strategic planning, in india, International forestry review, Commonwealth forestry association, UK Chapter 3 Organizing Existing Information

3.1 The Role of Existing Information

The ability to organize and use the existing information in planning and decision making is an important milestone in the capacity development. In the post- UNCED period, the scope of forestry has become very wide and covers ‘‘the entire range of environmental and development issues and opportunities, including the right to socio-economic development on a sustainable basis’’. This implies that forest planning must be carried out in a holistic manner by pooling together relevant data of other sectors such as climate, demography, economics and society. Another interesting development is the view of forests as an ecosystem, which adds time as an important dimension of forest observations. Modern tools like remote sensing and geographic information system (RS and GIS), database management system (DBMS), and modeling are being increasingly used to organize information from the following sources: • The existing forest inventory information and reports; • The existing research studies and scientific publications; and • The information in international libraries and documentation centres. It may be noted that the value of old information on land and forests has become more evident only recently. If field information had been well documented, data might become of great value which might not have been foreseen when the field information was gathered. The inventory specialists (many of them) might not have anticipated the importance of recording site information (such as latitude and longitude) and other influencing parameters (where, when, how!), to which the scientific community is placing a lot of value. Another major bottleneck in the reuse of information is the data secrecy policy of many governments. This may be a historic legacy, which may have to be relaxed to facilitate free flow of basic data and information to encourage research studies on many topics of national and global importance.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 23 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_3, Springer-Verlag Berlin Heidelberg 2013 24 3 Organizing Existing Information

The role of ‘‘analytical units’’ mentioned earlier becomes very important for integrating and organizing the existing data/information in a format, which lends smooth use of the information. The following sections will illustrates the use of various categories of the existing information ending with an important application of the existing country information by FAO to generate global forests assessment 1990 report.

3.2 The Existing Forest Inventories Data and Reports

One may use the existing forest inventory data for different purposes: • For knowing the spatial pattern of variation as guide to size and shape of sampling units and intensity of sampling. A comprehensive review of the spatial pattern of variation in tropical rainforests is presented based on the existing forest inventory reports in Chap. 11, as an example; • Getting auxiliary information for stratification of the survey area; • For making change analysis. This is a very important benefit of historic data, as basis for forest change analysis is being sought increasingly for use in climate change studies; and • As input to model studies on forest production and consumption studies. Many a time, when a fresh inventory is suggested, such as for making a pre- liminary investment decision, it is worthwhile to investigate if the existing information can be updated with some assumptions about changes during the intervening period. This has been done by the author several times with good outcomes. This option is important to consider as waiting for new inventory results implies postponing the decision for at least a few years. Forests change slowly. A comprehensive field trip, combined with a quick look to recent satellite images, may give an idea of land use or forest density changes in the recent past.

3.3 The Existing Forest Research Data

The construction of general volume equations, based on felled tree measurements, is a very expensive and laborious operation. With general disinclination to fell trees for research purposes, the existing general volume equations offer a good possibility of developing local volume equations, if sample tree measurement including diameter and height are made during a forest inventory. The Forest survey of India (FSI 1996) has developed nearly 750 volume equations for 198 species; more than one equation for some widely occurring species. These equations can be screened and the best ﬁtting model selected for transforming the same into local volume equations. 3.3 The Existing Forest Research Data 25

Table 3.1 The model and actual increment and prescription of thinning yield Stratum Name of the Age Standing volume Present increment Prescribed code stratum (years) (m3/ha) (m3/ha) treatment (m3/ha) Yield Actual Yield Actual Thinning Final table ﬁgure table ﬁgure crop I A Badsul 19 90 119 9.3 10.7 7.1 3.6 I B Budhikhamari 19 90 111 9.3 10.3 6.8 3.5 I C Jatipur 19 90 138 9.3 11.5 7.6 3.9 I D Swarupvilla 19 90 65 9.3 7.9 5.2 2.7 II Masinakati 16 80 119 8.7 10.6 7.0 3.6 III Sankhajod 13 68 102 8.5 10.4 6.9 3.5 IV Hatikote 20 94 111 9.3 10.1 6.7 3.4 V Kailashchand 16 80 99 8.7 9.7 6.4 3.3 Source Chaturvedi and Sharma (1980)

Yield tables exist for many timber and non-timber species, both in natural and planted conditions, which may be used in conjunction with forest inventory data to make forecasts of increments with some assumptions. The author investigated the use of the existing Yield Table for Coppice Sal prepared by Howard (1926), updated by Chaturvedi and Sharma (1980), in one of the management plan inventories and found them very useful for assessing the adequacy of the stocking and prescribing treatment to the crop, viz., the thinning intensity and the crop after thinning (see Table 3.1, the last two columns). The existing yield table data, based on systematic observation of forests over a long period of time, have a great value in assessing the adequacy of stocking and estimating climate determined productivity index. Chapter 14 will present methods for estimating potential productivity of a site based on existing climate data and sample plot measurements or existing yield tables.

3.4 National/International Libraries and Journals

Many interesting articles are appearing in national and international journals related to forest inventory and associated technologies, which have great value in suggesting new ideas. The FAO HQ Forestry Library maintains copies of all FAO- assisted forest inventory reports in country boxes, which generally relate to developing countries. Periodically, FAO has commissioned forestry papers in specialized ﬁelds of forest inventory. These are listed in the reference section. Forest inventory documentation also exists at CTFT and OFD Forestry Libraries. The FAO in the past has organized a series of capacity building training courses in various tropical regions and published proceedings in ofﬁcial FAO Languages. The lectures and exercises provide useful reference material. 26 3 Organizing Existing Information

3.5 Forest Dynamics Plots (FDP)

These plots have been laid in different tropical regions using a common approach and objectives, viz., systematic observations of forests coordinated by the Center for Tropical Forest Science (CTFS) of the Smithsonian Tropical Research Institute and maintained by the concerned national research institutions. The data are very valuable for the study of plot size and shape, growth and yield, and many other ecological studies. Plots are re-measured periodically and have large size, viz., 50 ha.

FDP Locations ASIA—Bukit Timah, Singapore—Doi Inthanon National Park, Thailand—Huai Kha Kaeng, Thailand—Kao Chong, Thailand—Lambir, Sarawak—Mudamalai, India—Nanjenshan Nature Reserve, Taiwan—Palanan, Philippines—Pasoh, Malaysia—Sinharaja, Sri Lanka LATIN AMERICA—Barro Colorado Island, Panama—La Planada Nature Reserve, Colombia—Luquillo Experiment Forest, Puerto Rico—Yasuní, National Park, Ecuador AFRICA—Ituri Forest/Okapi Faunal Reserve Democratic Republic of Congo—Korup National Park, Cameroon

The key features of FDP are the following (Singh 2005): • All plants down to 1 cm diameter at breast height are identiﬁed, using the best of taxonomic knowledge; tagged and geo-referenced; and their diameter measured to the nearest mm and re-measured on a periodic basis to permit dynamic modeling. • There is a common standard of measurement followed for all the plots in the network. This provides a unique opportunity for the study of intra- as well as inter-plot studies on spatial variation patterns; • There is a rich scientiﬁc literature of great theoretical interest based on the data of these plots relating to ecology, biogeography, evolution, and spatial pattern of variation in natural tropical forests. The data of the plot located in Mudamalai forests of Tamilnadu, India, were analyzed (Singh et al. 2009) for deciding the size and shape of the sample plot in a survey in Orissa. The species area equation obtained from the dynamic plot data was as follows:

S ¼ 3:0 A0:25 Here S is the total number of species and A is the sample plot area in m2. A graphical presentation of the equation follows in Fig. 3.1. The curve shows that for 1000 m2 sample plot area, the curve starts approaching a saturation value in the 3.5 Forest Dynamics Plots (FDP) 27

Fig. 3.1 Species area curve in dry deciduous forests of Mudumalai, India number of species found in the forest type. The investigations were conducted for plots rectangular in shape with a width of 10 m (5 m on either side of the survey line), in order to capture the maximum number of species optimally.

3.6 FAO FORIS: An Example of Organizing Country Data

The following example illustrates the organization and use of the existing country data for FAO Forest Resources Assessment 1990 (FAO 1993). The foremost challenge to FRA was to develop a statistically sound approach to make the best inferences about forest parameters, viz., 1980 and 1990 estimates of forest area and changes during 1980–1990 for the following three groups of countries: (i) Where two or more date consistent data were available; (ii) When only single date data were available; and (iii) When no data were available (for very few countries only). A database management system was developed with the acronym Forest Resources Information System (FORIS) to store and use existing country data for FAO reporting purposes. The database included three spatial databases covering the entire tropics at a sub-national level including: ecological zone boundaries, forest cover extent, and sub-national boundaries (see Fig. 3.2); and two statistical (attribute) databases: Census data since 1960; and forest area information since 1960. The idea of choosing 1960 at the starting year was to obtain as many inventory data at a province level as available in the countries. Incidentally, demographic data were systematically available for all countries and also found to be correlated with forest loss. Using a subset of sub-national data, which had multi-date forest area information, a model was developed to relate loss of forest as a function of forest area, ecological zone, and demographic changes over the assessment periods. 28 3 Organizing Existing Information

Fig. 3.2 Main geo-referenced data used in assessing the risk of loss of biodiversity in the tropics (originals are at 1:5,000,000 scale) 3.6 FAO FORIS: An Example of Organizing Country Data 29

The model had the form of Chapman-Richard function: dY ¼ b Yb2 b Y dP 1 3

Here b1,b2, and b3 are the model parameters. Y is the percentage of non- forested area in a sub-national unit. It is computed as:

Y ¼100 ðÞTotal geographic area Forest cover area =ðÞTotal geographic area P = ln (1 ? Population density) with population density expressed in persons per km2, In is the natural logarithm, and dY/dP is the derivative of Y with respect to P. The model seems to interpret the human/forest relation as a biological growth process. Very much like many biological growth processes, deforestation was observed to increase relatively slowly at initial stages of population growth, much faster at intermediate stages, and to slow down at final stages. The final computations at the sub-national level included the following steps: (i) Editing of sub-national data; (ii) Adjusting the most recent country data to standard dates, viz., 1980 and 1990 using the model; and (iii) Calculating default values, in case information was lacking for some sub-national units, or in special cases of partial inventories in a province. For preparing global report tables, the ‘‘standard’’ sub-national estimates were aggregated to obtain the ‘‘standard’’ national total, in turn, to form the basis for estimating the ‘‘standard’’global/regional totals. The tropical forest area in 1980 was estimated at 1910 million ha the annual deforestation during 1980–90 at 15.4 million ha. The recent FAO 2000 and FAO 2010 assessments confirm the magnitude of tropical deforestation estimated by FAO 1990 assessment and thereby confirm the soundness of the model approach.

Recommended Further Reading

Chaturvedi AN, Sharma RS (1980) Stand volume and yield tables for Sal (coppice). Indian Forester 106(6):383–396 FAO (1993) Forest resources assessment 1990: tropical countries, FAO forestry paper 112. FAO, Rome FSI (1996) Volume equations for forests of India, Nepal and Bhutan. Forest survey of India, Dehradun Howard SH (1926) Yield table for clear felled Sal (Shorea robusta) coppice, For. Rec. 124, FRI Dehradun Singh KD (2005) Forest biological diversity: assessment and conservation planning. WWF India and ITTO, Yokohama Singh K, Singh D, Sinha B, Sutar P (2009) Community based forest management with micro enterprise for poverty reduction, Bisen Singh Mahendra Pal Singh, Dehradun Chapter 4 Technology Transfer and Applications

4.1 The Role of Technology in Forest Inventory

During the last 30 years, technology has profoundly transformed forest inventory methods. It started with applications of micro-computing technology in the 1970s followed by space technology; and then the information and communication technology. All these technologies have direct bearing on forest inventory design, data collection, data processing, analysis, and reporting method. As presented in Table 1.2, satellite remote sensing is the main source of forest area information in most of tropical countries on account of ease of data gathering and associated cost; whereas regular field inventory is limited to very few (advanced) countries. On the thematic side too, forest inventory organizations are flooded with demands for new types of information relating to problems like: climate change, biological diversity, and land degradation. These issues are not only conceptually complex, but also global in dimension, difficult to comprehend and formulate in practical terms of data collection, analysis, and reporting. On account of the international dimension, harmonization of method and standard terminology is necessary. The capacity building, development of standards, and harmonization of methods and technology transfer need to be coordinated, reasonably robust inventory approaches developed and disseminated to enable countries to fulfill their national obligations.

4.2 A Classiﬁcation of Emerging Technologies

The technology of direct relevance to forest inventory could be grouped as under: (a) The Core Technologies: These include a subset of Computing, GPS, GIS, and Remote Sensing. Technology, considered essential for efﬁcient forest inventory planning and implementation. Within the broad area of remote sensing, the capacity for handling PC-assisted visual analysis is important.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 31 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_4, Springer-Verlag Berlin Heidelberg 2013 32 4 Technology Transfer and Applications

(b) Advance Technologies: Here included are high-end digital remote sensing and GIS technology. They require specialized hardware and software and investment in training. Hardware and software and training requirements change with time and in turn call for new investments in technology to maintain and develop expertise. Only a few developing countries such as Brazil, China, and India might afford and apply them, with a well-articulated strategy for technology application. But, here too, the technology spread is not uniform. A well-coordinated policy is required so that new technology is used at the province and local levels, without having to pay high development costs.

4.3 Strategy for Adopting New Technologies

4.3.1 Strengthening Core Competence

GPS have become indispensable tools of forest inventories, as all forest inventories involve accessing and locating field plots at some stage. There is also a trend towards simpler and cheaper devices and most of countries are making use of them. Remote sensing is currently the main technique for obtaining national and state level forest area information in the majority of tropical countries. It is, therefore, important to develop appropriate techniques and strengthen capacity for interpretation and further use of collected information using GIS. There is also the need for harmonization of classification and image interpretation methods used in countries, as output produced is the main source of information for global forest assessments. The term GIS is currently used in two different senses. In the narrow sense, it refers to computer hardware and software for handling geographic information. In the broader sense, it includes all GIS equipment and operations, including the hardware and software, input data, editing and storing, processing of data, and reporting (See Fig. 4.1). The value of GIS to deal with complex problems like land use and environmental planning, resources management, integrated area development, etc., has been known for a long time. However, recent advances in computer technology such as database management systems (DBMS), computer- assisted drafting and mapping (CAD/CAM) and modeling techniques have made GIS a very powerful and at the same time an affordable tool even at the individual level. Using relatively inexpensive PC, it is now possible to input, integrate, and store information in the form of data as well as maps, analyze and manipulate data and combine these with complex modeling algorithms, and display results in the form of computer graphs, maps, or tables. Together with the word processing capability of computers, the final report, in fact, can directly be produced. Experience suggests that input data for GIS has to be very well prepared if cost and time associated with a GIS study is to be minimized. In fact, to get the best out of a GIS, a total planning 4.3 Strategy for Adopting New Technologies 33

Fig. 4.1 Overview of a GIS system is required, starting with analysis of the problem, output requirements, followed by input design, reliable, and relevant data organization and modeling. In the above context, FAO global land cover information system (GLCN) with land cover classification system (LCCS), as an integral part, seems to be interesting. It includes tools for RS, GIS, and database management and, most important, a land/forest classification system national/regional/global in scope. An evaluation of the GLCN, made by the course participants of the Asia Pacific Region, from 12 countries, after a week training course was as follows (FAO 2006). • GLCN-LCCS could be used as an effective tool for harmonizing and stan- dardizing the classification and mapping of forest cover using remote sensing among Asian countries. • It could form a basis for launching a common programme for national forest cover mapping and continuous monitoring to meet emerging national needs and international obligations. In respect of the core technology related to GPS, GIS, and RS, the country and global strategy should be to maintain and develop overall Core Forest Inventory Expertise in one or more forest inventory institutions of the country. 34 4 Technology Transfer and Applications

4.4 Special Considerations in Technology Applications

Induction of a technology calls for a long-term policy to maintain the base, training of staff, and review of the ongoing methods and practices, briefly mentioned below. Temporal continuity: Often the need for maintaining consistency between procedures and databases produced during different periods is overlooked. This may happen either due to change of method or change of technology. This defeats the very purpose of multi-date mapping. Spatial continuity: A main application of satellite data is to make maps or collect data over a larger area. A change of interpretation/classification between adjoining sheets/reporting areas may cause hanging lines between areas and error in aggregation of data. In particular, for global applications, it is ideal if countries cooperate to adopt a common system of classification and mapping. Thematic integration: If the remote sensing forms a part of the larger forest inventory objectives, it is important to ensure that remote sensing (for area) and field derived data (volume per ha) are compatible and seamlessly integrated. This is important in the context of global studies, e.g., in the framework of climate change involving integration of area and growing stock information. Problem-oriented application: Remote sensing studies need to be carefully designed, in the same way as one makes design of a forest inventory to produce results with minimum cost/error. If, for example, change assessment is the primary goal, an interdependent interpretation of data of two dates would produce consistent and least error change information. Capacity strengthening: A meeting of Non-Annex I Kyoto Protocol Countries organized in 2009 at ICFRE, Dehradun, India, noted that ‘‘finance and strengthening of scientific and technical capabilities of developing countries are the basic requirements for an effective and efficient implementation of a REDD instrument. Capacity building is the most fundamental issue to move towards fulfilling various technical requirements of a future ‘‘REDD’’ (ICFRE 2009). It is recalled that the REDD instrument in the above statement refers to reliable reporting by countries on ‘‘Reduced Emission from Degradation and Deforesta- tion’’, the existing capacity for which is very weak. The delegates noted further that ‘‘while some countries like India, China and Brazil had institutional and infrastructural capacity, others needed more concrete efforts and support to build scientific and technical capacity and expertise for establishment of institutions to meet the requirements of REDD’’ (refer Sect. 1.5). Evidently, there is urgent need for global earth observation system of systems (GEOSS), FAO, ITTO, World Bank, and donor agencies to develop and implement a forest inventory capacity/ institution building initiative for effective participation of countries in the GEOSS Programme in order to realize the global databases and related outcome. Technological advancements contribute to efficiency of field inventories mainly. In no way do they replace them. On the contrary, systematically collected field observations, in particular representative sample plot data, serve invaluable 4.4 Special Considerations in Technology Applications 35 purpose for global change studies and as ground truths for remotely collected observations. They are also essential for providing basic knowledge on growth and yield information for tropical forest species, which generally lack annual rings.

4.5 FAO Remote Sensing Surveys of Tropical Forests

4.5.1 Background

FAO FRA1990 and FRA2000 Surveys of Tropical Forests and FRA2010 Survey of World Forests are examples of a well-planned application of remote sensing to solve issues of major global concern (FAO 1996, 2002, 2010). FRA1990/2000 Sample Survey of Tropics is presented here as it was specifically designed on the pattern of a continuous forest inventory to provide answers to the following questions about ongoing changes in the tropical forests: to find in a reliable manner, if the deforestation is slowing down, has remained unchanged, or is accelerating? This question required estimating change in the rate of change, i.e., the second differential of the forest area information. The second question was: What is happening to the deforested land? This required establishment of a geo-referenced database in the tropics to follow the forest and land-use changes in a location-specific manner. This mission was almost impossible, because no institution in the world had cared to maintain an archive of the first generation satellite (LANDSAT MSSS Chips) images of the tropics. Not only that. The equipments to produce color composited from the MSS chips were not in service any more. With the intervention of the USA Policy Makers, EDC made herculean efforts to dig out historic MSS chips of the tropics, assembled the equipment, and produced images for 1980 matching the sample images of 1990, 119 in total. The classification system and assessment method were specially designed to answer the two almost impossible questions. The specific objectives of the survey were as follows: • Achieve the highest possible level of consistency and precision in the assessment of forest cover state and change at global and regional levels; • Develop and disseminate a simple and robust monitoring technique for producing on a continuing and consistent basis the forest cover state and change estimates at global and regional levels with application also at national level; and • Provide spatial and statistical data for estimating class-to-class changes of land cover and forest cover categories between the two dates of interpretation at the sample locations and for producing change matrices at regional and global levels (see Table 4.1) needed for national and international reporting. The World Reference System 2 (WRS2) of the LANDSAT satellites was used to construct the sampling frame. LANDSAT scenes covering approximately 36 4 Technology Transfer and Applications

Table 4.1 Classiﬁcation used in FAO 1990 and 2000 Remote Sensing Assessments Main classes Additional classes Closed forest (canopy cover [40 %) High density (canopy cover [70 %) Medium density (canopy cover 40–70 %) Open Forest (canopy cover 10–40 %) Long Fallow (forest affected by long fallow shifting cultivation) Fragmented forest (mosaic of forest/non-forest) Dense fragmented (forest fraction 40–70 %) Sparse fragmented (forest fraction 10–40 %) Shrubs Dense shrubs (canopy cover [40 %) Sparse shrubs (canopy cover 10–40 %) Short fallow (agricultural areas with short fallow period) Other land cover Water Plantation (man-made woody vegetation) Forest plantation Agricultural plantation

3.4 million hectares to serve as sampling units. In view of cost-beneﬁt considerations, the sampling units have been selected from all LANDSAT scenes with a minimum land area of 1 million hectares and a forest cover of 10 % or more, estimated on the basis of existing vegetation maps. This has restricted the area surveyed to some 62 % of the total tropical land area but containing some 87 % of the total tropical forest. The main characteristics of the survey were as follows: • It covered the entire range of woody vegetation formations in the tropics, from the rainforests of the wet zone to the shrub and tree savannas of the dry zone. • It provided estimates of mean forest cover area and area change together with an estimate of associated error. • It was cost-effective through use of two-stage stratiﬁcation based on existing FAO global statistical and GIS database.

4.5.2 Methodology

The survey was conducted during 1992–1995 using stratiﬁed randomly chosen 117 sampling units with an intensity of nearly 10 %. The distribution of the selected sampling units by region was 47 in Africa, 30 in Asia, and 40 in Latin America. This sample size was expected to provide estimate of forest cover at global level with a standard error less than ±5 %. The same sites have been revisited again during 2002–2003 to provide change information, as a part of FAO global forest resources assessment 2000. 4.5 FAO Remote Sensing Surveys of Tropical Forests 37

At each sample location, satellite images of the best quality and appropriate season, separated by an approximate 10-year interval, were selected for study. The images close to 1990 in the first round provided assessment of the current state; whereas the area in common between the ‘‘1990’’ and ‘‘1980’’ images provided the assessment to be made of the changes over time. For the 2000 round also a set of two satellite images of 2000 and 1990 were interpreted for the same sites. A distinctive feature of the methodology (viz., interdependent image interpretation) lies in the fact that it provides not only forest cover change data of much higher precision than independent two-date assessments, but also maps and change matrices for each sample location. This enables estimations of class-to-class changes of land cover and forest categories between the two dates of interpretation at sample and its aggregation at the regional and global levels: this is essential information for understanding the complex processes such as deforestation, fragmentation, degradation, and afforestation. The salient features of the survey were the following: • Standard classification of various forest classes (closed, open, with shifting cultivation, fragmented) on a pan-tropical basis; • Interdependent interpretation procedure: this interpretation approach secures the highest level of thematic and spatial consistency between historical and recent image classification. This procedure is the most important element of the methodology since it reduces the error associated with the estimate of changes and makes production of change matrices possible; • Image archive: all images form part of a continuous time series archive. In future these images will be used to estimate the speed of change in conjunction with new data; • Low sophistication: in spite of its sound conceptual basis, the methods and procedures developed for this survey are simple and robust and require very low technological inputs; the methodology has been designed for implementation in average developing country conditions; • Flexibility: although applied here in a global survey context, the monitoring methodology can be applied at national or sub-national level without substantial modifications. The interpretation was implemented at selected regional and national forestry and/or remote sensing institutions, which had a good knowledge of the sample locations and are traditionally involved in forest resources assessment activities. With the two-fold objective of strengthening national capacities for forest monitoring and improving the quality of image interpretation, the Project organized three regional workshops and eight training sessions with national institutions, benefiting 27 countries and 81 participants. The interpretations made by local institutions were centrally reviewed and evaluated. A database was established and analyses carried out as is presented in the following sections. 38 4 Technology Transfer and Applications

Fig. 4.2 The changes during 1980–90 and 1990–2000

4.5.3 Main Findings

The 1990 Assessment Project’s Geographic Information System (GIS), with its multiple layers of geo-referenced information, was an integral element of the survey both at the stage of survey design planning and at the stage of analysis of results. A summary of net changes during the periods 1980–1990 and 1990–2000 by land cover classes at pan-tropical level (million ha) is given in Fig. 4.2. Transition matrices shown below summarize all the changes in land cover classes observed and reported in the interpretation overlays of the sampling unit during the studied periods (1980–90) and (1990–2000) (Table 4.2). The matrices refer only to the common part of all three images of the time series. This restriction of the study area allows comparing matrices from both periods. Similar matrices describing the changes observed in the common area to two consecutive images were also produced. The last row and column sums of the matrices give the area of classes, or states, at the times 1980, 1990, and 2000. The matrices are ideal for tracking changes but not very easy to read. To give a bird’s-eye view of the change information, FRA 1990 developed a pictorial presentation called a ﬂux diagram (see Fig. 4.3), which shows area changes on the x-axis and associated average biomass changes on the y-axis. Thus, the ﬁgure as a whole shows the total biomass change in the form of a rectangle. Such a presentation is easy to read and informative for studies on climate change. 4.5 FAO Remote Sensing Surveys of Tropical Forests 39

Table 4.2 Area transition matrices for the periods 1980–1990 and 1990–2000 at pan- tropical level (million ha)

Notes: The symbol ε indicates values below the displayed decimal point.

Fig. 4.3 Flux diagram showing the path of biomass change 40 4 Technology Transfer and Applications

Figure 4.3 shows clearly the complexity of the dynamics observed. Some remarks on the data gathered are mentioned below: • negative changes are far more prevalent than the positive ones; • almost all types of negative changes are represented; • the two largest area transitions, viz., closed forest ) other land cover and closed forest ) short fallow, also imply the largest amount of biomass loss; • for all other transitions, however, the biomass loss, that corresponds to the area of the rectangle indicating each transition, depends more on the biomass gra- dient than on the area of change; for example, in the third largest area transition (shrubs ) other land cover) the biomass loss has been much less than in other transitions involving less area; and • most of the new plantation areas (both agricultural and forestry) have been created on previous closed forest area, therefore implying, if the average biomass values are accurate, a loss of biomass. As will be presented in Chap. 19, the survey of tropical forests and study of change processes fulﬁlled a major information demand of the international community. Equally important was the fact that it gave a sense of certainty about the information provided by FAO on the tropical forests.

Recommended Further Reading

European Commission (2006) Global land cover validation: Recommendations for evaluation and accuracy assessment of global land cover maps, written by a team of international experts FAO (1996) Forest resources assessment 1990: survey of tropical forest cover and study of change processes, Forestry paper 130, FAO, Rome FAO (2002) Pan-tropical survey of forest cover changes 1980–2000, FAO, Rome FAO (2006) GLCN national workshop for india, FAO, Rome FAO (2010) Remote sensing survey: an outline of objectives, data and approach: forest resources assessment, Working paper 155, FAO Rome

On Web

ICFRE (2009) Summary report of the international workshop on national forest inventory: the experiences of non-annex I countries, Dehradun, India 27–29 April FAO 2006, Workshop on harmonization of forest and land cover classiﬁcation using LCCS for Asia Paciﬁc Region, 4–8 December 2006 Chapter 5 Capacity Building in Planning and Forest Assessments

5.1 The Problem Formulation

UNCED Agenda 21 introduced the basis for action for the Programme area D of Chap. 11 (‘‘planning and forest assessment’’) in the following words: ‘‘Assessment and systematic observations are essential components of long-term planning, for evaluating effects, quantitatively and qualitatively, and for rectifying inadequacies. This mechanism, however, is one of the often neglected aspects of forest resources, management, conservation and development. In many cases, even the basic information related to the area and type of forests, existing potential and volume of harvest is lacking. In many developing countries, there is a lack of structures and mechanisms to carry out these functions. There is an urgent need to rectify this situation for a better understanding of the role and importance of forests and to realistically plan for their effective conservation, management, regeneration, and sustainable development’’. The lack of capacity is as much a problem as the loss of existing capacity. When asked to give his impressions on the progress in country capacity building, one of the colleagues engaged for many years in global assessments and international cooperation, responded as follows: ‘‘Capacity building is a difﬁcult issue’’. ‘‘We’’ have built a lot of capacity but also lost it. Where it remains is in countries more developed. But in ‘‘countries’’ like Liberia, Ivory Coast, CAR, Congo where it once existed it is gone. We still face this problem. Donors will now make inventories for many countries; but don’t take their time to think in terms of continuity and capacity. The Finns are e.g. now making an inventory in Tanzania. In many countries, we would need to pay for a continuous inventory for decades to come’’.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 41 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_5, Springer-Verlag Berlin Heidelberg 2013 42 5 Capacity Building in Planning and Forest Assessments

5.2 Areas for Capacity Development in Forest Assessments

The UNDP in a recent publication defined Capacity as ‘‘the ability of individuals and organisations or organisational units to perform functions effectively, effi- ciently and sustainably’’. This implies that capacity is not a passive state but part of a continuing process and that human resources development is central to capacity building. The overall context (or the environment) within which organizations undertake their functions are also key considerations in capacity development programmes. Capacity is the power of something (a system, an organization, or a person) to perform or to produce effectively the mandated tasks and also look beyond. The following are some priority areas for capacity development: • Problem Formulation: The decision makers and the senior executives of forest inventory institutions work closely on Problem formulation/information need assessment: identify and describe the main developmental and environmental problems and information needed for sound decision making. • Capacity to organize existing information: including establishment of definitions, standards, norms, and methodology (refer recommendations of the UNEP/ FAO meetings at Kotka, Finland) including parameters such as biomass and biodiversity. • Human resources development planning: training/specialization in various aspects of FRA (forest inventory, GIS, remote sensing, application of information in planning) • Technology applications: identify, evaluate, and apply new technology appropriate to the problem and institutions. • Financial resources procurement: secure necessary financial resources for implementing country projects and programmes. Maintenance of human resource capabilities and its further development are key areas for capacity building. It is important that appropriate rules and regulations are enacted to contribute towards CCB and also prevent the loss of existing capacity. Some reasons for capacity stagnation/loss are the following: • High staff turnover. • Lack of training opportunities. • Lack of adequate data processing, remote sensing/GIS facilities. • Funding limitations for inventory updating. • Lack of opportunity to consult with international institutions in inventory design and analysis. • No practical applications of results or reports produced, as they are not problem oriented. • Lack of in-depth analysis of inventory data. • Static or out-of-date institutional objectives. Forest assessment techniques and technology are undergoing rapid changes. Challenges are also growing. It is important that sustained support is provided in 5.2 Areas for Capacity Development in Forest Assessments 43 the country to remain up-to-date and well informed. With the rising importance of global problems like climate change and biological diversity, S–S and N–S Cooperation/twinning may prove to be cost-effective mechanisms to enhance country capacity.

5.3 Integration of Planning with Forest Inventory: An Important Issue

Integration of forest inventory with planning and decision making is an important capacity building task in developing as well as developed countries. In this interface a topic of particular interest is analysis of consequences of alternative forest management programmes, including monitoring of effects. A brief discussion of this important topic follows (refer FAO 1994).

5.3.1 Long-Term Forestry Planning (Strategic Forestry Planning)

Here the concern is with issues of general relevance on the national or province level. Most issues emanate from the need for a consciously planned sustainable management of the natural resources and deal with identiﬁed problems relating to such management. Perhaps the most demanding information need in this context can be summarized in the following questions: what is the level of present and potential production (in a wide sense, including both goods and services) in forests and related lands, compared to the needs; how is it expected to change under alternative forest management programmes? A corresponding monitoring question is: what is the effect of implemented policies on the forest, and in particular its productive capacity? As strategic forestry planning has to deal not only with intra- but also inter- sectoral issues such as land use, energy, employment, social welfare, and environment, the scope of needed information is wide and includes variables linked with both production and consumption (need) of goods and services. This includes: land cover and land use; forest resources information such as volume, increment, and drain; areas suitable for afforestation; endangered species; ecological values; traditional/indigenous land-use values; biomass and productivity; demographic and socio-economic data affecting forest resources; need of land for agriculture. Assessments included under the generic name of ‘‘national forest inventories’’ are the most appropriate ones. They are also called ‘‘strategic forest inventories’’, as they help in formulating alternative programmes and projects for the development of the forestry sector. They cover the entire geographic area under 44 5 Capacity Building in Planning and Forest Assessments consideration which could be a region, province, or the whole country. The inventories are ideally conducted on a continuing basis to provide information on forest resources and land-use changes, growth, drain, and mortality.

5.3.2 Medium and Short-Term Forestry Planning (or Forest Management Planning)

This type of planning is oriented towards location-specific activities. The typical issue is choice of place and time for interventions. These interventions are, in the best case, part of a strategy developed for a larger area unit. The assessments or inventories for these purposes are designed to collect data for implementing projects in the framework of a long-term programme in a specified area (forested or not), e.g., a civil district or forest division. They provide specific prescriptions for treatments of forest stands and sites such as: choice of logging and regeneration techniques, thinning and tending of regenerated areas, wildlife and watershed conservation, etc. These are also called inventories for tactical planning. The survey intensity is rather high and mapping is an essential need. Information needed for forest management is specific to a forest and its present condition; and to an enterprise and its production goals. Included is information like: forest type, stand density, growing stock, and increment by main forest types, species or species groups and site quality, economic accessibility, wood quality of growing stock in relation to commercial acceptability, forest health and forest damage (by area, growing stock, species or species group, age or diameter class, extent and type of damage), stand and site treatment, afforestation, data on environmental, and other non-wood aspects of the forest under management. Besides data collected from forest inventories, long-term experimental observations on response of forests to various treatments are equally important. These serve as important inputs (viz. as a knowledge base) to model studies for simulating development of forest resources under alternative forest management options. Changes in information needs, in the developed and the developing countries alike, relate to environmental functions of forests in general and biomass and biodiversity, in particular. The environmental information has also global implications as will be described in Chapters 16–20. Considerable research on inventory techniques is needed before assessments can be carried out on an operational basis on these subjects. Information related to planning and implementation of national forestry action programmes (NFAP), are among those most needed currently in the developing countries. For this purpose national forest inventories could serve as a useful tool and contribute to preparation of an effective strategy for the forestry sector. The term ‘‘effective’’ is intended to convey the idea that a sound database is an essential requirement to relate investment with return, production with consumption, conservation with development, etc (Table 5.1). 5.3 Integration of Planning with Forest Inventory: An Important Issue 45

Table 5.1 Forest planning, information needs, and technology choices Planning tasks Information needs Technology and tools (i) Strategic planning (district Time series of production GIS and MIS, Policy and state level) and consumption data, Information, and statistical policy data data on other Sectors, Modeling (ii) Forest management Forest inventory and maps, Forest inventory; remote planning market and price data, sensing; GIS and MIS; funds available Area Production Model (iii) National Forest Forest resources statistics and Forest inventory, remote Inventory, Analysis maps related to multiple sensing, GIS, MIS, GPS, and reporting of uses of forests and modeling statistics

Establishment of fairly independent analytical units is very useful that can maintain and develop needed information systems and undertake necessary analysis. (Nilsson 1998).

5.4 Sustainable Non-Timber Forest Management: An Emerging Area

Forest Management in the tropics is traditionally designed for timber production mainly. Non-Timber Forest Produce (NTFP) has been treated so far mainly as a minor produce and not given due attention in the traditional forest research, planning, and control system, even though it meets major social needs. Moreover, NTFP utilization is mostly local and annual in nature. Therefore, annual record keeping in all the compartments is essential, compared to timber species, which mostly have a long rotation of almost 80–100 years and involve harvest operations at the end of rotation period and silvicultural operations take place say once in 10 years. The scientiﬁc basis of timber management in some tropical countries is well established thanks to research plots and experimental studies spread over more than 100 years. For NTFP, a de novo system of planning and control system has to be designed keeping in view their speciﬁc nature; control has to be in place to ensure that removals are within the carrying capacity of the area and are followed by regeneration (Fig. 5.1). In the NTFP context, new knowledge has to be developed to determine sustainable level of production, viz., the annual yield in terms of quantity and species that can be taken out in an area so as not to cause irreversible changes in the regeneration process. Need is also for an accounting system to keep record of what has been taken out, a plan to guide as to what can be taken out and interventions to ensure that balance between removals and new growth is maintained? Finally, forest protection needs to be legally and physically enforced in order that unplanned exploitation does not take place. In the ancient times, the moral sanctions served to self-regulate the process. In our times, we have to investigate: 46 5 Capacity Building in Planning and Forest Assessments

Fig. 5.1 Key issues in sustainable management of NTFP

what guidelines, what governance system, and what laws need to be there in place in order to achieve the same goals, viz., sustainable NTFP management?

5.5 European Experience with Capacity Development

According to Kuusela (1988), ‘‘the history of European forests, their development and utilization by eco-sociologic regions, illustrates the basic principles, methods and prerequisites to keep up, and to re-create if deteriorated, the delicate balance between forest, harvest of timber and other commodities and more intensive land- uses than forestry, in conditions of growing population and changing production techniques and value attitudes against forest. At the very first, there shall be a thorough knowledge gained by research, experience and reliable inventories of forest ecosystems, silviculture and land management. To build-up a strong and tenable silvicultural heritage takes 50–100 years. Whatever is the silvicultural knowledge on academic and professional levels, there 5.5 European Experience with Capacity Development 47 are almost un-surmountable obstacles in agrarian poor societies to reach conditions needed for forest management on the basis of sustained and progressive yield. National self-disciplines based on sufficient knowledge and material means is a prerequisite to create and keep up the balance between forest land and other more intensive land-uses as well as between valuable growing stock in full density and needs to harvest and consume timber. This includes also an administrative issue, e.g. the autonomy of provinces and federal states in economic matters may be an obstacle for a central government to formulate and execute efficient forest policies’’. The message is that a persistent effort is required to build, maintain, and further develop country capacity.

5.6 The Role of International/Regional Cooperation

Countries have sovereign rights over their forests, but also have obligations to provide information on their forests on a periodic basis as signatory to international Conventions and Protocols. This opens possibilities for seeking international assistance for capacity building, which may help in securing essential technical competence, not possible locally. FAO Global Forest Assessment makes use of statistical information provided by countries; and undertakes remote sensing survey on a sampling basis, which also requires countries support. Participation of countries in the global assessment

Geomatics Center of the Andhra Pradesh Forest Department, Hyderabad, India

GIS was totally unknown in the AP Forest Department (APFD), when the FAO/ World Bank sponsored Project started in 1994 as a part of the FAO Country Capacity Building Programme. In less than 4 years, the center became a prime user of GIS in the country and acquired sophisticated hardware and software and the skills to handle most of the spatial and non-spatial data. The center has presently all the necessary software for image processing, GIS and DBMS, etc.; and most important, it has seven well-trained staff. When the center started, it faced a lot of difficulties in database creation as well as processing. The difficulties faced were overcome by perseverance and hard work by the concerned staff. After overcoming the initial difficulties, it has grown into a national center of excellence and is disseminating RS, GIS, and GPS techniques and their applications in the state through three Regional GIS Centers. 48 5 Capacity Building in Planning and Forest Assessments

Geomatics Center of the Andhra Pradesh Forest Department, Hyderabad, India (contd.)

An Analysis of Causes for the Success: The subject is complex to analyze, but some of the contributory factors are brieﬂy touched upon. Initially, FAO provided a vision for the GIS development, application and its institutional growth on a continuing basis including establishing a main center at Dullapalli, close to Hyderabad, and the vision of the three regional sub- centers by ecological zone. All these are reality today. What the GIS Unit did not achieve: GIS was supposed to support a mod- ernized forestry planning system in the State. This goal has not been so far achieved. Reasons are many including the fact that a change in planning system requires approval of the Central Government. GIS and computer technology have not brought any major change in the working or planning environment of APFD, which continue to be done as in the past. No doubt, technologically, a visit to APFD Geomatics Centre at Hyderabad (http://apfdgis.com) is very refreshing. It is the testimony to the success of an FAO/World bank initiative in India. process provides an opportunity to know about technical and technological developments in other countries. South–South Cooperation, also called Technical Cooperation among developing Countries (TCDC), is an accepted method of capacity building. Countries located in analogous ecological regions tend to have very comparable forest formation. Geographic contiguity also makes the travel and meeting cost much less as compared to an interregional event. It has been observed in many countries that capacity built through short-term international or bilateral cooperation has been soon lost, when the project ended. Therefore, the international cooperation on a long-term basis could be the most effective to enhance the country capacity, as will be presented in Chap. 20.

Recommended Further Reading

FAO (1994) A review of FAO’s achievements in forest resources assessment and a strategy for future development, an auto-review by Singh KD, Janz K. FAO, Rome FAO (1998) Strengthening country capacity for planning, assessment and periodic evaluations of forests: support to programme development and coordination unit: project ﬁndings and recommendations. FAO, Rome ITTO (1988) Natural forest management for sustainable timber production, pre-project report [1] Kuusela K (1988) Long term development of lessons from the european forest resources by regions (unpublished report). FAO, Rome SIDA (1998) Development of national forest policies and strategies, introduction and comments by Nils-Erik. Nilsson, International Training Course, In Stockholm, Garpenberg, Skinnsk- atteberg and Vårdnäs, Sweden Part II Practice of Forest Inventory

Abstract This part describes the implementation phases of a forest inventory including: statistical planning, preparation of ﬁeld manuals, training of the crews, and collection of data including check-crew work, data processing, and report writing. A separate chapter is written on each phase, highlighting the need to see various phases as inter-connected and holistic. Finally, the need for leaving an organized database for future use is emphasized. Chapter 6 Statistical Planning

6.1 The Purpose of Statistical Planning

The main purpose of statistical planning is to ensure that survey estimates are reliable (i.e. they fall within the stated confidence interval); and obtained at the lowest cost for a given precision. In addition, some other benefits are the following: • Less Time: because only a fraction of the parent population is measured, the time involved in providing results is much less compared to a Census; • More Adaptability: Survey design and measurement procedures can be adapted to accessibility conditions and the available manpower; • Wider Scope: Innovative studies may be carried out by including additional measurements over only a subset of the sampled population. Fisher (1950) summarized the advantages of sampling in the following words: ‘‘I have made four claims for the sampling procedure. About the first three, adaptability, speed and economy, I need say nothing further. Too many examples are already available to show how much the method has to give in these ways. But, why do I say that it is more scientific than the only procedure with which it may sometimes be in competition, complete enumeration? The answer, in my view, lies in the primary process of designing and planning an enquiry by sampling. Rooted as it is in the mathematical theory of the errors of random sampling, the idea of precision is from the first in the forefront. The director of the survey plans from the first for a predetermined and known level of precision; it is a consideration of which he never loses sight, and precision actually attained, subject to well- understood precautions, is manifest from the results of the enquiry.’’

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 51 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_6, Springer-Verlag Berlin Heidelberg 2013 52 6 Statistical Planning

6.2 Role of the Forest Statistician

As the complexity of the survey increases, consultations with a qualified statistician will be found more and more useful: not only in the design of survey and analysis of data, but also consultations/discussions in various phases of the inventory. A quote from Cox (1951) is appropriate: ‘‘the statistician who expects that his contribution to the planning will involve statistical theory finds repeatedly that he makes his most valuable contribution simply by getting the investigator to explain why he is doing the experiment, to justify the experimental treatments and to defend his claim that the experiment, when completed will enable its objectives to be realized.’’ Besides specialized knowledge of statistics/sampling, a good knowledge of spatial variation patterns in the survey area and costs associated with alternative survey designs is required for a good planning of forest inventory. As an example, Chap. 11 presents spatial patterns of variation for tropical rainforests based on data collected from past inventories. The needed budget, manpower and available time for fieldwork are other important information for planning. An additional issue in the mixed tropical forests is the problem of identifying rare species, which invariably requires use of the local knowledge and consultation with a taxonomist. A forest survey institution with a qualified statistician and trained crews is a lasting asset, which confers many advantages for a country by providing: • Continuing information support to the decision makers based on available data and its analysis; • Using the existing data for the statistical planning of new inventories; and • Undertaking investigations/studies in the emerging areas of global change assessment and reporting.

6.3 Main Steps in the Sample Survey Design

There are many standard reference books on this topic. Therefore, the subject is presented here briefly. For further elaboration, standard textbooks mentioned in the reference may be consulted. Chapter 10 provides a detailed case study on the sample survey procedure implemented by the author related to assessment of non- timber forest produce in the state of Odisha, India. Statement of Objectives: A precise statement of objectives is the starting point in any survey. This has to be done in as precise (quantitative!) a manner as possible in order to guide the choice of the sampling design. Definition of Target Population: The definition of the Survey Population (also called Universe) is an important step, because once the survey area has been agreed upon, further investigations will be limited to this area only. For example, if the objective is to survey only forest areas in a district or a region, then trees outside forests will be excluded from the sample survey even though they may be known to making a major contribution to wood- production of a significant 6.3 Main Steps in the Sample Survey Design 53 amount. To carry out an up-to-date survey, it would be necessary to have an updated map, based on recent satellite data of the survey area. The geographical, administrative, and other specifications of the limits of the survey area have to be defined beyond any ambiguity. Sampling Frame: For the purpose of selecting a sample, it is necessary that the population is divided into a finite numbers of distinct and identifiable units called sampling units. The aggregation of all sampling units in a survey area is defined as the sampling frame. The definition of sampling units needs to be precise as they have to be later physically located and measured. Likewise, the sampling frame should cover the entire survey area without any gap or overlapping of the sampling units. Closely associated to decision on the sampling unit, is the sample plot configuration for carrying out measurements in the field. This topic is illustrated in Chap. 10 on Report Writing and further elaborated in Chap. 11 in relation to micro-pattern of spatial variation in forests. Sampling units constitute the building blocks of all statistical error calculations and the precision targets are set for main variables only like number of stems and volume per ha by species, also called stand and stock tables respectively. Sub- sampling is usually done for variables like height of trees, regeneration, crown size, etc., which are time-consuming and expensive to measure, without setting a precision goal. For the latter purposes, the sampling units are divided into recording units having only a small fraction of area of the sampling unit, say 1/4 or 1/10. Mea- surement of time-consuming parameters is limited to selected recording units only. The Sampling Design: The decision regarding the number of sampling units and their distribution in space requires considerable statistical skill and knowledge, as decision involves many considerations including: (i) the pattern of spatial variation in the survey area; (ii) the accessibility cost associated with reaching a sample plot; and (iii) availability of prior information on the sampling units such as existing map, satellite data, spatial variation pattern in the area, and sample plot accessibility. All these information help to formulate design alternatives along with indication of associated precision and cost, staff and training requirement and other field logistics. The ingenuity of the statistician lies in recommending an optimal number and distribution of sampling units to be measured in the survey space. If an appropriate sampling plan has been made, the final estimate will be not only reliable but also of least cost to achieve the precision goals. In the final analysis, it is not the sophistication of the design, but the ease of implementation as reflected in minimum bias and measurement errors that are the main deciding factors. A less sophisticated but robust design may deliver most of the gain that is likely to be achieved from an optimal design. The Sampling Intensity: The sample size is a function of the sampling design and inherent variability of the population to be surveyed. The sampling units may be distributed over the survey area using any of the procedures described earlier, depending on the availability of auxiliary information and the chosen survey schemes. For each design, there is a matching formula for estimating the mean and 54 6 Statistical Planning standard error of the survey. Inversely, the required number of sampling units for a given level of the standard error can be estimated. In most situations the past experience could be used as a guide in more complex situations. Alternatively, pilot studies may be undertaken, using a few sampling units to provide an idea of the variation pattern in the study area. In case the population is subdivided into homogeneous units, as in stratified sampling, allocation of sampling units to strata requires further consideration. One of two standard approaches is used: (i) sampling units are distributed to strata proportional to their area; or (ii) sampling units are allocated proportional to variance of strata. The latter is also known as optimal or Neyman allocation. The proportional allocation is useful, when the stratum means vary significantly from each other; whereas the optimal allocation is useful, when means and variances both vary significantly among strata.

6.4 Some Commonly Used Designs for Forest Assessments

For more information including formula for estimating the sampling error standard textbooks like Cochran (1963), FAO (1981a), Freese (1962), Loetsch and Haller (1964), Matern (1984), and Schreuder et al. (1995) are referred to. We start with a checklist of designs: 1.1 One-date assessment 1.1.1 No auxiliary variable used in inventory design Simple random sampling Systematic sampling Cluster sampling Block surveys (by topographic or geographic units) Distance and line intercept methods

1.1.2 An auxiliary variable available The term auxiliary information refers to data collected from sources other than the field inventory proper, e.g., remote sensing, past inventory data, socio-economic, or ecological surveys. The remote sensing information as auxiliary variable is the most common case. Again, auxiliary data may be available at the time of survey planning or data may become available only later when the field inventory has been completed. (i) The information available before start of survey Stratified sampling Multiphase sampling with Ratio estimates Regression estimates Probability proportional to size (PPS) 6.4 Some Commonly Used Designs for Forest Assessments 55

(ii) The information available after ﬁeld survey Post-stratiﬁcation of sampling units 1.2 Multi-date assessments (also called sampling on successive occasions) The choice of multi-date over single date survey is guided by the interest in assessment of changes. The best way to estimate changes in land use is remote sensing and that of growing stock is use of re-measured sample plots. Modern technology, such as geo-positioning system (GPS), assists in accurate location of sample plots at the time of next visit.

6.4.1 A Brief Description of Designs

Simple random Sampling: This design is very simple to implement. What is required is a sampling frame, where the sampling units have been serially num- bered 1 to N; and a Table of Random Numbers, to select n sampling units randomly out of N. The procedure does not make use of any existing information about the population being surveyed, which is a major limitation. From a conceptual point of view, the design is very interesting as it provides a basis for judging the ingenuity of the statistician to minimize the sampling error, using more sophisticated designs. Systematic Sampling: As the name implies, sampling units are distributed following a predetermined systematic pattern in the survey area, unlike simple random sampling, where there is no subjective element involved in the choice of location. The units may be linearly aligned, as in a systematic strip or line plot survey, or distributed in the form of a systematic grid form over the survey area. This design is most commonly used in the forest survey practice, in spite of a serious limitation, viz., lack of a statistically valid procedure to calculate the sampling error. Often, the simple random sampling formulae are used to obtain an indicative figure for the error. Stratified Sampling: The role of aerial photographs and satellite remote sensing for stratification is rather well established. Their main contributions are: stratification into forest and non-forest, thereby, elimination of field visit to certain sites; and reduction of variance by dividing the survey area into homogeneous sub- populations than otherwise would have been possible. Reduction of variance by stratification depends on the number of strata and correlation between the photo and ground variables, as given below (Cochran 1963):

VSTðÞ¼VSRðÞfR2=L2 þð1 R2Þg V (ST) and V (SR) are variances of stratified and simple random sampling, respectively, for the same number of sampling units; R is the correlation coefficient between the field and photo-variables; and L the number of strata. With the value of R = 0.8 and L = 4, the variance of stratified sampling is 40 % of the random sampling. 56 6 Statistical Planning

Double sampling or Two Phase Sampling: In this approach, the survey is conducted in two phases. In the less expensive phase one makes measurements of selected attributes over a large number of photo-plots (or the first phase units) and then in the second phase, a subset of these photo-plots are chosen for field measurements. The common sampling units, in the two phases, serve as the ‘‘bridge’’ to estimate (or predict) the values of all first phase sampling units. The critical factor is the cost ratio (Cg/Cp) between the second and the first phases; and the value of the correlation coefficient (R) between the variables observed in the two phases. Singh (2000) lists the limiting values as follows:

R2 ¼ 0:20:30:40:50:60:70:8 Cg=Cp [ 18 11 8 6:54:53:52:5 Two-Stage sampling (also called Cluster Sampling): In inaccessible forest areas of the tropics, reaching a sampling unit is often difficult and expensive, while measurement of more than one sampling unit may be feasible, once the survey crew reaches the spot. By clustering the survey plots, one is likely to cover less of variation, but gain in cost per unit. To realize such a scheme, the survey area is first divided into larger units, called primary sampling unit or the first stage sampling units; and dividing the first stage units into secondary (or second stage) sampling units. A subset of the primary units are selected first randomly, followed by the selection of a subset of secondary units for the survey proper. Primaries may vary in size and the sampled area in different primaries may also vary. Such modifications make the analysis more complex, but still results are valid statistically. For example, the Forest survey of India makes use of two-stage selection process in the layout of sample plots for the fieldwork, viz., choosing 60 districts randomly out of 600 districts of the country every 2 years; carrying out sample plot measurements in the chosen districts during a 2-year period; and returning to the same plot after 20 years! The advantage of two-stage sampling is that it may yield estimates of a given precision at a cost lower than that of a completely random sampling.

6.5 Survey of Trees Outside Forests

6.5.1 Introduction

In the wake of tropical deforestation, Trees Outside Forests (TOF) are serving to meet the local and even national needs increasingly. In 2010, TOF made nearly 82 % of the forest sector contribution to the national GDP in India. The wood produced in the farm was also costing much less to industries than that bought from forests: one metric cube of wood at mill site was priced around US$ 30 compared to US$ 60 per cubic meter from the national forests. 6.5 Survey of Trees Outside Forests 57

As the pattern of spatial distribution of TOF and their development are very different from forests, the inventory approaches also differ. Guidelines on National Inventory of Village Forests (Singh 2000) contain a systematic account of procedures for the survey of trees outside forests. Here problem formulation and procedures applied are presented brieﬂy.

6.5.2 The Formulation of Survey Objectives

The Bangladesh village forests inventory (BVFI) was a pioneer effort to undertake a nationwide assessment of trees outside forests. The main objective was ‘‘to undertake a national inventory of both the major and minor forest produce resources of the Village (or homestead) areas.’’ The sampling was planned to provide an estimate of the mean with sampling error about l0 % at a 95 % con- ﬁdence at a national level. Additional ancillary studies were required on village raw material supplies to selected industries (FAO 1981b). The Forest Survey of India (2011) is conducting surveys ‘‘to assess the availability of forest resources for the production of timber, fuel-wood and raw material for paper pulp, packing cases, essential oils, matchwood etc. in and outside reserve forests areas.’’ Speciﬁcations of the forest inventory by Forest Faculty, Bogor (1986), had the objective ‘‘to collect data on the biomass and fuelwood potential and fuelwood supply in the province of West Java, focusing particularly on fuelwood produced outside forest land.’’ The assessment was needed to identify areas of fuelwood surplus and shortage and for the preparation of biomass production options. This study included observations on degraded lands and inves- tigating the prospect for increasing fuelwood production.

6.5.3 Deﬁning Survey Universe

It is important to ensure that the survey universe of inventories (inside and outside forests) seamlessly merge to provide a holistic picture of the resource situation in the country. This may not be an easy task as national land-use mapping and national forest cover mapping are seldom done in a coordinated manner using a common legend. BVFI made use of Census Village Register to deﬁne the universe. FSI made use offresh mapping using high resolution satellite data to deﬁne the survey universe.

6.5.4 Survey Methodology

A comprehensive inventory design, treating TOF and forest areas as a continuum is the ideal solution. But, this may not be always possible. In case national inventories inside and outside forests are conducted independently, it is desirable that the 58 6 Statistical Planning concerned statisticians use a consistent deﬁnition, approach, and a common land-use classiﬁcation to obtain a national total in terms of area, growing stock, increment, and drain. Three approaches used for the TOF survey will be presented as examples.

6.5.5 Bangladesh National Inventory of Village Forests

The FAO/BGD Project (BGD/78/020) was designed in 1981 to undertake national inventory of village trees in Bangladesh. Conventional methods based on area could have been used but it would have been time-consuming and drawn out with difficult photo interpretation and mapping requirements at a single tree level. New techniques had to be developed to suit the inventory situation within certain cost and time constraints placed on the project FAO (1981b). The thrust of the Project was on inventory of both, viz., the major as well as minor forest produce in the Village Forests (homestead) of Bangladesh. Villages were chosen as the ultimate sampling unit, with the idea to make a production and consumption balance and to integrate inventory with socio-economic data for forestry development planning. The Bangladesh National Inventory of Village Forests used a two-stage sampling design. For the primary sampling, 10 % of the total of 440 Thana units (Sub- division of a district) were selected. Within each of these chosen Thana units, village units having a population between 700 and 1,000 (1974 census) were selected. A total of 267 village units, to serve as secondary sampling units, were randomly chosen giving equal weight on a population basis to all six strata. The overall sampling fraction was 0.0033 % which was constant for all six strata. A sub-sample of 27 village units, called pilot plots, was selected with the criteria that no more than one such plot be accepted from any individual Thana unit. The two-stage sampling design after stratification was found satisfactory. However, the first-stage sampling of 10 % (44 primary sampling units of a total of some 440) was too low and could have led to ‘‘bunching.’’ A minimum of 20 % would be more adequate to achieve a better primary sampling unit spread. A smaller size of primary sampling unit (and larger number) would have given a lower standard error when using ratio estimate, particularly at stratum level. In three of the six strata, the number of sample measured was rather small. A major assumption was that TOF are correlated with the population density. Detailed population data, at village level, was available. Therefore, population rather than area was considered as the basis of sampling design and implementation. The assumption made was that there is a relationship between population and trees. The outcome of the inventory was total volume and stand tables as is usual; but a special feature of this inventory was the use of volume per capita concept rather than volume per unit area values. In this study questionnaire data was of limited reliability and was avoided. This is particularly so with estimates of past harvest where the villager suspects he may be penalized (additional taxes?) if he gives a correct (and high) answer. 6.5 Survey of Trees Outside Forests 59

Also, reliability of questionnaire data very largely depends on the attitude and aptitude of the recorder. Rate of harvest and increment figures were required. It was ensured that a minimum number of permanent inventory plots are established and periodic measurements organized. Harvest figures can be obtained over short terms of 1–2 years, while increment data will of course take a little longer to be meaningful. In such a wide ranging inventory survey, additional data on consumption should be collected if not by the same field party then by another team in the same sampling unit (i.e. village) on a compatible basis. This will reduce time and make such data on consumption/resource directly comparable. Measurement definition and units should of course be the same for both sets of survey figures collected. In Bangladesh both consumption and inventory surveys were not coordinated at the planning stage with the result that two independent projects (under separate ministries) were formulated. Results from both were compared but it made correlation study more difficult and in addition, increased the cost involved in organizing two separate surveys. Results of the project constitute the basic information on this important resource so long lacking in Bangladesh. However, they will certainly not in themselves solve or alleviate the problems associated with the supply (and more importantly the anticipated required increased supply) contribution from the village areas. This should be done with the help of follow-up studies. The present project gives only the basic information on which development studies could be built upon. Some aspects that could receive more consideration in the future include: precise estimation of increment and drain. Some rudimentary figures on drain were collected by BGD/78/020 via questionnaire on the harvest over a 12-month period. Additional remeasurement on a sub-sample of plots will give a much more accurate estimation of increment and drain but only for the specific 12-month period. Therefore, permanent inventory plots need to be established to obtain this as a long-term requirement. As an interim measure increment borings may provide some a useful indication. Presupposing that the present available area is all but fully utilized, increased production is then only possible by introducing genetic improvements to existing species for either or preferably both fruit and wood production. In addition, new exotic species should be gradually introduced (first on a trial basis) to improve production. The homestead forest itself is probably near capacity with regard to forest cover. However, there may be additional areas not presently utilized that could be planted with species meeting specific requirements. Examples of these could be rice field boundaries, irrigation tank embankments and grazing areas. Depending on the above findings such strategies in relation to village forest development could be reviewed. 60 6 Statistical Planning

6.5.6 Survey of Trees Outside Forests in India

The design was developed at the beginning of 2000 and is still ongoing. It covers 60 new districts every 2 years. Forest covered areas are first masked out in the topographic sheets to show only the TOF areas. For survey purposes, a ‘‘district’’ constitutes the first-stage sampling units. A district is further distinguished in two strata, viz., (i) Rural; and (ii) Urban (FSI 2011). The rural stratum is mapped using very high resolution satellite images into three strata: Block plantations, trees in line, and scattered trees. The size of the plot and number of samples for each has been determined by FSI by conducting a pilot study. The plot sizes for Block and Linear strata are 0.1 ha and 10 9 125 m respectively. In case of scattered-trees stratum, the optimum size of sample plot has been fixed as 3.0 ha for non-hilly districts and 0.5 ha for hilly districts. The sample sizes for block, linear, and scattered domain are 35, 50, and 50, respectively, for non-hilly districts and 35, 50, and 95, respectively, for hilly districts. The sample points are randomly generated and the latitude and longitude of the plot in each stratum of district are prepared and provided to the concerned field parties. The data on dbh, crown diameter, species name, and category of plantation, etc., are collected in designed formats. The design of survey in urban areas is different. First, towns in the district are stratified into five population classes: [100,000; 50,000–100,000; 20,000–50,000; 10,000–20,000; and \10,000. The National Sample Survey Organization Urban Frame Survey (UFS) available for every town of the country containing, about 150–160 households, constitutes the sampling unit for urban tree survey. Based on the number of UFS, towns are classified into two classes: (i) having less than 500 UFS; and (ii) having 500 or larger UFS. In the first case, 10 % of blocks are selected for sampling with a minimum of 20. In the second case, 5 % of the blocks are selected with a minimum of 50, but not more than 60. After ascertaining the total number of sampling units to be measured, their distribution is done in each stratum proportional to the area using random numbers. A list of randomly selected blocks in each district is prepared and provided to the concerned field parties for carrying out complete enumeration of all the trees of 10 cm and above dbh in the prescribed formats.

6.5.7 Distance Method for Study of Discontinuous Vegetation of Andhra Pradesh, India

Stephen Smith (1992) used ‘‘line intersect sampling’’ to assess biomass by block (a sub-district unit) in Andhra Pradesh State of India characterized by discontinuous nature of vegetation. In each administrative block six transects of size 100 m by 200 m were measured. The observed large range of variation suggests that stratiﬁcation of blocks or use of probability proportional to estimated size could reduce the sampling error. 6.5 Survey of Trees Outside Forests 61

Thompson (1992) describes procedures, interesting for agroforest surveys, under the title ‘‘Detectability methods for elusive populations.’’ Matern (1984) also presents such methods. Trees outside forests can be or are often elusive. The procedure includes use of ‘‘distance methods,’’ wherein the nearest tree or trees from a randomly located point are located and distance measured. Line intercept sampling is also used, wherein distance of objects (here trees) from a randomly laid out line is recorded. From the measured distances, a distribution pattern of trees is generated. The ‘‘four-sides inventory’’ in China uses a strip of width 10 m and length 1,000 m to record trees and forests in villages, around houses, and along rivers and roads.

6.6 A Forest Inventory Planning Checklist

The purpose of a checklist is to ensure that all essential inventory steps (problems!) have been taken into account. Nothing important has been forgotten. In the words of Prof. Nilsson, the Forest Inventory checklist is the foundation on which the edifice of the Forest Inventory is built. The main items recommended by him are the following: 1. Identify the problem which inventory is expected to solve. 2. Review possible approaches to solve the problem. 3. Clarify what information is required. 4. Define the area to be inventoried. 5. Collect and systematize existing information and carefully assess its usefulness to solve the problem under consideration. 6. Decide whether or not a new inventory is necessary. Consider costs in view of existing information, financing, recruitment, institutional, and administrative difficulties and time. 7. List data to be collected and assessments to be made. 8. Is mapping required? 9. Define required accuracy for main parameters. 10. Clarify the need for special studies: volume functions, growth, and yield studies. 11. (Provisional) outline of design and estimation models. 12. Outline of measurement techniques. 13. Think about recruitment and training of staff, written field instruction. 14. Foresee administrative problems in organization of field and office routines. 15. Ensure control of fieldwork. 16. Expected end use of results in drafting final report and establishing a data bank. 17. Prepare an outline of results to be produced in order to solve the planning problem. 18. Data processing including recording, checking, compilation, and data storage. 62 6 Statistical Planning

19. Undertake necessary preparatory investigation. 20. Undertake pilot study. Control of the inventory plan by going systematically through the checklist would enable taking of corrective steps in time. Steps 1–3 relate to the inventory problem formulation. Steps 4 and 6 relate to the study of existing data and making decision whether or not a new inventory is required. The limitations of existing data have to be weighed against the associated new inventory, its cost, ﬁnances, recruitment of staff, time involved, institutional, and administrative problems, etc. Steps 7–17 relate to the proposed inventory and its elaboration. Once the elements of the new inventory are more or less known, a pilot study, as stated in Steps 18 and 19, could be conducted. Thus, the various steps mentioned above would contribute to wise planning and implementation of the new forest inventory.

Recommended Further Reading

Bogor Forestry Faculty (1986) Report of the study on fuelwood potential and supply in thewest province of Java, Bogor, Indonesia Cochran WG (1963) Sampling techniques. Wiley, New York Cox GM (1951) The value and usefulness of statistics in research, USDA Committee on experimental design, USA FAO (1981a) Manual of forest inventory with special reference to tropics, Forestry Paper 27. FAO, Rome FAO (1981b) Village forest inventory of Bangladesh, Inventory results by Hammermaster ET, BGD/78/020. FAO, Rome Fisher RA (1950) The Sub-Commission on Statistical Sampling of the United Nations, Bull. Int. Statist. Inst. 207–209 Loetsch F, Haller KE (1964) Forest inventory, vol 1. BLV Verlagsgesellschaft Munchen, Germany Matern B (1984) Four lectures on forest biometry. Swedish University of Agricultural Sciences, Umea Schreuder HT, Gregoire TG, Wood GB (1995) Sampling methods for multiresource forest inventory. Wiley, New York Singh KD (2000) Guidelines on National Inventory of Village Forests. CIFOR, Bogor, Indonesia Smith S, Farlane Mc, Phillips J (1992) Line transect sampling in discontinuous tropical vegetation report for CIDA Spurr SH (l952) Forest Inventory. Ronald Press Company, New York Thompson SK (1992) Sampling, John Wiley and Sons, New York

On Web

Freese Frank (1962) Elementary forest sampling. USDA Forest Service, Washington Chapter 7 Special Studies

7.1 The Scope of Special Studies

Special studies form an important component of forest inventories and they may involve signiﬁcant cost and time. They can be carried out separate from the sample survey proper or integrated with it depending on considerations of cost-effectiveness or the need for a specialized team. Examples are: accessibility and logging cost studies, market surveys, etc., in pre-investment surveys and development of volume equations in the forest inventories. In the context of climate change studies, biomass ﬁgures are also in great demand. Studies may be required for many other purposes such as estimation of leaf, fodder, fruit, seed, and bark yield, commonly referred to as non-timber-forest-produce (NTFP), also called Non- Wood- Forest-Products (NWFP). Information on bamboo may be needed in terms of weight in addition to number of culms. Demand for estimating yield in relation to NTFP species is growing, as the following utilization survey in rural India shows (Table 7.1). The wood products, as the above survey shows, constitute the most important use, but other produce have value too, for which suitable guidelines for estimation do not exist. Further research is needed as they are of major importance for the local dwellers.

7.2 The Planning of Special Studies

The planning of special studies involves similar steps as described earlier for the main sample survey, viz., statistical design, data collection, and data processing. In this section, the following topics will be covered:

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 63 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_7, Springer-Verlag Berlin Heidelberg 2013 64 7 Special Studies

Table 7.1 Utilization pattern of NTFP in rural India Utilization pattern Types of species use (%) Wood 100 Fruit 30 Bark 20 Gum 20 Leaf 12 Source: Compilation by author

Volume functions (general and local) Forest biomass functions NTFP yield estimation FAO has brought out several forestry papers on the subject with special reference to tropical regions. Equally valuable are reports from the FAO field projects carried out in the three tropical regions. But, access to them is difficult, as they are filed in the FAO Forestry Department Library or stored in Country Boxes, maintained by the Forest Inventory Unit at FAO Rome.

7.3 Development of Volume Equations

FAO Forestry Paper 22, Volume 1 presents detailed guidelines on definition and measurement of various volume categories, methods for tree and stand developments, and finally development of volume functions. The paper presents and illustrates various techniques in a simple form with illustrations (FAO 1980). As the fieldwork may involve destructive sampling (tree-felling, cross-cutting, and volume measurements), a careful planning of design becomes important. The number of trees to be felled may be determined from earlier experience or through a statistical procedure keeping view the cost of collecting volume data and the variance of volume for a given species, height, and diameter class. In many cases, it may not be possible to fell trees as statistical considerations may demand. A compromise may be made to take volume measurements on trees, which are being felled by the local communities or contractors, taking due care that trees selected for felling are well distributed and cover all ecological zones where the species occurs. But, sometimes this may not cover big size trees or leave small trees without proper representation. In such a case, some tree may be measured standing. Equations involving species, diameter, height, and volume are developed based on felled tree data, known as general (or regional) volume equations. They hold for the entire range of a species distribution; whereas equations involving species, volume, and diameter are called local volume equations and hold for one site class or locality only. The following sequence of steps is involved in volume equation preparation: 7.3 Development of Volume Equations 65

(i) Trees are felled and sectioned and the volume of each log (or section) is estimated using standard formulae (e.g. Newton, Huber, and Smalian); and a database on each tree is developed including standing tree measurements and felled tree measurements; (ii) A function is developed using tree volume (as dependent variable); and tree height, diameter, and species as independent variables. A common function used for a single species or group of species is: V ? a ? bD2.H. The equation can be more complex. The development of equations requires skill and experience. An uneven distribution of trees over the diameter and height classes makes the equation tilt: to predict too low volume (even negative volume) and/or a very high volume. A scatter diagram between dependent and independent variables would show the distribution of values over the variable range and outliers. After equations have been run, it is a good practice to compare the actual and predicted volume by diameter classes to make sure that variance over the diameter classes are uniform. The next task is to develop local volume equation, for estimating volume based on diameter measurements. For this purpose, data are collected on height, diameter and species on a sub-sample basis (also called sample tree data) in the main survey. Using the tree-height measured on sample plots, a local volume function by species or groups of species is developed. A common form of equation is: V = a ? bD? cD2. Again, the function can be more complex. In the above equation, V is the volume and D; a, b, and c are the regression coefﬁcients for a given species or groups of species. (iii) Local volume functions by species or groups of species are then used to estimate the volume of all trees enumerated on the sample plot. The following are some common problems in the volume equation development: • Large data scatter by species and size classes; • Few observations available per species in certain diameter classes; • Breakdown of assumptions for linear regression analysis regarding homogeneity of variance. The Forestry and Forest Industries Development Project Malaysia, (FAO 1974) during 1970–1975, carried out comprehensive studies on tree-form, industrial stem size distribution, and log quality by diameter classes. Data were collected on nine subplots of 503 sampling Units containing 51,724 trees belonging to 606 species above 20.3 cm (8 inches) diameter. Of these, 5,010 trees were felled and measured for volume and decay. The special study accounted for about 40 % of the total ﬁeldwork costs. Basic studies on development of volume functions and assessment of decay in form of special studies were carried out, which laid the foundation for a very reliable volume and quality assessment for the current and future projects. Figure 7.1 illustrates the great variability of tree height in the forest, between species and within the species for the same diameter. 66 7 Special Studies

Fig. 7.1 Variability of Industrial stem length for the same tree diameter

The data of 881 felled trees will be used (see Fig.7.2) to illustrate the variation in form factor (f) as estimated by the equation: v = f.d2.h, where: v: is the volume from stump height, i.e., 61 cm (or top of buttress) to the minimum top diameter of 30.5 cm or Crown Point whichever occurs earlier. d: is diameter of the tree measured at a height of 4.3 feet above ground on the uphill side of a tree or 2.3 feet above the top of buttress. h: is length of main stem measured between stump height or top of buttress to a minimum top diameter or Crown Point, whichever occurs earlier (Fig. 7.2). The above overview shows that the total stem length is the main determinant of the form.

All trees First level factor Second level factors Total trees

Total length 178 Total length 2 (0.59) 2-3 (0.56) 3 (0.54) 282

(0.53) Top diameter Total length 3-6 (0.52) 421 4-7 (0.50) 2 (0.49)

Fig. 7.2 Results of segmentation programme for determining the form of trees (Form factors are given in brackets) 7.4 Biomass Functions 67

7.4 Biomass Functions

Webster Dictionary defines ‘‘biomass’’ as ‘‘the total amount of living matter in the form of one or more kinds of organisms present in a particular habitat usually expressed as weight of organisms per unit area of habitat or as volume or weight of organisms per unit volume of habitat’’. This implies that both vegetation and animals are included in the broader biomass meaning. Usually, also dead tissues of living organisms such as dead branches, heartwood etc., are included, though the exact scope of this term may vary from one study to another. From a forest inventory perspective, the term biomass is defined as the total amount of aboveground living organic matter in trees expressed as oven-dry tons per unit area (FAO 1997). Hartig, the well-known German forester, as early as in 1888, stressed the importance of tree measurements expressed as dry weights. He wrote: ‘knowledge of the amount of dry matter contained in a given volume is of great scientific and practical importance, because dry weights are correlated with calorific values and with a wide range of important technical properties; and because from a knowledge of these values it is possible to obtain a picture of the natural laws that govern the production of organic matter’ (quoted in Parde 1980). In the context of climate change studies, forest assessments in terms of biomass have become a standard practice. The expressions such as ‘‘forest biomass’’, ‘‘wood biomass’’, or ‘‘tree biomass’’ are often used. We first provide a general definition for the term biomass and then a more specific usage related to a single tree or parts thereof (FAO 1986). Forest Biomass: The biomass of trees, shrubs, and lesser vegetation in a forest ecosystem including the belowground parts of the organisms. Woody Biomass: The biomass of the woody vegetation such as trees and shrubs including stump and roots. Tree Biomass: The biomass of vegetation classified as trees including foliage, stump, and roots. The true meaning of this depends on the definition of the tree. FAO (1997) defines ‘‘tree’’ as a woody plant having a main stem, which under normal conditions of growth, reaches a height of 5 m. Trees may have a single main stem in case of conifers (Excurrent tree form) or several branches (Decurrent tree form) in the case of broadleaved species (see Fig. 7.3). For conversion between volume, green weight, and dry weight the following formulae were used: Wood density = dry weight\green volume Moisture content (dry basis) = (Green weight—dry weight)/dry weight Moisture content (green basis) = (Green weight—dry weight)/green weigh Generally speaking, there are two approaches for estimating the biomass of forests: (i) use the existing inventory results available in the form of volume over bark (VOB per ha) and convert them into biomass (t/ha) terms using a conversion factor; or (ii) start with source inventory data or processed stand tables (stems/ha by diameter classes) and apply a biomass function (see FAO 1997). It must be realized that the available inventory data may not include trees below a certain 68 7 Special Studies

Fig. 7.3 The tree and its components

diameter class and non-woody biomass. To use either of these methods, the inventory must include all tree species. Brown and Lugo (1984) provide conversion factors for average stemwood volume into biomass for tropical hardwoods and softwoods (see Table 7.2). The basic density of wood from stumps, roots, and branches may differ from stemwood density. Though absolute amount of biomass per ha varies in different ecological zones, the relative distribution of dry weight by component, according to Crow (1978), seems to be fairly constant. Using the factors given in Table 7.3, the average standing volume of natural hardwood forests can be worked out by adding the contribution of components: Stem biomass, Branch biomass, foliage biomass, and Stump/root biomass. Carbon in soil, litter, and coarse woody debris, may either be monitored or estimated by establishing statistical relationships with forest and site characteristics such as stand age, land-use history, and climate, or with soil maps and land cover data. The method most commonly applied to estimate C at a site is to determine total organic carbon at different depths or globally for one or more horizons; and to transform the collected data in terms of a site estimate, taking into account the bulk density and stoniness of the soil. Until now there has been a lack of good, site- referenced data (FAO 2001). The estimation of C stock of the soil over an area is done at varying scales: the site or plot, the watershed, the region, a speciﬁc country or continent or the agro- ecological zone; and results expressed in the units of C kg/m2, t/ha or Gt (Pg). 7.4 Biomass Functions 69

Table 7.2 Average basic density for stemwood by region (ton/m3) Region Hardwood Softwood Tropical Asia 0.57 0.52 Tropical Africa 0.58 0.45 Tropical America 0.62 0.46 Overall mean 0.59 0.48 Source: Brown and Lugo (1984)

Table 7.3 Breakdown of total tree biomass into components Component Dry weight Dry weight (ton/ha) (%) Stem 265.1 65 Branches 79.2 19 Foliage 7.5 2 Stump/roots 59.3 14 All 411.1 100 Source: Cannell (1982)

A two-step process is used including: i) soil proﬁle analyses at representative sites; and ii) spatial extension of the sample site values using digitised soil map giving estimated areas of different soil units.

7.5 Non-Wood Forest Products

A classiﬁcation of forest produce including goods and environmental services follows: I. Commodities (goods) 1. Wood products • Timber • Fuel 2. Non-wood-products • Leaves/fodder • Fruits/ﬂowers • Bark • Gums/resin • Grass/Grazing • Fauna 70 7 Special Studies

II. Environmental Services 1. Abiotic • Windbreak/shelter-belt • Shade • Soil/water conservation • Carbon sequestration 2. Biotic • Biological diversity • Aesthetic effect Presently, there are no accepted procedures for assessment of most of the non- timber forest-products (NTFP). More research is needed, in the context of developing countries, for making production-consumption balance of NTFP (earlier called as minor forest produce). They are now of major importance for the forest fringe dwellers and other rural population, perhaps much more important than the timber products. Production and extraction of NTFP not only creates more job opportunities, but also opens the way for further local processing and may add to employment opportunities to people (and also conserve the traditional knowledge). Besides, many of these products are necessary for the daily life of the rural population. The procedure for preparing volume or growth functions of trees is well known and documented. But, the planning and implementation of special studies for many of the NTFP is little known.

7.5.1 Fruit/Seed/Pulp Yield

Bhat et al. (2003) monitored yield and yield attributes of 16 non-timber forest products species in the Western Ghats of India, during the years 1997, 1999, and 2000. Average fruit number per tree was estimated and yearly variation in fruit number was compared. Correlation coefficients were computed for yield with its attributes such as girth, crown size, and height. Further, the correlation coefficients among yield attributes were also computed. In most tree species observed by them, there were significant differences in fruit production between years, indicating that tropical tree species exhibit supra-annual cycle of reproduction. An important observation made was that in most species, the yield between two consecutive years was significantly different. Correlations among various yield attributes indicated that, though the coefficients among the yield attributes were consistent, correlation with yield was highly variable, indicating that annual variations may be because of the environmental attributes that influence yield. In the light of these observations, the authors recommend that yield and yield parameters be continuously monitored for at least 5 years, for understanding the patterns of fruiting in these species. 7.5 Non-Wood Forest Products 71

Fig. 7.4 Regression functions for yield of Amla in India (Emblica ofﬁcinalis)

Table 7.4 Average annual production of NTFP in Chhattisgarh, Central India Girth wise (in cm.) Yield of NTFP Per Tree (K.G./Per Year) Species 21–30 31–40 41–60 61–80 81–100 101–120 121–150 151–180 [180 Emblica ofﬁcinalis 2 5 11 20 17 Terminalia belerica 431 Terminalia chebula 316467191 Buchanania lanzjan 0.25 2 3 Syzygium cumini 11 25 42 74 86 Sleichera oleosa 81225437981 Tamarindus indica 848 Madhuca indica (ﬂower) 2 5 26 42 51 56 33 Madhuca indica seed 3 4 15 15 14 4 3 (seed and pulp) Aegle marmelos (pulp) 0.5 3 18 Semicarpus anacardium 0.25 2 5 4

A number of investigations on yields of fruit have been done by researchers in horticulture. In case of a forest species, an interesting result is cited here (Bawa 1995), which shows that fruit yield is closely related to tree diameter (see Fig. 7.4). The average NTFP production per tree in forests of Chhattisgarh India is given in Table 7.4 for information. Overall Comment: Presently, there are no accepted procedures for most non- timber-forest-products (NTFP). More research is needed, in the context of tropical countries, on making of production-consumption balance of the products as they are of major importance for the local dwellers and for the rural development. 72 7 Special Studies

Recommended Further Reading

Bawa K (1995) Economic botany vol. 50 p 274 Bhat, PR, KS Murali, GT Hegde, CM Shastri, DM Bhat, Murthy IK, Ravindranath NH (2003) Yield estimations in certain species of NTFP in Uttara Kannada, Western Ghats, Karnataka, Current Science, 85(9):1350–1355 Brown S, Lugo AE (1984) Biomass of tropical forests: A new estimate based on forest volumes. Science 223:1290–1293 Cannell MGR (1982) World forest biomass and primary production data. Academic Press, London CROW TR (1978) Common regression to estimate tree biomass in tropical stands. Forest Science 24:1:110–114 FAO (1974) A description of methodology and techniques used in the inventory of selected areas of mixed Dipterocarp Forest in Sarawak: DP/MAL/72/009, FAO Rome FAO (1980) Forest vole estimation and yield prediction, By Francis Cailliez, Forestry Paper 22/1. Rome, Italy FAO (1986) Wood biomass—a potential resource for energy and industry by Lars Gunnar Marklund, GCP/RAS/106/JPN FAO (1997) Estimating biomass and biomass change of tropical forests, Forestry paper 134 by Sandra Brown, Rome, Italy FAO (2001) Soil carbon sequestration for improved land management World Soil Resources Reports 96, Food and Agriculture Organisation of the United Nations, Rome Chapter 8 Data Collection

8.1 Classiﬁcation of Data Sources

The data sources in a forest inventory could be grouped as under: • The existing data • Remote sensing and mapping • Special studies • Field survey The first three information sources have been already presented in detail. This chapter will deal with the field survey planning and implementation. A few words may be in order to distinguish survey from studies. The term survey is used, when the objective is to estimate population characteristics based on sampling; whereas special studies (or simply studies) refer to fieldwork to develop equations or relations among forest parameters. Some examples of the latter are: collection of felled tree data and/or standing tree measurements to develop volume equations; estimation of non-timber products such as bark content, seed production, etc., from standing tree measurements; production-consumption studies, accessibility and cost studies in pre-investment surveys, etc. The fieldwork for surveys and studies is generally combined as they are related. The survey parameters enter as predictor variables in the equations developed from studies. Keeping this in view, this chapter presents field procedures for the two cases in a unified manner. In fact, information collected using all four data sources should be compatible to allow smooth integration and analysis. A holistic view is, therefore, called for in collection and compilation of all sources of data.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 73 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_8, Springer-Verlag Berlin Heidelberg 2013 74 8 Data Collection

8.2 Field Plan and Logistics

In case institutions already exist, planning and implementation of fieldwork is a routine affair. In case of de novo operations, however, the task is most challenging from technical as well as administrative perspectives. The field manuals have to be prepared and capacity developed to collect and provide accurate data in a timely manner. The recruitment and training of staff has to be started early, especially when they lack forestry background. The need for high importance given to the training and drill cannot be overemphasized, as the fieldwork constitutes the most expensive and important part of a forest inventory. The inaccessible and inhospitable work conditions in tropical forests add to the challenge. The field director has to cope with a number of mundane issues related to personnel, vehicles, camp establishment, and security. Planning of fieldwork for one of the most intensive forest inventory exercises conducted in Sarawak, Malaysia, with FAO technical assistance is given in Fig. 8.1 to serve as an illustration of field planning. A systematic planning of fieldwork can be started only when the design of survey and studies is final. By that time, the variables to be classified and/or measured and associated sample plot work is known too. To a large extent, the success of a forest inventory is determined by the quality of data collected in the field. If the data have been honestly collected following the field procedures combined with adequate technical supervision, there is no doubt that the final output will be of high quality. The saying goes: ‘‘garbage in, garbage out’’. Figure 8.1 presents a synoptic view of activities involved in forest survey and special studies.

8.3 Field Manual and Field Forms

Field manuals ensure that collection of field data is consistent and accurate. They provide definitions and detailed instructions for classifying and measuring the sample plots-related information. Field manuals include layout of pre-coded field

Fig. 8.1 Overview of field survey operations 8.3 Field Manual and Field Forms 75 forms for data recording and subsequent transfer to a digital media with minimum error or time. Pre-coded field forms and Code Number Manuals help in consistent data recording and subsequent consistent data entry to a computer device. To facilitate data collection, several types of forms are designed: • Plot Description • Plot Enumerations • Sample Tree • Felled Tree • Bamboo Data • Increment • Tree/Forest Products Data Field manuals contain the coding system, description, and use of measuring instruments and instruction on significant digits to be used in filling various fields. The following rules are useful for minimizing the errors of classification and measurements. 1. Define variables properly in the field manual to facilitate correct field measurements. 2. Make sure that variable can be objectively located. 3. Define the entity to be classified (or measured) and the area unit. 4. Clarify the processing/end use of data, its relevance. 5. Limit the significant digits in classes/measurement. 6. Ensure complete coverage of classes. 7. Balance between classes. 8. Reduce subjective influence. 9. Training of staff and control crew work. 10. Editing of field data (in field and office). 11. Post-field discussions (review). Examples of some field manuals available on the web are given in the Refer- ence Section. There is normally a hierarchy of forms for collecting various types of data. For example, qualitative data like forest type, slope, terrain, etc., are defined over a larger area unit than the sample tree measurements like height, crown diameter, etc., which are made on a sub-sampling basis only. Accordingly, different field forms and procedures are designed for recording different data categories. Plot Enumeration data • Species • DBH-OB • Saw log length • Saw log crown diameter OB 76 8 Data Collection

The data collected on a sub-sampling basis generally include the following: Sample Tree Data • Species • Diameter • Total height • Crown projection • Crown freedom • Sawlog length The tropical forest inventories need to take adequate cognizance of three complexities, not present in other regions, viz., multiplicity of species, products, and utilization pattern of forest species by the local communities. As an example, life-forms and species observed in the Dry Deciduous Forests of Odisha, India, occurring in an area of 250 ha are given in Table 8.1. Modern handheld electronic data loggers can improve the quality of ﬁeld data to a large extent since we avoid manual writing errors and can also provide for validity and consistency checks while still on the plot. The portable computer can easily be interfaced with the ordinary PC using remote devices or direct transfer. This is just a suggestion not based on ﬁeld trials.

8.4 Special Studies

The special studies traditionally include the following groups of data: (i) Felled tree data collected during special study including: volume, height, diameter, and species; (ii) Data collected during the sample survey including: height, diameter, and species. Due care is taken that trees selected for felling are well distributed and cover all ecological zones where the species occurs. The number of trees to be felled may be determined from earlier experience or through a statistical procedure keeping in view the cost of collecting volume data and the variance of volume for a given

Table 8.1 Life-form distribution in Compartment 13, Nayagarh Forest Division, Odisha, India Life-forms No. of Spp Stems/Ha Trees ([10 cm DBH) 69 589 Regeneration (\10 cm DBH) 44 91,250 Shrub 22 4,508 Climber 31 23,075 Coppice 51 25,492 Herbs 36 204,323 8.4 Special Studies 77 species and height and diameter measurements. It may be noted that the following manuals need to be prepared to guide the special studies: • Manual for felled tree data collection • Code numbers manual • Felled tree data control manual

8.5 Check-Crew Work

There are two ways to organize the check-crew work. The first is an independent visit to the sample plot. This can be done without crews meeting each other. The other approach is that the check-crew leader and the crew leader meet and discuss measurement and classifications problems together at the plot in order to adapt and rectify ambiguities in the mind of the crews or in the field manual. It is also desirable that the survey supervisor makes one day visits to the crew for education and for a company, as crews work for the whole season alone. The records of check-crews constitute an important feedback mechanism to ensure that the crew have correctly understood definitions and measurement procedures and to ascertain that instructions given in the field manual are clear without ambiguities. A mid-term change in the field manual is not advisable; therefore, initial planning has to be very thorough. It is also imperative that filled- in field-sheets are regularly checked in the camp-office at the end of each workday and data computer-edited soon thereafter. It is the experience of the author that computer editing serves a very useful purpose and might detects errors which may surprise even the best crew. Control of fieldwork is useful in many ways. It helps to identify crew leaders who have not understood definitions well and shows directions to guide them in the work. It also brings uniformity in definition used by various staff and thus improves the quality of forest inventory. Control results can also be used to describe the quality of the information gathered. Finally, in continuous inventories, the control results can be used to improve data collection techniques, field instructions, and identify the training needs, since they reveal weak points. The errors in field measurements could be random or systematic. Systematic errors have serious implications and result in under or overestimation of mean and also contribute to mean square error; whereas random errors tend to decrease with increasing sample size. Both types of errors deserve attention.

8.6 Computer-Assisted Editing and Data Archival Routines

During the ﬁeld inventory, there is a very important need that coded sheets are continuously and promptly checked for consistency and error, both manually and by computer, and a feedback mechanism provided to the concerned ﬁeld staff. 78 8 Data Collection

An appropriate data editing routine, for checking each item of the coding sheet as well as cross-relation among items, needs to be developed. Many errors come to light during editing. In a study, during a project, it was found that input data, before it was edited by the computer, was checked carefully and thoroughly; still, quite a few errors were brought to notice which could not be detected during the manual checking phase. The end product of the ﬁeld survey should be a clean and well-organized database.

Recommended Further Reading

AP (2006) Field manual for forest inventory. Andhra Pradesh Forest Department, India FAO (2003) Field Manual for Philippines National Forest Inventory, Compiled by A. Branthomme, FAO Rome FAO (2004) Field Manual Template of National forest inventory FAO (2009) Manual for integrated ﬁeld data collection. NFMA Working Paper No. 37/E, Rome FSI (2010) Manual for National Forest Inventory of India. Forest Survey of Dehradun, India Malaysia (2001) Field Manual for Forest Inventory Malaysia. 2001, Jabatan Perhutanan. Semenanjung: sustainable forest management and conservation project Chapter 9 Data Processing

9.1 Roles of Data Processing

As a starting point we make a distinction between data processing and automatic data processing. One may think of a forest inventory without automatic data processing, but one can never think of a forest inventory without data processing. The work which is carried out within an inventory project is really—or should be—data processing to a very large extent (Nilsson 1967). In fact, data processing (implying quantitative thinking) begins at the very inception of a forest inventory. At this stage the main concern is to structure the problem and identify the broad issues, define the data needs, and formulate a model for solving the problem. Once the inventory problem has been properly structured and worked out, preparation of inventory design can begin. During this phase, the Data Processing Unit (DPU) retrieves archived information useful for planning of survey design including data on variances, costs, and time information to facilitate formulation of cost-effective sampling designs. A necessary condition for this is that historic job-files are well documented and stored in an easily retrievable manner. Once the survey design has been finalized, the field manual has to be written. This involves work on classification, coding, and recording of data. A good design of field forms facilitates the data collection as well as its processing. Crew leaders have to be properly trained in filling forms in the prescribed format. The experience shows that training of crew members in the elementary data processing is always advantageous in the long run. During forest inventory operations, input data has to be continuously sent to Data Processing Units for editing and providing feedback. Many errors in field data recording come to light through computer editing, even when forms have been thoroughly checked beforehand by the field supervisors. Once the data have been doubly checked, data processing proper can begin. Especially, if data processing steps have been planned in advance, most of the tables and graphics can be printed for direct inclusion in the final report.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 79 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_9, Springer-Verlag Berlin Heidelberg 2013 80 9 Data Processing

This will considerably save the time of report writing and also reduce the likelihood of typographical errors. The foregoing overview highlights the important role, which data processing has in forest inventory.

9.2 Data Processing Operations in a Forest Inventory

Data processing operations in a forest inventory are usually shown with the help of a flowchart, which depicts the flow of data processing operations from beginning to end and linkages between them. From the past experience, the following five phases can be distinguished (see Fig. 9.1): • Manual and computer-assisted editing of field forms • Development of volume functions • Tree volume estimation and plot level summaries • Estimation of means and standard errors by strata • Final tabulations; data storage and retrieval system A brief description of the five phases follows (refer FAO PC-FIDAPS Appli- cation Guide 1988).

9.2.1 Phase I: Manual and Computer-Assisted Editing of Field Forms

The following manual checks can be done by the field crew, before computer assisted editing: • Look for gross errors in data recording • Search for format errors, blank columns, and other invalid entries • Deletion of unnecessary details on the field forms • Rewriting of illegible entries • Additional coding if necessary, e.g., species coding The computer-assisted consistency checking is a very important step. Invariably in the process, many errors in coding and recording of data come to light. Correction of records in error follows. Once all data have been edited and corrected, it is advised to re-run the check programme once again using the entire (corrected) master file to ensure doubly that no errors are noted and the summary of input records provided by plot, stratum, and the survey universe are correct. The clean master file is to be used as input for the next phase of the operations. 9.2 Data Processing Operations in a Forest Inventory 81

Fig. 9.1 Flowchart of data processing operations in a forest inventory 82 9 Data Processing

9.2.2 Phase II: Development of Volume Functions

This phase involves the following data processing steps:

1. Processing of felled tree data including (a) Volume calculation of single trees by types of volume. (b) Scatter diagram of volume against diameter and height. Generally D2His the most chosen compound variable. (c) Regression functions involving volume and a set of compound variables: D2H or DH etc. This step ends with a function of the form: V = f (D, H, Species) 2. Processing of Sample Tree data The sample tree measurements include information on height and diameter and species, intended for developing local volume equations. The main steps are: (a) Editing of data. (b) Calculation of sample tree volume using the chosen general volume functions. (c) Scatter diagram of volume against diameter for the chosen function. (d) Local volume equation trials involving only volume and diameter of the sampled trees. This step results in a function of the form: V = f (D, Species).

The function has only species and diameter as independent variables and can be used for calculating volume of all trees measured during the forest inventory work.

9.2.3 Phase III: Tree Volume Estimation and Plot Level Summaries for Error Calculation

Three tasks may be involved during the phase: (i) tree volume estimation (or any other parameter estimation such as basal area, quality volume) using pre-defined routines or equations; (ii) generating additional field values such as diameter classes from measured diameter; and (iii) summing up plot information by species (as rows) and diameter classes (as columns) on per ha basis using a ‘‘blow-up’’ factor (or expansion factor) for each record to get its contribution in volume per ha terms (see Table 9.1). During this phase, one may also define new statistical ID for the Input records, which may be needed for error calculation and for making post- stratification, when the stratum field might be created on the basis of plot description variables. 9.2 Data Processing Operations in a Forest Inventory 83

Table 9.1 Calculations of stems/ha, volume/ha, and biomass/ha by diameter classes Table 1 Stem volume/ha by diameter classes Table 2 Total stem volume by diameter classes Table 3 Total timber volume by diameter classes and by quality classes Table 4 Biomass weight/ha by diameter classes Table 5 Number of stems/ha by diameter classes and by utilization classes Table 6 Basal area/ha by diameter classes Table 7 Stem volume/ha and number of stems/ha by diameter classes

9.2.4 Phase IV: Estimation of Means and Standard Errors

The routines for mean and standard error calculations can be generalized. It is very important to keep in mind that error calculation is guided by strict statistical considerations and must be carried out in consultation with specialists or following standard text books. Freese (1962) provides a simple description and formulae for calculating mean and standard errors of different sampling procedures. For systematic sampling, as mentioned earlier, there is no statistically valid method of error calculation. Ma- tern (1984) discusses the problem and presents an approximate solution for calculation of the sampling error of a systematic survey. The experience in handling error calculations for FAO inventory projects indicates that most of the survey designs, except Double Sampling can be handled by making use of Ratio Estimators for calculating means and variances, as given below: P y Rb ¼ P i xi

where Rb is sample estimate of mean. yi = value of y for i-th sample xPi = value of x for i-th sample = summation of values 1 to n n = number of sampling units chosen The variance of mean is estimated by the expression: P b 2 ðyi R xiÞ Vb ¼ R nðn 1Þx2 where V = variance of sample mean bR x = value of mean of xi 84 9 Data Processing

9.2.5 Phase V: Final Tabulations and Database Storage and Archival Routines

At this stage, the area of each stratum needs to be entered for weighting per ha figures with areas by using an algorithm. Finally, a tabulation program can be used to produces the final tables. As mentioned earlier, data collected during a forest inventory has a long- term (rather a permanent) value. The information extracted, to meet the immediate needs, constitutes a very insignificant use of information. The inventory data have wide application and constitute a permanent reference record. It is important that they are properly stored in a digital media and well documented including definitions, manuals used, measurement techniques, and other relevant material. It should also be possible to reprocess data and create new output. Data of many forest inventories have been reprocessed by the author after a gap of several decades to produce new and very useful outputs!

9.3 Some Strategic Data Processing Questions

Some questions will be briefly discussed relating to strategy for strengthening of in-house data processing capabilities, computing environment, and use of standard versus tailor-made packages. A vibrant data processing unit is an asset to an organization. The aim should be to build in-house capabilities to do most of the routine data processing tasks and undertake regularly some statistical and GIS analysis tasks, not to forget statistical methods and packages by not using them! With relatively inexpensive PCs and software, it is now possible to input, integrate, and store information in the form of data as well as maps, analyze and manipulate data, and combine these with modeling algorithms and display results in the form of computer graphics, maps, or tables. Together with the word processing capability of computers, the final report, in fact, can directly be produced. Experience shows that larger GIS tasks have to be very carefully prepared if the cost and time associated with a GIS project is to be minimized. In fact, to get the best out of a major GIS project, total planning is required starting with survey and analysis of user’s output requirements, followed by input design, reliable and relevant data collection, and modeling. Keeping this in view, larger GIS tasks could be outsourced for the sake of efficiency. Still, developing minimum GIS and statistical capabilities in-house, in the form of an analytical unit mentioned earlier is an asset, no doubt. 9.4 Generalized Versus Tailor Made EDP Systems 85

9.4 Generalized Versus Tailor-Made EDP Systems

Advent of PCs and communication protocols has brought data processing and data transmission within the reach of a wider group of inventory organizations. However, a distinction needs to be made between a casual analysis and systematic data processing operation involving storage and analysis of a large volume of data on a repetitive basis. For such purposes, there are a number of options: (i) use an existing tested routine; or (ii) adapt an existing routine to meet the specific organizational needs; or (iii) develop de novo a tailor-made program. In this context, it may be noted that program writing and its debugging for a large volume inventory data processing is a time-consuming and expensive operation. It has been said that every program (especially of the size we are discussing here) has at least one error. Even till the end of the data processing phase, some bugs go undetected and cause problems in an unexpected moment. On the other hand, use of an existing program has inherent constraint set by the original program. Adaptations call for careful study of the original program, making of changes, and debugging. It remains a patchwork. Rapid developments and a computer-friendly environment in the computing field offer many advantages, but they also add to incompatibility problem. Work on a generalized Forest Inventory Data Processing System, FIDAPS, was started by FAO around 1973 using mainframe computing and FORTRAN as a language. The product was thought to be a companion to the FAO Manual of Forest Inventory written in 1973. The first version was ready by the middle of the 1970s. By 1980, when mainframe FIDAPS had become operational and a few training courses also had been organized, PCs had already arrived in the market, making the mainframe computing somewhat out-of-date. A decision was, therefore, made at the FAO HQ to develop a PC-FORTRAN version, which became operational by 1985. Besides a FORTRAN version, dbase and FOXPRO versions were also developed by field projects. In 1988, the whole program was rewritten to make it user-friendly. This brief history of a standard package shows problems associated with generalized packages. However, there are benefits of a generalized (in-house) routine because data processing steps in a forest inventory are very similar. Further, it is desirable that output formats of different inventories of the same organization are compatible. In fact, one should be able to expand the datasets across inventory projects in time and space. For such purposes, standard routines may save time and money and also offer compatibility. Further, a generalized system usually undergoes rigorous testing, making it comparatively more reliable and comprehensive. Finally, a generalized system is often well documented, which is of great advantage, when staff changes. 86 9 Data Processing

9.5 Case Study of FAO Forest Inventory Data Processing System (FIDAPS)

The PC-FIDAPS consists of the following programs/sub-routines:

Program Function DENT Data entry and simple checking CHECK Data consistency checking VOLTREE Volume estimation, summation RU-level GROUP2 Species and column grouping, summation SU-level STAT1 intermediate results, summation SS-level SELECT Selection before ﬁnal tabulation, area-merge WEIGHT Weights with areas, accumulates at higher level STAT2 Final tabulation / Error calculations

The following utilities also exist: FIDLIST SUBMENU: Utilities for unformatted data files FIDSORT SUBMENU: Sort / merge utilities FIDLSTID SUBMENU: Utilities for listing ID of records However, the user has to provide information on the volume functions and specify the reporting units (strata) for stems/volume cumulation. PC-FIDAPS cannot be used to construct volume Functions, which should be done separately. The software is written in FORTRAN 77, and consists of about ten different programs. At installation time, a few FORTRAN 77 subroutines had to be written, e.g., computing tree volumes for specified volume functions. The system is open, allowing the users to modify the source code so as to fit their own needs. In addition to the FORTRAN programs, a few utilities of MS-DOS operating system are used. However, the FORTRAN programs can run in another environment, even on a mini- or mainframe computer, if some minor adjustments are made.

9.5.1 Output and Input Speciﬁcations

The INPUT part of PC-FIDAPS accepts various kinds of enumeration data including fixed-area sample plots and variable relascope plots. The basic OUTPUT from PC-FIDAPS are tables where the rows refer to single species or species groups and the columns refer to volume, basal area, number of stems, etc., by diameter classes or by quality classes. The contents of each table are defined by the by the users. It has no limitations regarding the layout of different tables. The table format, using species/species groups as the row, is especially suitable for mixed tropical hardwoods. Tabulation can be made at any chosen levels, e.g., by forest 9.5 Case Study of FAO Forest Inventory Data Processing System (FIDAPS) 87 types within districts, by districts, by the whole inventory area, or by any selection of the lowest reporting units. While developing a generalized sampling error calculations algorithm for FIDAPS, it was found that the formula for estimating means and variances used in Ratio Estimation can be generalized for use in conjunction with most of the statistical designs, except that for Double Sampling. This makes error calculation procedure in FIDAPS very simple. The routine also prints values in different forms (such as standard errors, confidence intervals) along with the estimated values.

9.5.2 PC-FIDAPS documentation

PC-FIDAPS literature includes the following: (1) An Introductory paper; (2) Application Guide, which mainly is intended for inventory specialists (3) PC-FIDAPS handbook, giving ‘‘hands-on’’ instructions about the installation and use of PC-FIDAPS; (4) Programmer’s Guide, written for programmer responsible for a specific installation of PC-FIDAPS. It presents in detail all the options of PC-FIDAPS. • A user’s guide to program DENT • A user’s guide to program CHECK • A user’s guide to program VOLTREE • A user’s guide to program GROUP2 • Modifications of programs STAT1 - STAT2 to fit PC-FIDAPS (5) PC-FIDAPS case studies; (6) the FORTRAN 77 source code. As with any other software package, PC-FIDAPS needs to be regularly improved and updated.

9.5.3 Concluding Remarks

Statistical analysis and Data processing are twin activities. They go together. It is important that ﬁeld data is transformed in the form of a database soon after it has been ﬁeld checked and edited. Database management systems like EXCEL are good for initial editing and data transport, but the need for more powerful packages will become soon apparent. Microsoft ACCESS could be effectively used in less advanced data processing environment and for most database management operations including tabulations and error calculations. However, statistical analysis is also very important. Schreuder et al. (2004) recommend R-language for this 88 9 Data Processing purpose, because it provides an extensive array of well-documented statistical procedures in English, French, and Spanish and can be freely downloaded from the R-Project web site: www.r-project.org.

Recommended Further Reading

FAO (1967) Some views on data processing problems in forest inventories by N. E. Nilsson in HQ. Meeting of Forest Inventory Experts of UNDP/SF Projects GOI (1975) Data processing report. Preinvestment Survey of Forests, East Godavari report, FSI, Dehradun FAO (1973) A description of the data processing system used in the inventory of selected areas of the mixed dipterocarp forest in Sarawak. Malaysia. Working Paper No. 22. FO: DP/MAL/72/ 009 FAO (1978) The role of electronic data processing in forest inventory presented by K. D. Singh at FAO/SIDA/GOI seminar on forest resources appraisal in forestry and land use planning, Delhi/Dehradun FAO (1980) FIDAPS—a forest Inventory data processing system by K. D. Singh and J. P. Lanly, FAO Rome FAO (1988) PC-FIDAPS Application Guide, FO: MISC/88/11 Freese F (1962) Elementary forest sampling. USDA Forest Service, Washington Matern B (1984) Four lectures on forest biometry. Swedish University of Agricultural Sciences,Umea Chapter 10 The Report Writing

10.1 General Comments on Reporting

The report on findings of the survey constitutes the final phase of an inventory. Ideally, the report (viz. the last phase) should relate to problem formulation (viz. the first phase) and contain the background information or the concerns, which led to the inventory, followed by formulation of objectives, information need assessment, organization and analysis of the existing data, design of survey, data collection, data processing, and main findings. Processed tables could preferably be kept as a part of the Technical Annex, whereas the Technical Report should contain key findings of the inventory (including its reliability) and how they contribute to the decision-making process. The following narrative is taken from the report of a recent inventory project of the author in Odisha, India to evaluate community-based initiative to manage their forests on a sustainable basis (Singh et al. 2008). The study/report had two-fold objectives: (i) Evaluate the community achievements in respect of forest protection in terms of changes in forest area by density classes; and (ii) Draw inference from collected data regarding future forest management with special reference to three issues directly affecting the community, viz., (i) the sustainability of firewood removals; (ii) sustainable level of Sal- leaf removals; and (iii) conservation of overall biological diversity in terms of species composition.

10.2 Forest Inventory Problem Formulation

Budhikhamari Forest Federation was constituted in 1995 by 95 Communities in the Baripada Range of the Mayurbhanj district of Odisha. The objective was to protect and manage forests under their control in a democratic environment. The geographic area under consideration is 25,570 ha and the forest area 5550 ha as

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 89 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_10, Springer-Verlag Berlin Heidelberg 2013 90 10 The Report Writing assessed in the year 2000. With the total number of households at 5,800, per family holding comes to 1 ha/household approximately. The Federation has been in existence now for more than 15 years. However, there has been no reliable assessment of the forest changes using scientiﬁc methods such as multi-date remote sensing and/or statistically planned survey, to validate community achievement. Besides assessing the state of forests, the goal was also to develop sound (and simple) methods of inventory and planning, do-able by the community and enhance their capacity for taking conservation and management measures and in the process generate additional income in a sustainable manner. The multiple objectives called for establishment of a dynamic database to estimate the level of stocking and level of harvesting (which have implications for availability of resources for subsistence and/or commercial purposes), assessment of present and future subsistence needs, market demand, and prevailing prices. Further, yield estimates based on data of long-term research plots, were required to determine the sustainable level of harvesting from the present forests, without jeopardizing the future yield. This is how the concept of sustainability could be implemented in the practice.

10.3 The Statistical Planning

The traditional forest inventory and forest management techniques have been developed for application over large tracts of forests and large markets. When applied to community forest situations, they suffer from serious limitations because the management objectives in the latter case are location speciﬁc and require information on forest, community, and market in a connected manner, from where people collect forest produce, how much they consume, and where they sell. To cater to such information needs, a new concept of survey was developed, which views a forest patch, associated community, and market as one holistic entity, termed as Integrated Primary Sampling Unit (IPSU) as shown in Fig. 10.1(Singh et al. 2009). Objectives of the Survey: The Objectives of the survey, in the light of foregoing background, was established as follows: • Estimate the changes in the forest area and condition during the last 20 years; estimate the present consumption level and potential available to the community for enterprise development; • Assess the sustainable level of the yield versus actual consumption of forest products, market demand, and prices; and • Build expertise in the community to carry out such assessments in the other forest areas and updating the survey in their own areas at a later date. Population to be sampled: The population to be surveyed consisted of all patches of forests within the limits of Budhikhamari Forest Federation; associated 10.3 The Statistical Planning 91

Fig. 10.1 Integrated primary sampling unit (IPSU)

95 villages; and the markets, where products were sold. Using satellite images and visit, a location map of all such units was made in order to define the sampling universe. Data to Be Collected: The data to be collected consisted of four parts: Forest area of patches by density classes along with changes, if any, during the last 20 years; stocking of timber and non-timber species in surveyed patches; household consumption; and produce sold in the market. Degree of Precision required: The overall standard error of volume per ha in the survey area of trees above 10 cm diameter was expected to be around ± 5%. The Survey Techniques: The method consisted of three phases: remote sensing survey, field inventory, and special studies, as described here. The remote sensing and GIS consisted of the following steps: interpretation of multi-date remote sensing images, preparation of change map and its validation in the field, overlaying of village boundary layer, and associating attributes like census and sample survey statistics to the forest maps. The final output was a comprehensive statistical and spatial database on forest cover state and changes during the periods 1985, 1990, and 2000 by forest patch and associated villages. This formed the basis for forest inventory (Fig. 10.2).

10.3.1 The Sampling Design

In statistical parlance, assessment can be described as a stratified random sampling. Five distinct patches of forests were identified through remote sensing. The patch (1) after the field visit was further divided into four sub-strata. Thus, in all, eight strata were identified and assessed independently. 12 strip plots were systematically laid in each stratum. 92 10 The Report Writing

Fig. 10.2 Methods of remote sensing and GIS

The existing data of Mudumalai Forest Dynamics Plots was used to configure the ideal sample plot layout, which was found to be a strip plot 10 9 100 m. This was divided further into ten recording units (also called sub-plots) of the size 10 9 10 m. Within each patch (stratum), 12 sample plots were laid systematically at intervals of 100 m, both along and between the lines with a random start. The following data were collected on different life-forms using specified sub-plot sizes: • Trees, stumps (cut trees), samplings, and climbers on all sub-plots 1-10; • Seedling, coppice shoots, and shrubs on sub-plots 1 and 10 only; and • Herbs in four 1 m 91 m quadrates laid in sub-plots 1 and 10, two in each of them as shown in Fig.10.3. The collection of herbs and shrub data, in particular, species identification presented a challenge. The problem was solved by using an elderly knowledgeable person in the villages to identify plants by local name and making a herbarium, which was used for identification/verification by the botanist for the scientific names. Data on trees outside forests (TOF) were also collected. This refers to agricultural bunds, homestead, and agricultural plots adjoining the forest area. For this purpose, line-plots were used in continuation of and in the same orientation as forested plots, viz., NW direction. Instead of rectangular shape, plots were circular with an area of 100 m2, laid 100 m apart extending 1 km in length. 10.3 The Statistical Planning 93

Fig. 10.3 Recording unit layout for different life-forms

10.3.2 Special Studies

The household consumption survey was carried out in 13 villages of IPSU-1 and only one village in each of other IPSUs, i.e., in all 17 villages. A low sampling intensity in other IPSUs was used in view of the low level of variability of NTFP usage among villages. The variables measured are listed in Table 10.1. As Sal leaf collection and marketing was taking place on a large scale throughout the Sal-belt all over the state and in the other adjoining states, there was a growing concern about overexploitation of leaves, beyond the carrying capacity of the forests and the negative effect it would have on the ecosystem and the leaf production itself. Keeping the controversial nature of the issue, a reliable methodology was developed to estimate the Sal-Leaf production and consumption.

Table 10.1 Database on Village NTFP consumption and marketing Community ID Number household Species Local consumption Commercial use Price/unit Quantities sold by market centers 94 10 The Report Writing

10.4 Main Findings of the Survey

10.4.1 The Land Cover and Forest Changes

The community-led efforts toward protection started in 1985. The change assessment was therefore carried out with 1985 as the baseline. A problem was encountered as 1985 satellite images belonged to Landsat MSS generation and were found not suitable for monitoring changes of the type under consideration. Fortunately, the Forest Survey of India had in their archive forest type/cover maps for 1985 obtained from interpretation of aerial photographs, which was procured to constitute the baseline. The relevant information is given in Table 10.2 along with comparable information for the year 2000 based on study of TM images. A comparison of statistics for the two dates reveals that the forest cover in the study area increased from 16 % in 1985 to 22 % in 2000; and more important, the proportion of closed forests increased from 26 % in 1985 to 62 % in 2000. These are very remarkable achievements, remembering that the decision to protect forests is mostly made by the individual hamlets; the Forest Federation and the Forest Department are mainly regulatory agencies. This shows a very strong community inclination, most likely a part of their tradition, to protect forests. The community forests not only increased in area, but seemed to be putting on more volume increment and more foliage production. For a 20-year Sal coppice crop, according to yield table, most of the trees should fall in the diameter class 10–20 cm, whereas in the present case, the size classes 30–40 cm and higher have signiﬁcant presence (see Table 10.3). The distribution of stems by diameter class in various strata provides interesting information on the history of forest protection. Sri S. Jee, the Working Plan ofﬁcer of the Mayurbhanj District for the period 1953–1954 to 1972–1973, wrote as follows: ‘‘Over exploitation, depletion of mature Sal trees and steady encroach- ment of miscellaneous species have been described in the outgoing plan to be the

Table 10.2 Land cover and forest type changes in the study area during 1985–2000 Land cover/Forest type 1985 2000 Ha (%) Ha (%) Closed forests 1,100 4.3 3,413 13.4 Sal (more than 50 %) 800 3.1 Sal mixed (21–50 %) 300 1.2 Open forest (coppice regeneration) 3,132 12.2 2,133 8.3 Forest (closed 1 open) 4,232 16.4 5,546 21.7 Scrub 40 0.2 500 2.0 Pastureland/wasteland/grassland 19,140 74.3 10,899 42.6 Agriculture and water 2,360 9.2 8,626 33.7 Total 25,772 100 25,571 100 Source 1985 data is from aerial photographs and 2000 data is from LANDSAT TM of 2000 10.4 Main Findings of the Survey 95

Table 10.3 Stocking in community managed forests (CF) compared to yield table (YT) IPSU FPU Area Age Crop Diameter cm Stems/ha Volume m3/ha (ha) height (m) Years CF YT CF YT CF YT CF YT I–A Badsul 1,505 19 12.8 12.1 18.6 12.2 575 1,419 119 99 I–B Budhikhamari 1,570 19 11.3 12.1 14.3 12.2 1,305 1,419 111 99 I–C Jatipur 227 19 12.1 12.1 15.3 12.2 1,346 1,419 138 99 I–D Swarupvilla 940 19 11.9 12.1 17.0 12.2 440 1,419 65 99 II Masinakati 578 16 10.4 11.0 20.1 10.1 508 1,308 119 77 III Sankhajod 183 13 11.3 10.4 10.4 10.1 1,172 1,308 102 51 IV Hatikote 149 20 12.6 12.6 14.0 12.9 1,183 1,455 111 117 V Kailashchandpur 123 16 12.3 10.4 14.7 10.1 960 1,308 99 77 Source Value for average yield tables for coppice Sal is taken from Chaturvedi and Sharma (1980) results of the work under leases’’. Jee quotes the inspection report of Mr. R. C. Ramsay, the political Agent to the Orissa Feudatory States during his visit in 1920–1921: ‘‘The forests contain many ﬁne trees and there is a vast quantity of Sal timber available, but at the time it is apparent to any one that the State will sooner or later—and rather sooner than later—be faced with a long period, when there will be no Sal of merchantable size’’. This forecast came true for the state forests in general. But, the historic trend seems to have been reversed in the community managed forests in the last 20 years!

10.4.2 The Condition of the Forest Floor

Statistics on the forest ﬂoor vegetation is given in Table 10.4. This includes seedlings (less than knee-height), saplings (more than knee-height but less than DBH of 10 cm), and coppice shoots, both Sal and non-Sal. They all add to the biological diversity (given in parenthesis for all life-forms) and to the resource potential for the community livelihood. The distribution of seedlings and saplings of Sal shows that forests are relatively well protected. Noteworthy is the number of Sal coppice shoots, which are the main source of leaf to manufacture and trade cups and plates as the main source of cash income.

10.4.3 Trees Outside Forests

The community grows trees in the homestead gardens and agriculture ﬁelds to complement fuelwood and NTFP gathering from the forests (see Table 10.5). 96 10 The Report Writing

Table 10.4 Frequency of seedlings, saplings, and coppice shoots of sal and all species IPSU FPU Sal (stems/ha) All (stems/ha) Seedlings Saplings Coppice Seedling Saplings Coppice I–A Badsul 9,360 126 5,985 21,150 (31) 566 (23) 6,285 I–B Budhikhamari 3,785 248 4,405 19,725 (17) 465 (19) 6,825 I–3 Jatipur 3,755 89 4,600 19,755 (21) 1,109 (22) 6,915 I–4 Swarupvilla 9,970 628 7,450 22,570 (31) 785 (16) 8,050 II Masinakati 36,890 467 6,757 82,930 (22) 1,169 (16) 7,727 III Sankhajod 14,040 596 4,893 32,545 (20) 997 (19) 6,528 IV Hatikote 3,715 321 4,862 18,525 (13) 2,367 (18) 5,997 V Kailashchandrapur 3,585 235 5,488 18,640 (18) 2,509 (23) 6,018

Table 10.5 Trees outside forests in the study area Botanical name (local name) Stems/ha Mean diameter (cm) Mean height (m) Volume number (m3/ha) Terminalia bellerica (Bahada) 10 30 12.0 3.7 Buchanania lanzan (Char) 160 19 6.0 11.4 Pogamia pinnata (Karanj) 40 19 6.0 2.7 Diospyrus melanoxylon (Kendu) 15 12 4.0 0.2 Madhuca indica (Mahula) 295 22 9.0 39.5 Azadirachta indica (Neem) 25 20 7.0 2.3 Butea monosperma (Palash) 50 15 7.0 2.3 Shorea robusta (Sal) 90 29 11.0 27.6 Tamarindus indica (Tentuli) 30 17 5.0 1.2 All species 715 21 7.9 91.0

10.4.4 Comparison with Other Forests in the District

Table 10.6 presents a distribution of the number of stems by diameter classes observed in a district-wide survey covering an area of 428,391 ha conducted by FSI in 1992–1993 along with ﬁndings of the present survey for ‘‘All species’’ and ‘‘Sal’’. The forests of the district (as whole) are characterized by very adverse biotic factors. As much as 71 % of the area was found affected by grazing, 41.63 % had inadequate regeneration, and 59.14 % was subjected to heavy soil erosion, which all contribute to a lower stocking of state forests in the lower diameter classes. This reinforces the ﬁndings that community protection of forests under Budhikhamari federation has been quite effective. The species diversity was relatively higher in the community forests than the forest areas in the district. Therefore, it could be inferred that implementing a management regime as in the community forests will help in moving towards a well-stocked forest condition with a higher Sal proportion and overall higher number of species and stems per ha. 10.4 Main Findings of the Survey 97

Table 10.6 Stand structure in the forests of the Mayurbhanj District and community forests Diameter Mayurbhanj district Study area class (cm) All Sal All Sal (Stems/ha) (Stems/ha) (Stems/ha) (Stems/ha) 10–15 119 71 52 31 15–20 26 12 772 626 20–25 15 6 103 85 25–30 9 4 7 3 30–35 6 4 2 1 35–40 3 2 1 0 40–50 4 2 1 0 50–60 1 0 0 0 All 183 101 938 746 Source Forest survey of India report for Mayurbhanj District (1992)

10.4.5 Livelihood and Resource-Use Pattern

In terms of the monetary value, fuelwood dominates with 56.5 % of the total beneﬁt, followed by Sal leaves sale constituting 27.2 % of the total (see Table 10.7). The main resource used by the community, viz., ﬁrewood, will be further discussed from a sustainability angle in the following sections.

Table 10.7 Resource-use pattern in the study area based on household survey Main Items/species Quantity (Kg/hh/yr) Monetary value (INR) Scientiﬁc Local name Subsist Commercial Total Subsist Commercial Total (%) name All tree species Fuelwood 3,545 384 3,929 5,317 577 5,894 56.54 Shorea robusta Sal leavesa 5,417 354,385 359,802 66 2,769 2,602 27.20 Dioscorea spp. Pani Alu 47 12 59 379 95 474 4.55 Agaricus spp. Chhatu Mushroom 6 3 9 247 98 345 3.31 Madhuca Mahua 2 16 18 20 125 145 1.39 latifolia Flower Others 462 269 964 7.01 Total (32 items) 6,491 3,933 10,424 100.00 a Values are in absolute terms; 95 % of Sal leaves collected are used as pali and only 5 % are used as plate 98 10 The Report Writing

10.4.6 Fuelwood Gathering and Sal Leaf Plucking

Fuelwood need in the study area is generally met from the dead and fallen trees and branches and to a limited extent from the standing stock. The important question is whether the current level of wood harvesting is sustainable. Is the current harvesting level leading to improved stocking of the forest or resulting in its systematic reduction? These questions are analyzed with the help of statistics presented in Table 10.8 which gives yield table data along with actual volume for the same age and site quality. The actual volume except for one IPSU is consis- tently higher than those given by the yield table. This shows that community forest management has been rather conservative. This fact also became very apparent during the discussions with the community members, who seemed less interested in timber harvesting, but more in annual use of NTFP including fuelwood. The following model proposed by Nilsson (1992) was used to estimate the current level of mean annual increment in different IPSU.

b Ia ¼ Im ðÞVa=Vm where I stands for increment and V for the growing stock and the subscripts ‘‘a’’ and ‘‘m’’ refer to actual and model values. The superscript ‘‘b’’ is an adjustment factor arrived by experience, usually it has value 0.6. The estimated value of the actual increment ranges from 7.9 to 11.5 m3/ha/year depending on the stand parameters. The safe limit for annual removals (in the form of thinning yield given in column 8) could range from 5.2 to 7.1 m3/ha/year. Converting these values in weight terms, using a factor of 0.7 (Pandey 2002), the yield in the form of thinning ranges between 3.6 and 5.3 tons/ha/year with mean around 4.7 tons/ha/year. The level of current use as ﬁrewood (see Table 10.7)is 3.9 tons/ha/year, which shows that the current ﬁrewood use is within the sustainable limits, except for one IPSU.

Table 10.8 The standing volume, actual increment, and prescribed treatment by strata as per the yield table I Forest Age Volume Increment Recommended P protection unit (years) (m3/ha) (m3/ha) management (m3/ha) S Yield table Actual Yield table Actual Thinning Final felling U yield yield I A Badsul 19 90 119 9.3 10.7 7.1 3.6 I B Budhikhamari 19 90 111 9.3 10.3 6.8 3.5 I C Jatipur 19 90 138 9.3 11.5 7.6 3.9 I D Swarupvilla 19 90 65 9.3 7.9 5.2 2.7 II Masinakati 16 80 119 8.7 10.6 7.0 3.6 III Sankhajod 13 68 102 8.5 10.4 6.9 3.5 IV Hatikote 20 94 111 9.3 10.1 6.7 3.4 V Kailashchand 16 80 99 8.7 9.7 6.4 3.3 Source Chaturvedi and Sharma (1980) 10.4 Main Findings of the Survey 99

The stump (cut trees) inventory (see Table 10.9) gives further evidence of green removals from the forest. It may be noted that stump inventory accounts for only the removal of green trees from the forest; the dead and dying trees constitute another source for meeting the fuelwood needs of the community. The sal leaf balance is presented in Table 10.10. Out of 8 months of active collection of Sal leaves by the community, the bushes constitute the main source. The collection from trees is conﬁned to about 1–1.5 months, when bushes do not carry adequate leaves. It is estimated that 80 % of the total consumption of leaves is from bushes and 20 % from trees more than 10 cm diameter and saplings (see Table 10.9). The values of potential production for trees and saplings have been combined in Table 10.10. The analysis (using higher estimates in Table 10.10), shows that the use of Sal leaf from trees and saplings varies from 6 to 12 % of the total tree/sapling leaves produced in a year. In other words, an average of about 6–12 leaves out of 100 leaves. As Sal trees shed leaves on a cycle of 10 months, the level of foliage removals is unlikely to affect tree growth. The extent of Sal leaves extraction from the bushes is generally 4–8 %; however, in open forest strata, it could be as high as 30 %, as the number of bushes is very high. The community, speaking in general terms, does not make a bush bare, but collects only 20 % of leaves at a time,

Table 10.9 Stump inventory to estimate fuelwood removal IPSU Stratum D (cm) Stems/ha H (m) Volume (m3/ha) I–A Badsul 17.1 18.8 12.8 2.4 I–B Budhikhamari 13.4 3.8 11.3 0.3 I–C Jatipur 14.0 1.7 12.1 0.1 I–D Swarupvilla 15.8 27.3 11.9 2.8 II Masinakati 14.7 18.8 11.8 1.6 III Sankhajod 13.3 8.6 11.3 0.6 IV Hatikote 12.7 10.6 12.6 0.7 V Kailashchandrapur 13.4 21.6 12.3 1.6

Table 10.10 Sal leaf production-consumption pool by life-form I Sal leaf consumption by Production from Production potential Leaf harvest from P origin (Kg) Coppice bushes from trees and trees and saplings S saplings (Kg) (%) U Total Coppice Trees Potential Harvested Higher Lower Higher Lower bushes and (Kg) (%) estimate estimate estimate estimate saplings I 922 737 184 2,576 29 3,494 2,966 5 6 II 377 302 75 5,972 5 2,611 1,972 3 4 III 132 106 26 2,955 4 3,718 3,085 1 1 IV 282 226 56 2,905 8 3,504 3,195 2 2 V 1,456 1,165 291 3,919 30 2,710 2,504 11 12 100 10 The Report Writing i.e., revisit a bush 5 times in a year. The higher extraction of Sal leaves within IPSU I & V is because of proportionately higher population pressure, open forest condition, and also because of close proximity to the road. More scientiﬁc studies are required on the subject, as the issue is important for the local livelihood. Through this analysis, one could conclude that current extraction of Sal leaf is not causing degradation (except one stratum) or impacting on growth of Sal trees. Any further necessary steps towards processing of Sal leaves should enhance the livelihood without compromising on sustainability.

10.5 Survey Evaluation and Recommendations

The statistical evaluation of the survey design and main ﬁndings presented in Table 10.11 show that the survey design achieved the stated precision goals and provided estimates of the overall mean volume per ha with a standard error less than 5 %, as originally planned. The entire survey operations were implemented by the community, except remote sensing and data processing, consistent with local capacity building objectives. The survey provided useful information concerning effectiveness of community forest management, possibilities of community-based enterprise development and replicating such surveys in other forest areas, which were the main goals of the project funded by the Department of Science and Technology, India. Three kinds of livelihood initiatives are recommended for the future: (1) Technical support to update remote sensing survey and forest inventory every 5 years so that changes in growing stock, increment, and regeneration status of NTFP and timber species can be obtained to guide the community on proper level of harvesting and undertaking of natural regeneration measures. (2) Marketing and enterprise development requires that the Forest Department and Rural Development Departments work hand in hand so that (i) resources

Table 10.11 Estimated mean values and precision of stems and volume per ha by strata (dbh [ 10 cm) IPSU Forest protecting unit Area Stems Volume CV SE Number of (ha) no/ha m3/ha % (%) plots I–A Badsul (degraded) 1,505 575 119 97 28 12 I–B Budhikhamari (good) 1,570 1,305 111 21 6 12 I–C Jatipur (good forest) 227 1,346 138 20 6 12 I–D Swarupvilla (Degraded) 940 440 65 94 27 12 II Masinakati (good) 578 508 119 37 11 12 III Sankhajod (good forest) 183 1,172 102 28 8 12 IV Hatikote (good) 149 1,183 111 31 9 12 V Kailashchandrapur 123 960 99 38 11 12 (good) All All strata 5,225 910 113 36 4.6 96 10.5 Survey Evaluation and Recommendations 101

utilized are what is ecologically sustainable; and (ii) the benefit of processing and enterprise based on forest resources flow to local communities. This will require needed capacity building of the community, pre-feasibility/feasibility studies, and investment support to the community. (3) Finally, the governance in the forest fringe areas must become responsive to link other development measures like education, health, banking, etc., to promote all-round development of the community. Only then, the poverty in the region can be alleviated in a time bound manner. This was the key consideration to open discussions with concerned district development authorities including the District Magistrate to sensitize the officials to take progressive measures in this direction. Ideal would be to nominate the Forest Department as the coordinating agency for all developments in the forest fringe areas.

Recommended Further Reading

Singh KD, Sinha B, Sutar P (2009) Community based forest management with micro-enterprise for poverty reduction. Bisen Singh Mahendra Pal Singh, Dehradun Part III South–South Cooperation

Abstract This part highlights the role of South–South Cooperation in the capacity building process. This is obvious from the common patterns of spatial variation observed in the main forest types across the continents; and comparable social and economic situations in the countries, indicating that countries of the region may collaborate with mutual benefit and learn from the past experience of each other including: the use of different sampling designs; applications of remote sensing technology; growth and yield studies; assessments of environmental functions of forests; and land evaluation techniques. Chapter 11 Common Patterns of Spatial Variations in the Tropics

11.1 Similarities in Forest Formations Across the Continents

The ﬁndings presented in the chapter are based on research conducted by the author on the common patterns of spatial variation in tropical forests across the three continents. What seems to be a bewildering chaos of vegetation does have in fact an organized structure. Ecological studies also show evidence of striking similarities among the formations. We will take in this chapter the most complex forest type, viz., the rainforests as an illustration. According to Richards (1971), they are composed of similar synusiae or communities of tree species with homogeneous characteristics and show great similarity in their spatial arrangement. Species of corresponding synusiae in different geographical regions, besides being alike in life-form, are to a considerable extent alike in physiognomy (see Fig. 11.1). The fundamental pattern of structure is thus the same throughout the whole extent of the rain forest. Thanks to this, the knowledge acquired in one region or country could serve a very useful purpose for survey planning in another region or country. The chapter would limit to rain forests, the most complex of all forest formations. Similar information is available on forest formations in the drier regions of the tropics in the FAO Forestry Paper 51/1.

Fig. 11.1 Distribution of tropical rain forest in the equatorial belt (Source: Smithsonian)

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 105 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_11, Springer-Verlag Berlin Heidelberg 2013 106 11 Common Patterns of Spatial Variations in the Tropics

11.2 Macro-Variation Patterns and Their Signiﬁcance for Stratiﬁcation

The factors causing major changes at formation level can be divided into two broad groups, natural and cultural. Among the natural factors climate is relatively uniform. Soil and topography appear as the main agents for causing variation. Some common types are:

• Climax rain forest (usually on well-drained soils and below elevations of 500–800 m). • Mangrove forest. • Swamp forest. • Periodically flooded forest. • Riparian forest. • Hill forest (usually at altitudes above 500–800 m). • Forest on tropical podsols. • Forest on coastal sand. Cultural factors (positive or negative influence of man) cause further changes in the natural rain forest vegetation. The following broad types of land use can be distinguished: • Primary rain forest (influenced less by man). • Logged-over forest. • Secondary forest (resulting from abandoned shifting cultivation). • Mixed forest-grassland association. • Managed forest (natural and manmade). • Other non-forest classes. The gains in precision reported due to stratification by the FAO Sarawak Forest Inventory Project at macro-level are shown in Table 11.1 (FAO 1973). The precision refers to the net industrial volume of all species 18 in. (45 cm) and more in diameter (95 % confidence level).The gain of precision has a range from 20 % to almost 100 %, a very significant gain indeed.

Table 11.1 Gains from stratification Sampling scheme Precision estimates for the survey by region (%) 1234567 8 Un-stratified sampling 6.7 8.1 9.6 6.6 7.2 7.9 10.4 10.0 (Post-) Stratified 4.5 6.8 7.6 4.0 3.7 4.2 5.4 5.6 11.2 Macro-Variation Patterns and Their Significance for Stratification 107

11.3 Meso-Variation Patterns and Their Signiﬁcance for the Sampling Design

Meso-variation is a relative concept. Here it refers to variation within strata, more speciﬁcally variation among sampling units within strata. This depends on the survey area under consideration as well as on the size and shape of the sampling units. The design objective is to reduce the variance component between sampling units as compared to that within. In the planning phase of the Sarawak forest inventory an analysis of variance was carried out to discover the relative component of variation within and among sampling units within strata, to determine the optimum number of sub-plots in the cluster. The sampling unit consisted of a nine-sub-plot cluster, as shown in Fig. 11.2. Data from two inventory units (1 and 3) were used, covering an area of 176,000 and 225,000 ha respectively (Table 11.2). 2 2 2 2 The value of R , computed as Sb /(Sb ? Sw ), was found to be 0.3, showing relatively low variation between sampling units compared to the variations within a sampling unit. A similar high level of variation within individual sampling units was also observed by the author’s analysis of FAO’s survey data of peninsular Malaysia, implemented later. For the rain forest of the Amazon valley, Ramos et al. (1972) found the value of R2 to be 0.3 in one case and 0.5 in another. The sampling unit consisted of four closely spaced rectangles of 250 9 10 m located 10 m from the cluster center and pointing in all four directions. Boon (1962) studied the variation pattern of Eusideroxylon zwageri in Indo- nesia by dividing the survey area into blocks of variable sizes, 2 9 1, 4 9 1, and 4 9 2 km. Each block was further split into two sub-blocks to compare the variance among and within the blocks. Analysis showed that variance between the blocks was not signiﬁcantly different from that within blocks. This was observed for all the blocks and all the plot sizes.

Fig. 11.2 Nine-point cluster, Sarawak Forest Inventory 108 11 Common Patterns of Spatial Variations in the Tropics

Table 11.2 Analysts of variance for gross volume/acre 1 of all species 45 cm ? diameter Source of Degrees of Sum of Mean sum of Variance variation freedom squares squares components 2 Among samples 81 431448800 5326528.3 Sb = 474223.86 2 Within samples 656 694383300 1058511.1 Sw = 1058511.10 1Mean: 2,031 ft3/acre (142 m3/ha)

Giudicelli et al. (1972) studied the spatial correlation for the number of stems/ha for three important species and one species group in a forest area of 1,500 ha in Cameroon using a plot size 50 9 50 m. The ‘‘correlogramme’’ observed by them is given in Fig. 11.3, showing high variability within small distances in respect of all the species. Furthermore, curves show concavity toward the origin. In such cases systematic sampling is expected to be cost-effective than random sampling. The above results clearly bring out an important feature of the rain forest: high variability within small areas and relative uniformity over large areas. This indicates that the use of some type of cluster sampling may be cost-effective. While reviewing the book, Prof. Nilsson informed about the use of the term ‘‘autocorrelation’’ in the Swedish national forest inventory. The autocorrelation between plots is much higher in north Sweden than in south, for both forest land and no-forest lands. He informs as follows: ‘‘We arrived at 100 m between plots in region 5 (southernmost Sweden) and 500 m in region 1 (upper Norrland). The size of the sampling unit (viz. a tract) was 1,400 m in region 5 with 28 plots in each and 2,200 m in region 1 with 16 plots in each unit. These data are as originally used. Later, the tracts were reduced for cost reasons resulting in higher standard error’’.

Fig. 11.3 Correlogram for selected species in the semievergreen forest of Cameroon 11.3 Meso-Variation Patterns and Their Signiﬁcance for the Sampling Design 109

11.4 Micro-Variation Patterns and Their Signiﬁcance for Plot Size and Shape

The complexity of the micro-variation depends to a great extent on the forest formation and the species grouping. To study the influence of the size and shape of the sampling unit on precision, a pilot survey was carried out in India. There were 39 samples in all spread over two forest types, covering an area of about 80,000 ha. Results showed that in the rain forest both size and shape had important influences on the coefficient of variation. On the other hand, in the adjoining moist deciduous forest, size and shape had little effect on precision, as shown in Table 11.3. Survey experiments carried out in the rain forest of Indonesia, with 20 m wide rectangular plots of variable length, showed that coefficient of variation decreased exponentially with increase in the plot size. It tended to approach a stabilized value of 16 % beyond a size of 1.5 ha (see Nyyssönen, 1961). Heinsdijk (1961) investigated the spatial arrangement of species in the Amazon valley. He found that the distribution of many species approached a Poisson distribution pattern but there were often exceptions where sampling units showed a high number of particular species, indicating a tendency toward contagious distribution. Utilizing the survey data from the Amazon valley, Surinam and Sarawak, the author carried out a series of investigations. For the most important 10–20 species, the mean number of stems per sampling unit (m) and its coefficient of variation (cv) were plotted, and the relationship between two parameters was studied. For a strict Poisson distribution the points relating to mean and coefficient of variation should fall on a straight line inclined 45 to the X-axis. This is because of an important feature of a Poisson distribution which is mean equals variance. In all cases, the species tended to deviate upward from the expected straight line relation, indicating grouping tendencies. Investigations of Boon (1962) for Eusideroxylon zwageri show that this species has a marked grouping tendency. Jack (1961) studied the spatial arrangement of

Table 11.3 Effect of size and shape of the sampling unit on coefficient of variation in Kerala, India (the character under study is total volume of all species) Forest type Mean volume Coefficient of variation (%) Sample size m3/ha Square plot Rectangular plot 03 ha 0.6 ha 0.3 ha 0.6 ha (10 9 300 m) (20 9 300 m) Rain forest 393 98 82 75 48 14 Moist deciduous 177 63 60 60 58 25 Source Unasylva (1973) 110 11 Common Patterns of Spatial Variations in the Tropics the seven most important commercial species of the semievergreen forest of Ghana, using sampling units of various sizes and shapes. For Triplochiton scleroxylon, Khaya ivorensis and Entandrophragma cylindricum he found a significant departure from the Poisson distribution and a good fit with negative binomial distribution. Entandrophragma utile and Mimusops heckelli showed grouping to be observed only with plot sizes larger than 2.0 ha, whereas Chlorophora excelsa, with change of plot sizes, exhibited a vacillating pattern between random and contagious. Giudicelli et al. (1972), in a similar type of forest in Cameroon, observed a pronounced contagious distribution for Triplochiton scleroxylon and Terminalia superba but an approximately random distribution in the case of Entandrophragma cylindricum. They also found that the distribution pattern, when 10 important species were grouped together, tended to become regular. Studies of non-poisson distributions are complicated by the fact that the pattern, as revealed from the observed data, is confounded partly with the size and shape of the plot. According to Greig-Smith (1964), many types of nonrandom patterns appear random when sampled by very small or large plots, but are nonrandom with intermediate sizes. This shows the need for caution in drawing a general conclusion with patterns of species. The above consideration, together with ecological studies quoted earlier, clearly indicates, however, that within the range of plot sizes used many dominant species of the rain forest show a contagious distribution. An inventory person is not so much interested in the patterns as much as he is in their implications for the survey design. For pure randomly distributed populations (viz. Poisson) the planning job is rather simple. Here, precision of estimate depends only on the number of individuals counted. The standard error of estimate will be the same whether many small plots or few large ones or even a single big plot is used (Greig-Smith 1964). The shape of the plot has no effect. Similarly, survey precision will hardly be affected if plots are randomly or systematically arranged, or if they are clustered or not clustered. The situation is complicated for contagious distributions. One consequence has already been mentioned: lack of definite correlation between the plant density and coefficient of variation. Secondly, both the size and the shape of the sampling unit influence the precision of estimates. It is commonly observed that for the same sampling intensity small but well-dispersed plots give a higher precision than a few compact big plots (Dawkins 1952; Lanly and Vannière 1969; Greig-Smith 1964). Species differ in their response to the increase in the plot size. Their differential behavior can be correlated with their spatial arrangement. Triplochiton and Terminalia species, both of which have marked contagious distribution, show, relatively speaking, more reduction of error with the increase in plot size than Entandrophragma utile, whose spatial arrangement approximates Poisson distribution. Eusideroxylon zwageri and Khaya ivorensis, both of which also show aggregation tendency, behave in a similar way. 11.4 Micro-Variation Patterns and Their Significance for Plot Size and Shape 111

In the case of contagious distribution, elongated sampling units result in more precise estimates than compact units of the same size. This fact can be easily visu- alized: the former is more likely to fall in both dense and rare patches than the latter. Investigations by Jack (1961) show that long and narrow sampling units with general orientation across local drainage give more precise results than relatively compact ones. Similar experiences with ecological studies have also been reported by many investigators, e.g., Greig-Smith (1964, p. 29) and Aubréville (1971). So far attention has been confined to the study of contagious distributions only, because of their wide occurrence in the rain forests. There are many other possibilities in which plants can be arranged in space. In the light of cited studies it appears that an inventory in the rain forest at the level of individual species would be very expensive compared to that for ‘‘all species’’. As an alternative, it would be interesting to know what happens at the species group level. An expected effect of species grouping is a decrease in the coefficient of variation. Mean and coefficient of variation for various species and levels of species grouping for the mixed dipterocarp forests of Sarawak are presented in Table 11.4. Results clearly show that significant economy in sampling is achieved by working at the species groups level. At this level there is much less variability. The number of sampling units with zero stocking for species decline and the asym- metry in the frequency curve decreases. The effect of the size and the shape of the sampling unit on the coefficient of variation is still marked, as mentioned earlier, and continues to be so even at the ‘‘all species level’’.

11.5 Rain Forest Loss and Change

As has been mentioned earlier, FRA 1990 carried out tropical forest assessments using two techniques: (i) Model Assessment using country data available since 1960 and other GIS/RS databases; and (ii) Pan-tropical Remote Sensing Survey

Table 11.4 Mean and coefﬁcient of variation of basal area/ha at groups and species level for trees 18 in. (45 cm) and more in diameter (83 samples, unit 1) Groups/species Mean basal area Coefﬁcient of variation m2/ha (%) All species 14.70 27 Dipterocarps 9.89 32 Nondipterocarps 4.68 39 Red merantis 4.10 40 Yellow merantis 1.28 85 Keruings 1.16 82 Kapurs 1.21 141 Shorea parvifolia 2.46 111 Shorea rubra 0.95 142 Source Unasylva (1973) 112 11 Common Patterns of Spatial Variations in the Tropics using sample images close to 1980 and 1990. Using FORIS database, a time series analysis of rain forest losses since 1960 was carried out. Table 11.5 gives the historic area (at end 1960) and area at 2000 on regional and pan-tropical basis. The loss during 1960–2000 works out to 159 million ha, compared to 96 million ha prior to 1960, bringing out the fact that most of the rainforest loss occurred during the post-1960 period. Regional breakdown of loss during 1960–2000 are also given in Table 11.5. Table 11.6, based on remote sensing survey, presents estimates of rain forest changes during 1980–1990 grouped into two broad categories: (i) deforestation; and (ii) forest degradation. The former refers to outright change from forests to non-forest class; whereas the latter refers to changes within forest area changes. Two conclusions are obvious from the survey results: (i) Forest degradation rate (at 36.5 million ha) is higher by 50 % of the deforestation rate (-23.4 million ha); and (ii) the relative deforestation in Tropical Asia is very high (2.5 times of Brazil in proportion to the remaining rain forests). Forest degradation is higher than South America in absolute terms, viz., 16.9 million ha compared to 15.9 million ha respectively. The pan-tropical estimates of forest areas by model and remote sensing survey differ by 1.2 % only, whereas the regional values differ between 3 and 18 %. Deforestation rates, as expected, differ by a wider margin. The two methods used by FRA 1990 serve a complementary function. The pan-tropical remote sensing survey is based on a statistically planned inventory design to furnish estimates of state and changes in the forest area on a continuing basis. This is not possible by

Table 11.5 The historic and rain forest areas at end 1960 and 2000 Region The historic area Area Area Loss (Million ha) end 1960 end 2000 1960–2000 (Million ha) (Million ha) (Million ha) Africa 127 100 81 -19 Asia 308 245 159 -86 S. America 500 495 440 -55 Total 935 839 680 -159 Source Marzoli (2002)

Table 11.6 Rain forest changes by region (in million ha), 1980–1990 Region Remote sensing survey-based results Rain forest areas Deforestation Degradation 1980 1990 1980–1990 1980–1990 Africa 91 89 -2.0 3.76 Asia 161 151 -9.9 16.9 S. America 500 488 -11.5 15.9 All regions 752 728 -23.4 36.5 Source FAO 1996 p 151 11.5 Rain Forest Loss and Change 113 aggregating country statistics, even if their source is wall-to-wall remote sensing, because the aggregated error is high and ‘‘unknown’’. The contributing factors to error are: country to country variations in technology used, interpretation quality, timeliness of data, and differences in the classiﬁcation system. FORIS has some built-in procedures to minimize the effect of such variations by making time series analysis of country statistics at the sub-national level and using expert knowledge on country classiﬁcation and data collection method. The model also makes use of additional natural and cultural variables, correlated with forest changes, to reduce error in change estimation. But the fact remains that the precision of regional estimates based on country data can be improved only through improvements in the country assessments, which is possible only through country capacity building and inter-country harmonization of methods used. Information on forest degradation is a very important plus point in favor of the pan-tropical sample survey method, not presently possible in model-based assessments. The precision of sample survey techniques for estimating changes could further be improved by incorporating natural and cultural information at the sample sites in the analytical procedure, as done by the model. Further gains are possible by integrating wall-to-wall information obtained by using coarse resolution satellite data in the sample survey based on high resolution data.

Recommended Further Reading

Aubréville A (1971) Regeneration patterns in the closed forests of Ivory Coast. In: Eyre SR (ed) World vegetation types. London, Macmillan, p 41 Boon DA (1962) Plot size and variability. Delft, International Training Centre for Aerial Survey. ITC Publication, Series B. No. 17 Dawkins HC (1952) Experiments in low percentage enumerations of tropical high forest with special reference to Uganda. Emp For Rev 31:131 Dawkins HC (1958) The management of natural tropical high forest with special reference to Uganda. Oxford, Imperial Forestry Institute paper no. 34 FAO (1957) Report to the Government of Brazil on the forest inventory in the Amazon valley. FAO/ETAP report no. 601 FAO (1965) Report to the Government of Brazil on forest inventories in the Amazon. FAO/EPTA report no. 2080 FAO (1970) Inventory of the Ebini-ltaki area. Guyana. Forest Industries Development Survey. Technical report no. 9. FO-UNDP/SF GUY/9 FAO (1970) Preinvestment survey of forest resources. Technical report no. 1. FO-UNDP/SF IND/ 23 FAO (1973) The timber species of the mixed dipterocarp forest of Sarawak and their distribution. Malaysia. Working paper no. 21. FO: DP/MAL/72/009 FAO (1996) Forest resources assessment 1990: survey of tropical forest cover and study of change processes, Forestry paper 130 Giudicelli X, Lanly JP, Ouakam JB, Pietry M (1972) Application de la théorie des processus aléatoires a l’estimation de la précision d’un inventaire forestier par échantillonnage systématique. Annls Sci For 30:267 Greig-Smith P (1964) Quantitative plant ecology. Butterworths, London Heinsdijk D (1961) Forest survey in the Amazon valley. Unasylva 15:167 114 11 Common Patterns of Spatial Variations in the Tropics

Jack WH (1961) The spatial distribution of tree stems in a tropical high forest. Emp For Rev 40:234 Lanly JP, Vannière B (1969) Precision d’un inventaire forestier en fonction de certaines de ses caractéristiques. Bois Forêts Trop No 125:35 Nyyssönen A (1961) Survey methods of tropical forests. Rome, FAO Ramos AA, Vieira AN, Vivacqua CA, Alencar JC, Barros JCM, Péllico Netto S (1972) Inventario ﬂorestal do distrito agropecuario da zona Franca De Manus. Revta Flor 4(1):40–53 Richards PW (1952) The tropical rain forest. Cambridge University Press, Cambridge, p 10, 249 Richards PW (1971) The structure of tropical rain forest. In: Eyre SR (ed) World vegetation types. London, Macmillan, p 26 Schimper AFW (1903) Plant-geography upon a physiological basis. (trans: Fischer WR). In: Groom P, Balfour IB (eds) Clarendon Press, Oxford Singh KD (1973) Spatial variation patterns in the tropical rainforests, Unasylva, 26(106). FAO, Rome Singh KD, Marzoli A (1995) Deforestation trends in the tropics: a time series analysis. In: Proceedings of World Wildlife Fund/International Institute for tropical forestry conference: the potential impact of climate change on tropical forest ecosystem. Puerto Rico Chapter 12 Remote Sensing Applications in Forest Inventory

12.1 On Rapid Developments in Remote Sensing Technology

Technological developments in the ﬁeld of remote sensing have been phenomenal. Almost every year, new Earth Observation Systems (air-borne and satellites) are being announced with some novel features. The Google website mentions 68 such systems currently in orbit, which provide imagery of higher spatial, spectral, and temporal resolutions than earlier. A simple classiﬁcation of technologies relevant to forestry is: 1. Earth observation satellites with optical sensors (Landsat, SPOT, IRS, CBERS, etc.) 2. Multispectral, thermal, and hyperspectral sensors 3. Radar and lidar sensors 4. Aerial photographs (from air and space). Table 12.1 presents an overview of operational remote sensing technologies and their potential applications taking one important ecological formation, viz., mangroves, as an example. In 12 years’ time since the paper was published, each category of sensors has undergone several rounds of improvements. Still the paper provides a fairly good overview. Among the technologies mentioned, satellite images and aerial photographs are the most used products for forest inventory purposes. The satellite data has taken over aerial photographs on account of lower cost and easy availability. Both visual and/or digital image interpretation techniques can be used. Other geo-referenced data, like local administrative boundaries and topographic information, can be combined to make the product useful for planning.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 115 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_12, Springer-Verlag Berlin Heidelberg 2013 Table 12.1 Applications of different remote sensing technology illustrated with an example of mapping mangrove Inventory Forest in Applications Sensing Remote 12 116 Sensor, spectral Swath and Pixel size Criteria of Detected objects Applications resolution repetition rate mapping scale discrimination NOAA AVHRR Width 2700 km 1.1 km x 1.1 km Broad physiognomy Broad distribution Not much used (Sunderbans) Ch. 1: 0.58–0.68 l Daily 1/1000000 of large mangroves Ch. 2: 0.72–1.1 l

Landsat MSS 185 km x 185 km 60 m x 80 m Broad physiognomy One or two dominant Few applications in global Ch. 5: 0.61–0.69 l 16 days 1/250000 sub-classes of dense distribution map, regional Ch. 7: 0.8–1.1 l mangrove, openinter-tidal studies (Gulf of Guinea) ecosystems Landsat TM 185 km x 185 km 30 x 30 m Density, Phenology, Dense mangrove ,Open Cartography, Inventories, Ch. 3: 0.63–0.69 l 16 days 1/25 000 to 1/50000 Hydrology, mangroves,Mixed grassy and Management and conservation Ch. 4: 0.76–0.90 l Human impact woody halophytes,Barren dry planning saltish soils, Mudﬂats, Degraded stages of agriculture and aquaculture SPOT HRV 60 km x 60 km 20 x 20 m or 10 x 10 m, As above As above As above XS. 2: 0.61–0.68 l 26 to 3 days 1/25 000 to XS. 3: 0.76–0.89 l 1/50000

IRS-P6-LISS III 140 x 140 23.5 x 23.5 As above As above As above Ch. 3 : 0.62–0.68 5 days Ch. 4: 0.77–0.86

Aerial Photos Long intervals l/5000 to 1/25,000 As above + Height As above + tall dense In forest working plans (Infrared, colour or 5–20 years Classes + Structure Rhizophora stand [ 15 m Panchromatic) The information is based on Blasco et al. (1998) 12.1 On Rapid Developments in Remote Sensing Technology 117

Even within earth observation satellites, improvements are constantly taking place, almost forcing the pace of application. Table 12.2 gives an example of changes in the mapping system of Forest Survey of India during 1987–2005. The data procurement period for the whole country for LISS 3 data is now only 1 year. Currently, forest cover map is being produced in three classes with a minimum mapping unit of 1 ha using digital classiﬁcation techniques.

Table 12.2 Technological changes in forest cover mapping of India in the past Years Data Sensor Resolution Scale Mapping Method of period (m) unit (ha) interpretation 1983 1972–75 Landsat MSS 80.0 1:1000,000 400 Visual 1983 1980–82 Landsat MSS 80.0 1:1000,000 400 Visual 1987 1981–83 Landsat MSS 80.0 1:1000,000 400 Visual 1989 1985–87 Landsat TM 30.0 1:250,000 25 Visual 1995 1991–93 IRS-1B LISS-2 36.25 1:250,000 25 Visual & Digital 1999 1996–98 IRS1C/1D LISS-3 23.5 1:250,000 25 Visual & Digital 2001 2000 IRS1C/1D LISS-3 23.5 1:50,000 1 Digital (Raster) 2003 2002 IRS-1D LISS-3 23.5 1:50,000 1 Digital (Raster) 2005 2004 IRS-P6 LISS-3 23.5 1:50,000 1 Digital (Raster) 2009 2007 IRS-P6 LISS-3 23.5 1:50,000 1 Digitas (Raster)

There is, however, need for great care in introducing changes in the technology, interpretation method, classification, and mapping procedures at short intervals. It should only be done after taking care that the latest outputs are comparable with the past. The Symposium on Forest Monitoring organized by Global Earth Observation System of Systems (GEOSS) in September 2009 in Brazil sums up the current applications as follows: ‘‘Remote sensing has been demonstrated to be an effective tool for mapping and monitoring changes in the extent of the forest cover. When combined with ground measurements, these changes can be used to estimate carbon emissions and stocks. It is important to note that existing ground measurements are often limited, particularly in developing countries. In this context, the Symposium endorses the proposed GEO Task on Forest Carbon Tracking and recommends strong links to ongoing related Tasks on forest and land cover and on capacity building’’. The Symposium concluded that ‘‘several existing products clearly demonstrate the ability of earth observations to improve global forest monitoring. However, creating sustained operational systems based on these capabilities will require considerable additional efforts and major investments in capabilities and capacity. These improvements in forest monitoring will be more widely realized with the adoption of stronger open-data policies. Current remote sensing assets now support wild land fire management for tactical and strategic purposes. Global and national estimates of burned areas are needed, and there is a critical need for sensors with better spatial, temporal and radiometric resolutions than the current assets, which are not designed specifically to detect fires. Another area where operational earth 118 12 Remote Sensing Applications in Forest Inventory observation capabilities could be implemented is conservation and biodiversity assessment. The effectiveness of protected areas and progress towards the Con- vention on Biological Diversity (CBD) 2010 targets could be monitored with remote sensing combined with ground truth’’.

12.2 Lidar Potentials in Forest Inventory

The radar data complement optical data in areas with high cloud frequencies. Airborne laser scanning (ALS) also called LiDAR (Light Detection And Ranging) seems a promising technology, but still in an R&D state, which permits observation of the vertical structure of forests. This ability distinguishes ALS from conventional remote sensing approaches. LiDAR data seem to have a great advantage because laser pulses penetrate even through a dense multi-layered canopy. In a recently conducted study done in Laos, it has been claimed that two- phase sampling with LiDAR data provided estimates of aboveground biomass on an area of 0.5 ha with a relative root mean square error of 25–35 %. A recent example of data use in Laos will be described here. At the plot level (the area of a single plot was 400 sq. m), the satellite-based estimation of the entire area yielded root mean square error (RMSE) in measuring different parameters as follows: 19.2 % for dominant diameter, 18.5 % for dominant height, 26.5 % for basal area, and 34.4 % for stem biomass when using the surrogate plots as a training set. The estimation error drops dramatically when values are calculated at the scale of 0.4 and 1 hectare. For the 1-ha segments the values of observed RMSE were as follows: for dominant diameter 8.3 %, for dominant height 9.21 %, for basal area 19.5 %, and for stem biomass 23.9 %. It is stressed that the LiDAR is still in an R&D stage and brief mention has been made for the sake of com- pleteness of the information on technology (Table 12.3).

Table 12.3 Potential contributions of lidar remote sensing to forest inventory Forest characteristic Lidar derivation Canopy height Direct retrieval Subcanopy topography Direct retrieval Vertical distribution of intercepted surfaces Direct retrieval Aboveground biomass Modelled Basal area Modelled Canopy volume Modelled Mean stem diameter Modelled Large tree canopy density Inferred Canopy cover, LAI Fusion with other sensors Life form diversity Fusion with other sensors Source Ralph O. Dubayah and Jason B. Drake (2000) 12.2 Lidar Potentials in Forest Inventory 119

12.3 Applications of Aerial Photographs in Forest Inventory

This section will present a summary of research studies on applications of aerial photographs in forest inventory conducted by the Pre-Investment Survey of Forest resources Organization (PIS) during 1965–1985. The ﬁndings of studies are very relevant as remote sensing technology is moving toward stand level automatic interpretation. The earlier efforts in measuring height, crown diameter, and stand height using aerial photographs and correlating the values with ground measurements may be very interesting and useful. Survey designs used by PIS possess a good feature in that results can be complied on the basis of ground data alone, or after supplementing the ground data with information available from aerial photographs. This makes it possible to carry out some interesting comparisons among alternative survey designs. A major portion of the results quoted here are available only in the experimental ﬁles or cyclostyled reports of the Pre-investment Survey of Forest Resources (PIS) and the Indian Photo Interpretation Institute (IPI), now called Indian Institute of Remote Sensing (IIRS). The main aim of this chapter is to make these investigations known to a wider circle of research workers. The discussion on the subject has been divided into the following two sections.

Complete Photo Interpretation

The topics included are the following: • Use of serial photographs for stratification of survey area into forest and non- forest; • Use of aerial photographs for stratification of forest area into broad volume/ha classes; • Use of aerial photographs for species-types stratification; and • Stratification by crown density and height classes.

Point Photo Interpretation (Synonym: Plot Photo Interpretation)

The topics included are the following: • photo-interpretation for measuring: (i) height; and (ii) crown density; • studies on correlation of crown density and height with volume/ha and • development of aerial stand volume equations. 120 12 Remote Sensing Applications in Forest Inventory

12.3.1 Complete Photo Interpretation

(i) Broad stratification of the survey area into forest and other land-use classes In Table 12.4 survey results of four different projects of PIS are presented, which used aerial photographs for locating forested plots and stratification of the survey area into forest and non-forest classes. In fact, reliable identification of forested plots and their location in the field saves fieldwork costs significantly and has been a regular practice in all PIS surveys. In all the cases, the difference in the forest area estimate by the two alternative methods is rather small. This is true, more or less, irrespective of the scale of the aerial photographs used which range from 1:15,000 to 1:60,000. (ii) Use of aerial photographs for the stratification of forest area into broad volume/ha classes. In several projects of IPI (see Table 12.5), forest areas have been further stratified into three broad volume/ha classes to reduce the sampling error. The results of the various field surveys presented here look very consistent. It has been the experience that broad volume stratification can be done, with more or less the same degree of efficiency at photo scales varying from 1:20,000 to 1:60,000.

Table 12.4 Comparison of forest area estimates from ground and photo survey Survey Estimated forest No. of Difference Photo Name of survey project area area (%) ground 2 (%) scale (km ) plots Ground Photo survey interpret Bastar, MP (Tropical mixed 29,790 67.30 66.20 +1.1 2944 1:15,000 deciduous forest) Jhelum Valley, J&K. 10,196 25.89 25.82 +0.07 1274 1:40,000 (temperate forests) Chenab Valley, J&K. 9,846 36.48 38.00 -1.32 1217 1:43,000 (Temperate forest) Chanda, Maharashtra 13,351 69.03 67.60 +1.43 2684 1:20,000 (Trop. mixed deciduous) Source PIS report of various zones

Table 12.5 Mean volume (m3/ha) in various strata The survey areas East Godavari Bastar Chanda Timli Karimnagar (PIS) (PIS) (PIS) (IPI) (IPI) Strata (Mean volume (m3/ha) in the strata) 1 98.6 83.3 79.3 249.0 80.0 2 70.9 58.0 59.4 127.0 61.0 3 48.3 45.6 40.3 67.0 19.0 Source Inventory Reports of PIS and IPI for the survey areas 12.3 Applications of Aerial Photographs in Forest Inventory 121

In the East Godavari, Bastar, and Chanda surveys of PIS, the variance of mean volumes in the three volume strata had approximately the same magnitude, which indicates that stratified sampling, with proportional allocation of ground plots to various strata, was a good method of allocating the ground plots. Only in the case of Timli (UP) the three volume classes also had large differences in the variances (variances for the strata 1, 2, and 3 were 60, 39, and 13 respectively). This shows the possibility of further gains by distributing the ground plots according to Neyman allocation (optimum allocation) (iii) Use of aerial photographs for species-type stratification. In PIS and IPI, a number of studies have been made on further classification of the forest area by species types. The results have been very varied. In the case of the coniferous forests of J&K the differences between the estimate of the fir and spruce areas by ground survey and photo interpretation (scale 1:40,000) were very close, as given in Table 12.6. The problem of species-type recognition is not only in photo interpretation but also in the field, as the following study shows (Table 12.7). Form the above statistics, one can see that even on the ground classification of forest types of the same plot differs from crew to crew. In most cases the difference in classification is small, e,g., Sal with miscellaneous or miscellaneous with Sal.

Table 12.6 Species-type recognition by photo interpretation Species type Estimation of area (km2) Difference (%) Ground survey Photo interpretation Fir/Spruce 1352 1336 1.2 Bluepine/Deodar 1016 1025 0.9 Source PIS Inventory Report of J&K 1971

Table 12.7 Replicated observations of forest type of the same plot by different observers using the ground and photo survey methods, in the tropical mixed deciduous forests of Bastar, M.P., India (Wahlstrom 1967) Plot No. Observer’s serial number 1234567 Ground survey Photo interpretation Forest-type code recorded by various observers 1 0004554 2 2225223 3 0000004 4 5254555 5 5252550 6 5640000 7 5404445 8 4445444 9 5254444 Forest-type codes: 0 Sal, 1 Teak, 2 Miscellaneous medium and good, 3 Miscellaneous poor in outcrop area, 4 Sal with miscellaneous, 5 Miscellaneous with Sal, 6 Teak with miscellaneous 122 12 Remote Sensing Applications in Forest Inventory

Teak classification in plot No. 6 is presumably a mistake in filling up the form. In plot No. 7, on the other hand, we can see that the opinions differ from pure Sal to miscellaneous with Sal. In photo classification the variation is more frequent and the amplitude wider. Note also that all the interpreters have classified plot No 6 as pure Sal, where the corresponding field classification is miscellaneous with Sal or Sal with miscellaneous. These differences highlight the problem associated with definition of species type from person to person. It shows the need for adequate care in evolving suitable definition and classification. (iv) Stratification by crown density and height classes In several surveys of PIS/IPI, where photographs on a scale larger than 1:25,000 were available, efforts were made to stratify the forest area by species, height, and density classes. The height was usually divided at intervals of 2–5 metres, and the density into 3–5 classes. After delineation of species, height, and density strata, a number of ground plots were laid in the field for the estimation of the mean volume. Subsequent analysis has, however, shown that many of such strata do not differ significantly with respect to the mean volume. It has been generally found that for the purpose of mean volume estimation the number of strata could be reduced to 3 (Parthasarathy 1975). To investigate the optimum class interval for height and crown density measurement, the data of 672 ground plots of Bastar forests (Central India) were tabulated as given in Table 12.8. In the above table there are only three crown density classes with the levels of volume/ha at 47, 77, and 95 m3 respectively. It is also observed that two-way classification of height and crown density classes does not, in general, lead to good stratification of volume/ha. The differences between the density classes 0.4–0.6 and 0.7–1.0 are especially very small. However, the density class 0.1–0.3 shows out as a distinct stratum in all the height classes. The above table also shows height as a better variable for stratification of the mean volume than crown density. Height strata have a range 10–200 m3/ha. While developing a classification system for height, one has to keep in view the two kinds of errors involved, viz., sampling and non-sampling. Further, in large- scale inventories, in which PIS has been mostly engaged, a third component, namely variability among the interpreters, is also to be kept in view. Studies conducted by PIS show significant variation among the Interpreters in the measurement of height (Job Nos., 14 and 19 of PIS, Central Zone, Bastar).

Table 12.8 Volume m3/ha by height and density classes (based on ground data) Density/height 10 12 14 16 18 20 20+ All 0.1–0.3 6 26 44 70 77 87 100 47 0.4–0.6 10 22 46 68 110 159 265 77 0.7–1.0 22 26 58 77 113 160 218 95 Source Report on trial results Jagdalpur, U.E. Wahlstrom 1967 12.3 Applications of Aerial Photographs in Forest Inventory 123

12.3.2 Point Photo Interpretation

The main aim of point photo interpretation is to estimate ground volume/ha of a photo point on the basis of photo measurement like height and crown density (sometimes other parameters also for example, crown size). The following steps are involved in the process; (i) height measurement or estimation; (ii) crown density estimation; (iii) correlation of height and crown density with volume; and (iv) construction of aerial stand volume equation (s). (i) Height measurement or estimation Results of an investigation conducted by PIS in the tropical mixed forests of Bastar, Central India, are given below for illustration (Table 12.9). The above comparison shows mean and standard deviation of height measured on 294 plots on the ground and on the photo. The mean and standard deviation values in both the cases are very comparable. However, the interesting item is the standard deviation of the difference (ground height—photo height); this is of the order of 5.4 m, which is of the same magnitude as the standard deviation of the height in the population. If the photo-height and the ground-height measurements were identical, the standard deviation of the difference would have a low value and should be equal to zero in the ideal cases. The present magnitude, which seems rather high could be attributed to two factors; firstly, random differences in the location of trees on ground and photo plots and, secondly, the difference among the interpreters. While developing the serial stand volume equations it is assumed that the independent variables (in this case height and crown density) are free from error of measurement. This does not appear to be justified on the basis of the present investigations. (ii) Estimation of crown density The questions associated with the crown density measurement have already been discussed earlier. It was especially indicated that more than three crown density classes may not lead to significant improvement in the precision of mean volume estimator. The Crown density estimations are also subject to non-sampling

Table 12.9 Comparison of ground and photo-height measurements in the tropical mixed forests of Bastar, M.P Technique No. of observations Mean (m) Standard deviation ±2 SE (m) Ground height (GM) 294 16.2 4.1 0.47 Photo measured (PM) 294 16.3 5.2 0.59 Photo estimated (PE) 294 15.4 5.5 0.62 Difference (GM-PM) 294 -0.1 5.4 0.62 Difference (GM-PE) 294 -0.7 4.6 0.53 Difference (PM-PE) 294 0.8 5.9 0.68 Source PIS, Central Zone, 1966–68 124 12 Remote Sensing Applications in Forest Inventory

Table 12.10 Crown density classiﬁcation on the ground and photo as recorded by different observers (10 = Full Density; O = Blank) Plot Serial number of the observer Volume Forest-type code m3/ha 1234567 Ground survey Photo interpretation Crown density codes of the same plot by different observers 1555444690Sal 2 6 6 5 4 4 6 8 150 Miscellaneous 3 7 9 7 8 8 10 10 250 Sal 4 5 5 3 2 2 4 6 40 Misc. with Sal 5 6 7 7 6 6 8 8 140 -do- 6 7 8 6 6 8 8 10 110 -do- 7 5 6 5 2 2 4 6 100 Sal with Misc. 8 7 7 6 4 6 6 4 230 -do- 9 6 7 5 6 6 6 10 90 Misc. with Sal Source PIS study in Bastar, Central India errors. To illustrate the point, the experimental results of the Bastar survey are given in Table 12.10. In this study, there is appreciable variation in the crown density classiﬁcation of the same plot from observer to observer. This is an indication of the poor correlation between volume and crown density. Plot No. 1 has lower density than plot No. 9, but the same volume. Similarly plot No. 6 has higher density, but lower volume than plot No.5. (iii) Correlation of height and crown density with mean volume For examining this subject, values of multiple correlation (R2) of height and crown density with volume, as observed in the North Zone forests of PIS, are given in Table 12.11. This has a range from 0.36 to 0.72. R2 is the lowest in the case of mixed hardwoods at the low altitude, viz., 0.36, and has maximum value of 0.72 for the mixed hardwood forests at the high altitude. The coniferous forests have a

Table 12.11 Correlation between ground height and ground crown density with mean volume of the same plot in sub-tropical/temperate forests of India Forest type Multiple correlation coefficient (R2) No. of observations Top Crown Height and height density density Mixed hardwood at low- 0.20 0.10 0.36 46 altitude Chir 0.46 0.18 0.56 48 Bluepine 0.53 0.66 0.72 46 Deodar 0.46 0.46 0.72 48 Fir 0.38 0.24 0.49 45 Mixed hardwoods at high- 0.69 0.25 0.72 74 altitude PIS—Ground sampling data of the North Zone 12.3 Applications of Aerial Photographs in Forest Inventory 125 range between 0.60 and 0.73. In the case of photo measurements, due to various types of errors, the value of R2 is expected to be lower than indicated above. Some of the common sources of error are listed below: • Quality and scale of aerial photographs. • Differences in the concept of forest type on the photo survey and ground survey. • Differences in the size and shape of plots used for the ground and photo. • Differences in the location of ground and photo plots. • Differences among the Interpretations. (iv) Development of aerial stand volume equations Some aerial stand volume equations developed in India are listed in Table 12.12. The above results show multiple correlation coefficient (R2) between volume on the ground as dependent, and photo height and crowd density as independent variables. The range is between 0.15 and 0.62. The low value of correlation coefficient implies a big restriction on the use of photo interpretation as part of a double sampling design, unless the ratio of the cost of fieldwork as compared to photo interpretation is very high (see the following section). The effect of scale of aerial photograph on the value of R2 in the case of tropical mixed forests of Bastar, MP, was Investigated in detail by Joshi (1972). His findings are listed in Table 12.13.

Table 12.12 Volume equations reported by various investigators in India Name of Forest type Photo scale Signiﬁcant No. of ground Correlation investigator variables plots coefﬁcient PIS (1969) Sal 1:15,000 H, C 19 0.47 Good misc. 00 H, C 242 0.14 Poor misc. 00 H, C 44 0.25 Teak 00 H, C 27 0.19 All species 00 H, C 332 0.20 Joshi (1972) Misc. 1:20,000 H, C – 055 Gupta (1973) Teak 1:10,000 H 51 0.46 Tiwari (1978) Sal 1:10,000 H, C 51 0.50 Notation: H = height; C = crown density

Table 12.13 Correlation coefﬁcient between photo and ground variables Scale Photo variables R2 1: 5 000 H, C 0.62 1: 7 500 H, C 0.57 1:10 000 H, C 0.46 1:20 000 H, C 0.55 1:25 000 H, C 0.35 Source Joshi (1972) 126 12 Remote Sensing Applications in Forest Inventory

The above results do not reﬂect an appreciable difference in R2 for a scale between 1:7 500 and 1:20 000; the scale 1:25 000 does appear to have signiﬁcantly lower values of R2 as compared to 1:20 000.

12.4 Estimating Cost-Effectiveness of Remote Sensing

The efﬁciency of one survey design over another E1, 2 could be expressed by the ratio: (cf. Cochran 1963)

E 1;2 ¼ C2V2= C1V1 where,

C1 = cost of measuring one sampling unit for the survey design (1); C2 = cost of measuring one sampling unit for the survey design (2); V1 = variance of mean associated with the survey design (1); and V2 = variance of mean associated with the survey design (2); Example: To illustrate the gain of precisian, the following results are reproduced from the PIS report of East Godavari, India. In the case of the East Godavari survey, the cost of fieldwork covering an area of about 10 000 km2 came to 61.00 Indian Rupees/km2. The cost of complete photo interpretation at a photo scale of 1:60,000 showing three broad volume/ha classes, came to about 12.50/km2. The cost for the combined survey is, therefore, Indian Rupees 73.50 Indian Rupees/km2 (= 61.00 ? 12.50). The ratio of variances: V (combined survey)/V(field survey) was 1/3 (see Table 12.14). The efficiency of ground survey as compared to combined survey is, therefore, 73:50 1 1 E ¼ ¼ 1;2 61:00 3 2:5 The above calculation thus shows that, in East Godavari, photo interpretation made a substantial contribution to the precision of the survey. This comparison holds true only if the main concern of the survey is total volume (of all species). The gain of precision at the level of total forest area and total volume would also, however, improve the precision at the species level. Therefore, the ratio is representative of the whole survey.

Table 12.14 Comparison of ground with combined survey standard errors SE% for survey options Stratum Variance ratio Ground Combined High volume 7.5 3.9 0.27 Medium volume 9.8 5.1 0.26 Low volume 12.6 7.1 0.31 Shifting cultivation 30.2 24.4 0.65 Source PIS Data Processing Report for East Godavari 1975 12.4 Estimating Cost-Effectiveness of Remote Sensing 127

In the Bastar survey of PIS, 27,000 photo plots (0.4 ha area) were laid 1km9 1 km apart on photograph at the scale of 1:15 000. For all the plots, top height was measured in 2 m classes and crown density estimated in 20 % classes. Then 376 photo plots were located on the ground and their volume/ha were accurately measured. Based on this basic data, aerial stand volume equations were developed. The values of R2 for all the equations have already been referred to in Table 12.12. As can be seen, the overall value of R2 is 0.25. For studying the cost-effectiveness of the survey, a formula given by Cochran (1963) was used. This double sampling approach would be economical if: Rffiffiffiffiffiffiffiffi2 hCg=Cpi [ p 2 ðÞ1 1R2 where Cg is the cost of measuring one ground sample, Cp the cost of measuring one photo sample, and R2 is the value of the correlation coefficient between the photo variables and ground volume. The ratio Cg/Cp has been listed below for various values of R2 ranging from 0.2 to 0.8 (see Chap. 6).

R2 ¼ 0:20:30:40:50:60:70:8

Cg = Cp ¼ 18:011:08:06:54:53:52:5

2 For an overall value of R at 0.25, the value of the ratio Cg/Cp has to be more than 18 for the double sampling to be cost-effective. In the Bastar survey the sampling unit on the ground consisted of a cluster of three plots, each 0.1 ha and 500 m apart from each other. They were systematically distributed at intervals of 9 9 9 km. The cost of fieldwork came approximately to 600.00 Indian Rupees/ sampling unit (1969 estimates). The cost per plot was, therefore, INR 200.00.The cost of aerial photography at scale 1:15,000 in the same year was reported to be 15.50 Indian Rupees/km2 and the cost of photo interpretation 1.80 INR/photo plot. The total cost/photo plot was, therefore INR 17.30. The ratio between field and photo plot survey is accordingly: 1:12. Based on the above statistics, double sampling does not seem to be a cost-effective design. Nichols et al. (1975) investigated the use of remote sensing in combination with field survey (double sampling) for volume prediction in the USA Plumas National 2 Forests. The observed value of R was 0.6 and Cg/Cp was 18/1. Consequently, the use of double sampling was appropriate.

12.5 A Case Study of FSI State of Forest Report

12.5.1 Assessment Method

The methodology used by the Forest Survey of India in the State of Forest Report released in 2009, compared to last SFR, has many new features: vector-based approach has replaced that based on raster with focus on the polygon where change 128 12 Remote Sensing Applications in Forest Inventory in forest cover has occurred, either due to improvement or degradation. Due to this change in the methodology, the subjectivity in the interpretation has been greatly reduced. The vector output provides improved cartographic presentation and facilitates GIS analysis. Operational steps involved are the following: First, geometric rectification of the satellite imageries is carried out with the help of scanned and geo-referenced Survey of India (SOI) topographic sheets on 1:50,000 scale. The methodology of interpretation involves a hybrid approach in which unsupervised classification (ISODATA algorithm) is aided by on-screen visual interpretation of forest cover. In order to make the present maps comparable with the earlier one (viz. the year 2005), the entire dataset of the earlier assessment was reinterpreted using identical methodology. Compatibility of geometric rectification was ensured with the help of a new module of the ERDAS Imagine software called Auto Sync. Normalized Difference Vegetation Index (NDVI) transformation was used for removing non- vegetated areas from the scene. Interpretation is followed by extensive ground verification. All the necessary corrections are subsequently incorporated. Reference data collected through ground truth and field experience of the interpreter plays an important role in classifying and delineating the forest cover into the following canopy density classes. • Very Dense Forest: lands with tree cover of canopy density of 70 % and above. • Moderately Dense Forest: lands with tree cover of canopy density between 40 and 70 %. • Open Forest: lands with tree cover of canopy density between 10 and 40 %. • Scrub: Degraded forest lands with canopy density less than 10 %. • Non-forest: Any area not included in the above classes.

12.5.2 Accuracy Assessment

For accuracy assessment, 4302 control points were randomly chosen distributed over the whole country for validating the map classification through field visit as well as reference to very high resolution satellite data of 5.8 m resolution. The two datasets (viz. original classification and ground truth) are tabulated in the form of a matrix (see Table 12.15). From this table, a number of statements about the reliability of the map output can be calculated. The first is overall accuracy. This is the ratio: (sum of all diagonal elements of the matrix, viz., correctly classified points 3,962)/(total number of control points, viz., 4,302). This comes to 92.10 %. If the mapping is considered in two classes only, i.e., forest cover and non-forest, the overall accuracy rises to 95.72 %. Producers’ accuracy tells us the probability (in percentage terms) of field points/ plots of a given vegetation or land-use type (such as forest, agriculture land, grass land, etc.) getting correctly mapped. This is given for each class in the last row of the matrix calculated by dividing the number of correctly classified points by the total number of field points/plots of that type. The users’ accuracy is given in the 12.5 A Case Study of FSI State of Forest Report 129

Table 12.15 Error matrix for interpreted classes in FSI maps (FSI 2005) Reference data classes (ground truth) Interpreted classes User’s accuracy Very Dense Open Scrub Non-forest Total (%) dense Very dense 197 4 0 0 1 202 97.52 Medium dense 7 1,350 63 10 22 1,452 92.98 Open forest 3 42 925 9 34 1,013 91.31 Scrub 0 7 3 168 9 187 89.84 Non-forest 3 40 55 28 1,322 1,388 91.30 Total land 210 1,443 1,044 215 1,388 4,302 100.00

Producer’s 93.81 93.56 88.43 78.14 95.24 accuracy (%) last column of the matrix, calculated by dividing the number of correctly classified map points/plots for a given class by the total number of map points/plots belonging to that class. The two types of accuracy statements, viz., that of producer and of the user serve different purposes (Table 12.15). There is another statistics k (called khat), which serves as an indicator of the extent to which the overall (%) accuracy of an error matrix is due to ‘‘true’’ agreement versus ‘‘chance’’ agreement. This is on account of the fact that even a completely random assignment of pixels to classes will produce correct percentage values in the error matrix. In fact, such a random assignment could result in a surprisingly good apparent classification result. The k statistic is a measure of the difference between the actual agreement between reference data and an automated classifier and the chance agreement between the reference data and random classifier. In contrast to overall accuracy, k takes into account the non-diagonal elements of the error matrix. For further details Lillesand and Kiefer (2004) may be referred to. In the case of FSI map, the observed value of k is 0.89, showing that map classification is very robust.

Recommended Further Reading

Blasco F, Gauquelin T, Rasolofoharinoro M, Denis J, Aizpuru M, Caldairou V (1998) Recent advances in mangrove studies using remote sensing data. Mar Freshw Res 49:287–296 Cochran WG (1963) Sampling Techniques. Wiley, New York Dubayah RO, Drake JB (2000) Lidar remote sensing for forestry applications. J For 98(6):44–46 Gupta PN (1973) In: Proceedings of symposium IUFRO S 6.05, Freiburg, Federal Republic of Germany FSI (2005) State of forest report. Forest Survey of India, Dehradun FSI (2011) State of forest report. Forest Survey of India, Dehradun Joshi SC (1972) Optimum scales of black and white aerial photography for growing stock assessment in Bastar Region, India Lillesand TM, Kiefer RW, Chipman JW (2004) Remote sensing and image interpretation, Wiley (A newer edition is available) 130 12 Remote Sensing Applications in Forest Inventory

Parthasarathy V (1975) A note on resources survey of Karimnagar East Forest Division IPI PIS (1971) Pre-investment survey of forest resources Chenab Valley (J&K) PIS (1969) Pre-investment survey of forest resources of Bastar District, Central Zone (Job Nos., 14 and 19) PIS (1975) Data Processing report on forest resource of East Godavari (AP) PISFR (1971) Kashmir valley, J&K. Inventory results. Pre-investment survey of forest resources PISFR (1975) East Godavari (AP). Pre-investment survey of forest resources PISFR (1975–1976) Chenab Valley (J&K). Report on forest resources, vol 1. Pre-investment survey of forest resources PISFR (1976) Ballarashah Catchment. Pre-investment survey of forest resources, Chanda District (Maharashtra) Nicols J, Sharp J, Titus S, Orme W, Gialdini M (1975) Cost-effectiveness comparison of existing and Landsat based total timber resources inventory system, a research paper, University of Berkeley, California Tiwari KP (1978) Aerial stand volume equations for Sal forests of Timli, UP Wahlstrom UE (1967) Report on trial ﬁeld results. Pre-investment survey of forest resources, Jagdalpur Chapter 13 Growth and Yield Studies

13.1 Special Growth and Yield Conditions in the Tropics

We will ﬁrst present a conceptual model, commonly used in genetics, for describing biological development of an individual or a population, viz.,

P ¼ G þ E ð13:1Þ where P is the observed value of a characteristic of an individual or a population, also called phenotypic value; G is the share of heredity in its realization, also called genotypic value; and E the environmental determination. This model is applicable to all crops, annual and perennial, woody or herbaceous. After adding time as a factor, the model could be expressed in the following form:

P ¼ fGðÞ; E; t ð13:2Þ A distinguishing feature of forest crop, compared to agriculture, is the long period involved in its development. Decisions on ﬁnal cutting and intermediate operations, such as tending and thinning, are needed to ensure that the forest produces desired outputs. These important decisions are generally made comparing ‘‘the actual forest’’ with ‘‘model forest’’, the latter are constructed as a part of long-term growth and yield research, of which reliable and representative ﬁeld observations form an essential part. In the tropics, there is a further complexity: almost 90 % of forests are multi- species, uneven-aged, and multi-storied (see Table 13.1). The reported number of tree species (30 cm+) in a rain forest plot of 50 ha in Paso, Malaysia, was 357. The other formations have lesser but still high species richness. For example, Barro Colorado Islands, Panama, had 229 species; and Mudumalai, India, had 52 species for the same plot size. The forest canopy structure is also more complex than the temperate zone. A further problem is that the majority of tree species lack annual rings. It is not possible to estimate the age of species and, therefore, calculate the

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 131 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_13, Springer-Verlag Berlin Heidelberg 2013 132 13 Growth and Yield Studies

Table 13.1 Canopy structure and species richness by forest formations in the tropics Location Forest type Dry months Canopy layers Stems Number of species per ha in 50 ha Plot Number 10 cm+ 10 cm+ 30 cm+ Pasoh, Malaysia Rain forest 0 5 529 683 377 BCI, Panama Moist forest 4 3 429 229 146 Mudumalai, India Dry forests 6 1 301 63 52 Source Studies on 50 ha plot data. See Condit et al. 1996 mean annual increment. The complexity requires long-term growth and yield research projects, institutional capabilities, and ﬁnancial resources. Taking into account the complex factors, the above model (13.2) can be reformulated as follows:

P ¼ fGðÞ; A; B; t ð13:3Þ where G is the genotype under study, A is the abiotic, and B the biotic environment. Abiotic environment, mainly including climate and soil factors in a study site, can be taken as more or less constant. In contrast, the biotic environment of each tree in the mixed tropical forests is extremely variable, in vertical as well as horizontal planes. This makes growth of a ‘‘genotype’’ very variable depending on the neighbor. Keeping the above considerations in view, discussions in this chapter have been organized in the following order in the increasing order of variability: • Pure even-aged forests • Mixed uneven-aged forests • Individual tree level studies Pure even-aged forests are taken up ﬁrst, because they are the simplest formation in the tropics and provide insight into the methods of forest growth and yield research in general. One can also draw upon the vast knowledge and experience available for forests in the temperate regions, where most of the forests are even-aged by nature. With the increasing importance of Carbon Trading, the demand for reliable growth and yield data on tropical plantations is also expected to increase rapidly.

13.2 Growth and Yield of Tropical Plantations

13.2.1 Methods of Study

Studies generally make use of the following sources of data, viz., (i) Permanent Sample Plots (PSP), which should ideally cover the entire range of species distribution, site quality, and treatments classes; and (ii) Forest Inventory Plots (FIP), 13.2 Growth and Yield of Tropical Plantations 133 ideally re-measured plots, but quite often temporary plots have been used. The output is generally in the form of classical yield tables or mathematical functions. FAO Forestry Papers on Methods of Volume Estimation and Yield Prediction provides information on sample plot design, procedure for data collection, data storage, and analysis and validation of G&Y models, which may be referred to for getting detailed guidance on the subject (FAO 1980a, b). The following aspects of the PSP measurements will be brieﬂy mentioned here: • Size of plots and their shape • Number of plots and their distribution • Re-measurement policy The size of sample plots, depending on the variability in the plantations, may range from 1/40 to 1/10 ha. Generally circular or square plots are used. Higher the variability in the population, larger the plot size and sampling intensity may be required. As a rough guide 100–200 plots, uniformly distributed over all age, site quality, and stocking classes, should be measured to get a good growth equation. It is desirable to have, at least, part of the sample plots (say 20–30 %) measured over a long period, say minimum three times, to obtain an accurate estimation of the shape of the growth curve and, therefore, family of site quality curves. Timely editing of sample plot data and comprehensive analysis using a good statistical package is highly recommended. Advantage should be taken of research reports published by the concerned IUFRO Working Party and papers in international journals.

13.2.2 The Current State of Knowledge

The yield research began in India about 100 years back, associated with large-scale conversion of natural forests into uniform even-aged managed forests including: Sal (Shorea robusta), Teak (Tectona grandis), Chir-pine (Pinus roxburghii), and Deodar (Cedrus deodara). By 1930, provisional yield tables for all main species were already available. As the area under indigenous and exotic plantations increased, more sample plots were laid and tables/equations were updated. Pandey (1983), in a desk study under FAO auspices, covered 76 countries and 8 million ha of tropical plantation areas. He found 120 publications on 29 main species. The quality of yield tables for seven species was judged ‘‘good’’, 14 species ‘‘average’’, and eight species ‘‘poor’’. The same author, based on Forestry Abstracts reports 12 years later, found very few new publications although the planted area in the tropics during the period 1980 and 1990 had almost doubled. The report concluded that ‘‘the number of existing reliable growth and yield models against their requirement are so few that to estimate the total potential yield of the entire plantation area, even at the national level and more so at global level, is impossible’’. To improve the quality of yield table studies, the following recommendations were made: 134 13 Growth and Yield Studies

• The investigators must include in their report description of site factors of their study area such as climate, topography, etc. • Provide basic information about the crop/seed origin and treatment, fertilized or unfertilized, irrigated or un-irrigated, etc. • Inform about the nature of sample plots, viz., permanent or temporary or a part of plantation survey, their size, number, age series, and numbers of measurements, etc. • The concept of volume used must be deﬁned, whether under bark/over bark, inclusive or exclusive of stump and branch wood, diameter measurement limit, and method of volume measurement. • Site quality assessment method needs to be provided, whether based on dominant or average height and what reference age. • Information on stand densities per hectare at different ages if thinning or natural removal taken place must also be quoted. • While quoting yield data, information about the thinning yield must be made. • Sample plots should be established with equal representation of different age classes. • Use of IUFRO standard symbols would make reports more understandable and comparable among countries. • Long-term international cooperation in growth and yield studies among the tropical countries (S–S Cooperation) and between Tropical and Temperate Region Countries (N–S Cooperation) will be very useful.

13.3 Growth and Yield Research in the Temperate Zone

13.3.1 The Current Status

Pretzsch (2009) summarizes the European situation in the following words: ‘‘With a history of more than 200 years, yield tables for pure stands may be considered the oldest models in forestry science and forest management. They reflect stand growth over defined rotation periods and are based on long-term periodic measurements of diameter, height, biomass, etc. Despite a number of drawbacks, yield tables still form the backbone of sustainable forest management planning. ……..’’ Beginning with yield tables for large regions as a basis for taxation and planning, model development led on to regional and site-specific yield tables and culminated in the construction of growth simulators for the evaluation of stand development under different management schemes, mainly on account of availability of data over a long period of time, continuity of research, and accumulated knowledge. An attempt was made in the year 2000 to link the single country runs to one dynamic full-scale European model. The evolution of European forest management illustrates the important role played by knowledge and information emanating from long-term observation plots, reliable inventories of forest resources, and research on silviculture. The second very important lesson is growing cross-country cooperation 13.3 Growth and Yield Research in the Temperate Zone 135 in research. The tropical countries should soon initiate measures for sharing G&Y data and knowledge among them, which may bear fruits say 10–20 years hence. Stage (1976) and Hägglund (1979) observed systematic differences between growth curves developed from the data of yield research and forest inventory sample plots. This could be due to special treatment (e.g. location, measurement, thinning etc.), often accorded to the research plot. One way to minimize bias is to monument the sample plots in a sufficiently inconspicuous manner, so that a normal worker is not made aware of them while carrying out forestry operations. This, however, increases the risk of error in relocation of the sample plots at the time of re-measurement. In a personal discussion, Nyyssönen (1981) drew attention toward systematic ‘‘bias caused’’ by use of small-sized sample plots. N. E. Nilsson (1980), in a personal communication, provided the following comments on PSP studies, in relation to their application in forest management planning: (i) Research plots almost always represent growing conditions, which are much too ideal in comparison with what can be found under practical conditions. (ii) Temporary sample plots can be useful, if their age is known, but in some cases they can lead to bad results. For instance, in case of a rapid development over age with respect stand establishment procedures or similar causes, which may invalidate past development functions based on one-time measurements. Forest inventory techniques, based on one-time plots, have the same disadvantage. (iii) It may be possible to develop models based on observations from certain homogeneous stands, but the results do not refer to any other stand condition. Number of plots and selection procedure should represent actual field conditions. There is right argument against yield research based on stands, which are not representative of the real forest conditions. Research findings should include the description of the source data of plots in annex to judge the applicability of research findings. (iv) Young stands may perform much better than say 10–15 year older stands. Climatic fluctuations may influence on the growth during the entire life cycle of a short rotation crop. Every year of plantations need to be followed up individually.

13.3.2 Forest Plantation’s Development Modeling

Relative Production Function (RPF) was developed by Nils-Erik Nilsson, Sweden, for making potential cut calculations for Swedish forest species in the framework of the Swedish Forestry Commission, 1973. It uses Chapman-Richard Function with Age and Yield expressed in relative terms as (%) of Rotation Age and Rotation Yield; a, b, and c are parameters (see Table 13.2 for terminology). Nilsson (1996) presents a detailed account of the Function. 136 13 Growth and Yield Studies

Table 13.2 Symbols and parameters of the relative production function Item Description Age: A Speciﬁed by 10-year intervals at mid class 5, 10, 15 etc. Volume:V (m3/ha) The volume standing in the forest excluding thinning Accumulated yield: W (m3/ha) Total volume over bark including past thinnings and standing volume Mean annual increment: The accumulated yield (m3/ha) divided by the age MAI (m3/ha/year) Current annual increment Difference in the accumulated yield between two ages CAI (m3/ha/year) divided by the time interval between two measurements Max MAI (m3/ha/year) The maximum value of mean annual increment Rotation age: T (Years) The age for maximal mean annual increment Rotation yield: P The Accumulated Yield at the Rotation Age Relative age: R Age expressed as % of the Age of maximum MAI (viz. rotation age) Relative yield: Y Accumulated yield in an age class as % of the total accumulated yield at the rotation age Thinning fraction: G Accumulated thinning yield as a fraction of the total accumulated yield at the rotation age Final cut volume: S The standing volume left for ﬁnal cutting at the rotation age as % of the accumulated yield

The form of the RPF and its ﬁrst differential follow: ÀÁc Y ¼ a 1 bR=100 ÀÁc1 dYðÞ=dRðÞ¼0:01 a c lnðÞb bR=100 1 bR=100

The Area Production Model (APM), developed for long-term forestry and land- use planning in India in 1976, made use of RPF to describe the growth functions of most of the planted species, both coniferous and broad leaved. The model forecasts were very close to observed values (see Table 13.3 and Fig. 13.1). It may further be noted that the relative production function assumes that the thinning does not inﬂuence the total yield. This would hold true only if the thinning is not too heavy as compared with the growing stock at the time for thinning. Accordingly, Y = G ? S. This fact was also validated using actual sample plot data for many species. Should the stand volume in real forest deviate from the model volume, to estimate actual increment, model increment has to be adjusted. 13.3 Growth and Yield Research in the Temperate Zone 137

Table 13.3 Relative production functions of temperate and tropical zone species Species Site quality MAI Rotation age Production at relative rot. ages Y(50 %) Y(80 %) Y (100 %) 1. Shorea robusta, India I 4.5 100 38.6 77.5 100.0 2. Cedrus deodara, India I 5.6 85 38.2 77.2 100.0 3. Eucalyptus hybrid, India I 15.5 8 38.7 79.2 100.0 4. Picea excelsa, Sweden G20:4 5.5 150 38.5 78.0 100.0 5. P. excelsa, Sweden G28:4 9.2 105 41.6 79.1 100.0 6. Pinus sylvestris, Norway E/III 3.2 115 39.2 77.5 100.0 7. P. sylvestris, Norway D/III 4.5 100 39.3 77.1 100.0 9. P. contorta, Sweden H50 = 20 6.4 68 37.5 78.4 100.0 Relative Production All Sites 1.0 100 38.0 77.7 100.0

Fig. 13.1 Relative production model

Increment and Volume Relation

The following empirical relation has been found useful while working with the model

a If ¼ Im ðÞVf=Vm

Vm is volume of the model; Vf is volume in the stand; Im is Increment in the model; If is increment of the Stand; and a : is a constant. if a is 1.0, there is direct relationship; if a is 0.5, there is square root relationship. Generally a has been found to vary between 0.7 and 0.8. 138 13 Growth and Yield Studies

13.4 Growth and Yield of Mixed Tropical Forests

The G&Y research on mixed tropical forests had a beginning during 1910–1920 in India and Malaysia; and later during 1930–1950 in Nigeria, Uganda, Ghana, and erstwhile Rhodesia. The network of forest plots in India is very wide and covers diverse forest types and environmental conditions. Long Term Research Sites (LTRS) in India have three different names such as Linear Tree Increment (LTI) plots; Linear Increment plots (LIP), and Linear Sample plots (LSP). Some are also called Permanent Preservation Plot (PPP). The common feature of all these plots has been the diameter measurement of individual trees at periodic intervals. Most G&Y studies have been conducted to provide information needed for forest management decisions such as: (i) the time of passage of trees through successive diameter classes; and (ii) effects of alternate harvesting and silvicultural treatments on growth of trees. For collecting data, one of the following two types of plots have been generally used, viz., Linear Increment Plots (LIP) (in form of transects or strip) for taking increment data on individual trees; and Permanent Sample Plots for collecting growth data/for trees and crops. A comprehensive description of laying out and measuring sample plots, logging, and sivicultural experimental plots in the mixed forest environment is contained in FAO (1981). As a rough guide, 100–200 sample plots, of 1 ha area and square in shape, uniformly distributed over the whole forest area are recommended. It has also been a common practice to select subjectively, number and measure individual dominant trees in forest blocks under varying logging/silvicultural treatments. The following publications present comprehensive information on G&Y Studies in the mixed tropical forests including collection, analysis, and interpretation of sample plot data: FAO (1989), Alder and Synnott (1992), Vanclay (1995) and Alder (1995). Reading of them in entirety is recommended. Commenting on the current state of knowledge, Vanclay (1995) considers it as ‘‘ironic that the most sophisticated models exist only for the most simple forest ecosystems (pure even-aged plantations), and that models for the most complex of forests, the tropical moist forest, tend to be rather primitive or lacking’’. In the above context, the network of Forest Dynamics Plots (FDP) established through the initiative of the Center for Tropical Forest Science (CTFS) of the Smithsonian Tropical Research Institute, in conjunction with scientiﬁc collaborators around the world, could play a valuable role.

13.5 Growth and Yield of Individual Trees

13.5.1 Methods of Research

Growth and yield at individual tree level may relate to forests and/or trees. Brandis (1858) is the earliest recorded study of growth rate of teak in forests of Myanmar, 13.5 Growth and Yield of Individual Trees 139 based on annual rings counting. Trees occurring on the farmland, trees in line or grown in homestead gardens are all parts of such studies. Almost 80 % of GDP contribution of the forest sector in India is accounted for by agro-forestry. From a methodological angle, the system of permanent plots needs to be replaced by a system of permanent trees as done in linear increment plots with a complete documentation of site factors and farming practices. The mathematical functions like Chapman-Richards may be still useful in agro-forest conditions as they relate to individual organisms as well as populations. The single tree observations could be further distinguished under the following methods. • Annual ring studies including: – Stump and stem analysis; – Increment borings; • Linear Increment plots The Silviculture Research Code (Vol. 3) by Grifﬁth and Bakshi (1947) gives detailed procedures for the collection of stump/stem analysis data and increment borings, equipment needs, staff, and labor requirement.

13.5.2 Stump and Stem Analysis

Techniques under this group are applicable to only those species which show distinct annual rings. With recent technological improvements, according to a report, growth—rings can be observed on a larger number of species, than was possible earlier. In all such cases the technique, no doubt, is most useful, as it provides time-series information based only on one-time measurement. A stump section of Cedrus exhibited in the FRI Museum provides growth records dating back to the thirteenth century. The stump and stem analysis, integrated with inventory, requires destructive sampling, which is a serious limitation in the application. In stump analysis, the method can provide information on the rate of diameter growth at stump height only. However, with the help of some additional measurements like seedling height and age and taper data, a relation between growth rate at stump and breast height can be established. In stem analysis annual ring measurements are done at a number of sections along the tree stem from which complete growth information about a tree, viz., growth rate of diameter, height, and taper (and, therefore, the volume) can be obtained. The information on age—height relation obtained from stem analysis is especially useful for construction of site index curves. Detailed procedure is available in the FAO (1981). Ideally, selection of trees for stump and stem analysis should be done prior to felling. In most cases, however, advantage is taken of ongoing felling operations in the forest. This, however, is likely to introduce bias of an unknown extent. 140 13 Growth and Yield Studies

13.5.3 Increment Borings

Increment core is generally taken on standing trees at breast height, which does not involve tree felling of any kind. From the core, information about rate of growth at the reference height is directly obtained (cf. stump analysis). Compared to stem analysis, however, the information provided by increment cores is much less, being limited to only one variable, viz., diameter and to one section only, viz., the breast height. As the technique is very simple and quick it can easily be integrated with other measurements in a forest inventory. This is a very desirable feature of the technique, which enables collection of growth data along with other inventory variables and also a correlation of growth with site and stand parameters. Some important points in carrying out studies with the help of annual ring measurements are: (i) Selection of trees for study should be done preferably through a sampling procedure as subjective selection is likely to introduce bias, (ii) The relevant stand and site characteristics of the sample trees should also be recorded. The best way is to conduct growth studies as an integral part of a forest inventory design. This will allow correlation of growth data with other inventory parameters (like stand, stocking and size classes, dominant height, site quality etc.). (iii) To reduce cost, only a fraction of trees standing on a ‘‘sample plots’’ should be selected as ‘‘sample trees’’. (iv) The desired number sampling units (ﬁrst stage selection) and number of trees in each unit (second stage selection) depend on the variability of stand and site conditions on one hand; and precision requirements on the other. FAO (1981), as a rough guide, makes the following recommendations (Table 13.4):

Table 13.4 Recommended number of bore cones for reliable growth modeling Objective of study Total number of Number of trees per sampling units sampling unit Growth modeling with the help of 300 5–6 increment core data Site index curves with the help of stem 50 1–2 analysis

Around 100 bore cones, well distributed in a sub-region (as a part of National Forest Survey), should be adequate for the purposes of dendro-climatological studies. While setting the precision goal it is desirable to take an integrated view of all possible sources of error in estimating growth with the help of annual ring measurements. The two important sources of systematic error are: improper boring direction and de-formation of core due to boring pressure. They are discussed in detail by IUFRO (1966). Several investigations have been carried out on random error arising due to sampling. Matérn (1961) reports the following results based on 13.5 Growth and Yield of Individual Trees 141 observation of 37 sample trees taken from the experimental plots of the Swedish Forest Research Institute (Table 13.5):

Table 13.5 Standard errors reported in dendrological studies Situation Standard error in the estimate of diameter growth (%) One increment core from each tree and section 19.0 Two cores on each section for each tree • taken at right angles 10.0 • taken in opposite direction 12.5

If the object is only to determine the periodic current annual increment observation of cores aided with simple pocket lens in the field should suffice. The expected measuring error will be of the order of 1 mm. For dendro-climatological studies a precise measurement of annual rings, under laboratory conditions is essential. The reported measuring error under laboratory conditions is usually of the order of 0.01 mm. It is estimated that one technician can measure about 30 complete increment cores in one day. Some problems connected with measurement of growth and age in tropical forests need mention here. (i) In the early ages (say first 5 years) growth rings are not well-defined. (ii) False and sometimes very narrow rings are formed as it also happens in temperate forests. (iii) Irregularities in the development of annual rings, within the same species from tree to tree have also been reported. Weck (1955) found that P. halepensis, P. merkusii, and Podocarpus cupressina, which in sub-tropical/temperate conditions show clear annual rings, exhibit irregular pattern under the tropical moist conditions. The following observations of Costor (1927), based on a study on 45 trees of Hevea brasiliensis, are quoted by him to provide an example of behavior of a broad leaved species. 21 stems had clear annual rings 5 stems had periods of 9–11 months 1 stem had a period of 13 months 18 stems had very irregular patterns. Under Swedish conditions the variation of annual growth can be of the order of 60–70 % from one year to the other and the difference between two consecutive 5-year periods can be up to 20 % under extreme conditions (1934–1939 and 1940–1944 in a subregion of South Sweden). A test was done to compare growth patterns using bore cones from coniferous forest in northern India with that in Sweden. The magnitude of variation was found to be of the same order. The annual variation of growth is of a much higher relevance in the tropical zone than in the temperate region for the reason that the crop rotation periods are much 142 13 Growth and Yield Studies shorter. It is a fact that results from permanent sample plots that are fairly valueless unless the variation in growth is not calibrated for.

13.6 Applications of G&Y Research in Forest Management Planning

Figure 13.2 shows a typical forest management decision-making process: the information on forest characteristics is obtained by forest inventories; and the information on development of growth and yield of trees/stands is provided by the G&Y research plots; and the two groups of information along with forest management options are used together to study implications of various forest management decisions. The outputs of model studies then guide the decision making process. Among the earliest examples of a combined forest inventory and growth and yield study in tropics is due to Dr. Brandis in 1856 to estimate annual allowable cut of teak trees (above certain selection size) on a sustainable basis in the mixed forests of Burma. The Brandis Method of yield calculation is well known to all tropical foresters and is still used as a model for yield regulation. The FAO/UNEP Forest Resources Assessment 1980 Project reported 41.3 million ha under scientiﬁc management, which constituted 4.3 % of the total reported forest area. Tropical Asia contributed 39.1 million ha, of which 32.5 million ha was located in one country, viz., India. The rest of 8.8 million ha were distributed over 18 other countries. The term forest management was used covering only wood production and implying ‘‘controlled application of harvesting regulations, complemented by appropriate silvicultural and protective measures designed to maintain and improve the productivity of the forests’’. The current forest area under management according to June 2011 Report of ITTO was 30.6 million ha. Another important issue is the lack of connectivity of growth and yield research with the reality. The problem is generic, but the following examples are given to illustrate the situation.

Fig. 13.2 The role of growth and yield studies in forest management 13.6 Applications of G&Y Research in Forest Management Planning 143

13.6.1 Hill Dipterocarp Forests of Malaysia

The observations appear in a report of Malaysian-German Sustainable Forest Management and Conservation Project (1994–2005): ‘‘The total volume of all trees, with dbh 5 cm+, in the virgin forest is about 340 m3 per hectare. In the areas logged during 1991–1996, the estimated stock is 220 m3, and in the areas logged during 1982–1991 is 240 m3. Thus, the standing volume is reduced approximately by 120 m3 after logging. However, only 45 m3 of that volume actually reached the sawmills. This clearly indicates that the practice of using the so-called ‘‘cutting intensity’’, i.e., the log volume reaching the mills as the indicator for logging intensity severely underestimates the real impact logging has on the stands, and on the ecosystem, because it neither considers the damage to the residual stand nor the non-utilized timber residue. To maintain sustainability, the forest growth has to compensate for the reduction in stocking volume. Considering the available data on growth rates in hill dipterocarp forests and, especially, current logging practices, this is unlikely to happen within the harvesting cycle of 30 years prescribed by the Forestry Department’s Selective Management System (SMS). Enforcing RIL and introducing LIL may help to resolve this problem. Earlier, timber harvesting has caused a shift of species composition from dipterocarp-dominated forest to non-dipterocarp-dominated forest, and within the non-dipterocarps a shift from high-value to low-value species. Silvicultural treatment of the young dipterocarp regeneration (saplings and small poles) should aim at restoring the dominance of the original species. It is obvious that TOR of the growth and yield research needs to take into account the stand composition after harvesting; and ﬁnally measures for the res- toration of the original ecosystem.

13.6.2 Mixed Tropical Forests, Indonesia

The soft landing forest policy was issued in 2002 with the aim of saving the remaining natural forests by reducing the national annual allowable cut. In actual practice, however, the policy could not be implemented as it was really intended. The constraint, as one may guess, was the scarcity of growth data which is needed for determining the right annual allowable cut. In the absence of the critical data, the policy was implemented in a modiﬁed fashion. Ideally, the annual allowable cut of each single management unit must be calculated to come up with the aggregate national allowable cut. However, since it was not possible, the approach was the other way around, the national allowable cut was somehow determined, which was further disaggregated by province, and ﬁnally at the local level. It seems that the national forest inventory system with a network of permanent and temporary plots might provide a robust approach for guiding sustainable forest management. 144 13 Growth and Yield Studies

Since the Stockholm Conference on Human Environment, 1972, the environmental functions of forests have got international acceptance. The G&Y research has to give them due attention, as a meeting organized by ITTO and CIFOR in 2005 in Bogor, Indonesia, has rightly pointed out. ‘‘In the future other aspects (viz. non-timber) related to permanent sample plots are becoming more and more important and it is timely that we discuss in this workshop all the aspects related to permanent sample plots; not only growth and yield aspect, but all aspects related to data, information that we are gathering from the permanent sample plots’’. These are very important and challenging calls, as there are many unsolved issues related to quantiﬁcation of ecosystem services CIFOR (2005). Using instruments of REDD+, efforts are being made to provide support to sustainable forest management. The growth and yield research on mixed tropical forests and plantations will form an integral part of this initiative and required resources provided for such purposes.

Recommended Further Reading

Alder D, Synott TJ (1992) Permanent sample plot techniques for tropical Mixed Forests, Oxford Forestry Institute, Oxford, UK Alder D (1995) Growth Modeling for Mixed Tropical Forests by. Oxford Forestry Institute, University of Oxford, UK Condit RG, Hubbell SP, Lafrankie JV, Sukumar R, Manokaran N, Foster RB, Ashton MS (1996) Species-area and species-individual relationships for tropical trees: a comparison of three 50- ha plots. J Ecol 84:549–562 Costor CH (1927) Quoted in Weck 1955 FAO (1980) Forest volume estimation and yield prediction, FP 22/1: volume estimation, Rome FAO (1980) Forest volume estimation and yield prediction, FP 22/2: yield prediction, Rome FAO (1981) Manual of forest inventory with special reference to tropics, Forestry Paper 27, FAO Rome FAO (1989) Studies on the volume and yield of tropical forest stands: 1: Dry forest formations, Forestry Paper (51/1) FAO Rome Grifﬁth AL, Bakshi SR (1947) The silviculture research code, v.2. The statistical manual; v.3. The tree and crop measurement manual. ICFRE, Dehradun Hägglund B (1979) Evaluation of site quality in connection with a model for large scale forecasting of timber yields. Department of Forest Survey, Royal College of Forestry, Sweden CIFOR (2005) Report on meeting at Bogor, Indonesia IUFRO (1966) Fehler bei der zuwachs bestimmung in besondere der Zuwachsbestmmung. IUFRO, Vienna Matérn B (1961) On precision of diameter growth from increment Borer, IUFRO 13 Congress Band 2 Matérn. In: Bo Ranneby (ed) Published by Swedish University of Agricultural Sciences Section of Forest Biometry, Umeo Nilsson NE (1980) Personal Communication Nilsson NE (1996) An alley model for forest resources planning. In: Bo Ranneby (ed) Essays in honour of Bertil Matérn. Swedish University of Agricultural Sciences, Section of Forest Biometry, Umea, Sweden Nyyssönen A (1981) Information on biased growth curve resulting due to use of small plot sizes (personal communication) Recommended Further Reading 145

Pandey D (1983) Growth and yield of plantation species in the tropics. FAO Misc paper Pretzsch H (2009) Forest dynamics, growth and yield: from measurement to model. Springer, Germany Stage AR (1976) Forest inventory data and construction of growth models, IUFRO 1978 XVI Congress Vanclay JK (1995) Growth models for tropical forests: a synthesis of models and methods, Forest Science vol 41/1 Weck (1955) Forstliche Zuwachs und Ertragslehre. Neumann Verlag, Germany Chapter 14 Estimating Potential Productivity of Forests

14.1 The Need for Potential Productivity Estimation

In 1956, a Swedish geographer S. S. Paterson published a paper with the title: ‘‘The Forest Area of the World and its Potential Productivity’’. In this paper, the author presented an index to derive forest potential productivity for all climatic regions of the world from only a small amount of easily obtainable climatic parameters. The FAO Forestry Department was obviously very interested in the paper and requested the International Union of Forest Research Organizations (IUFRO) to check the validity of Paterson’s index. In November 1958, a working group was formed including eight scientists from the main climatic regions of the world under the chairmanship of Prof. J. Weck of Germany. Since time was rather short (to the 1961 IUFRO Congress in Vienna), the working party could check the validity of the index only in those regions, where climatic and forest productivity data were already available (Weck 1961, 1962). Members of the working party soon came to the conclusion that an index based on climatic data alone cannot completely explain the quality class of specific forest sites, because the site quality does not depend on climatic influences alone, but also on the characteristics of the soil and the genetic make-up of species growing at the site. Still, it was of great interest (with special reference to tropical regions) to determine, whether an index based on climatic data alone could be developed for estimating average potential productivity of forests in regions with a certain climatic uniformity and normal distribution of soil quality classes. Obviously, this assumption could be tested statistically only in the temperate and boreal zones, on account of non-availability of climatic and forest productivity data in the tropics. Weck (1955) had already carried out studies for the forest regions of Germany on climate and forest productivity relation and found good correlation coefficient of 0.83 between mean values of sustained yield of forest and the corresponding values of a climatic index proposed by him. Altering the index value by 100 units

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 147 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_14, Springer-Verlag Berlin Heidelberg 2013 148 14 Estimating Potential Productivity of Forests altered mean annual yield of the forest area by 0.69 ± 0.14 m3 per year and hectare. The unanimous view of the working party submitted to FAO was as follows: • In general, the quality class of individual forest sites cannot be determined accurately enough solely from a climatic index. • For restricted regions there are indications that it may be possible to ﬁnd empirical functions, based on physiological principles, which permit sufﬁciently accurate correlation between climatic index and the mean potential production of forest areas within those regions. The situation regarding availability of climatic data in the tropics has since then considerably improved. What is lacking presently is the permanent sample plot network, which needs the immediate attention of countries. Regional cooperation in sharing such data would be most useful, which may be facilitated by FAO.

14.2 Description of Climatic Indices

The indices proposed by Paterson and Weck are the two most used ones for estimation of potential productivity of a forest area based on climatic variables. Their approach is empirical and consists in estimating an index by multiplying positive factors and dividing by presumed negative factors. The climatic index needs to be correlated with forest productivity based on sample plot data through a regression function. The following section presents a brief account of the models and studies to validate them.

14.2.1 Paterson Climate–Vegetation–Productivity Index

The Climate-Vegetation-Productivity (CVP) index based on climate data has the form (Paterson 1956): Tv P G E I ¼ Ta 12 100 The variables used are the following: I = The productivity Index Tv = Mean temperature of warmest month (C) Ta = Mean annual range of temperature between coldest and warmest month (C) P = Mean annual precipitation (mm) G = Length of growing season in months E = Evapotranspiration reducer (based on latitude and giving generalized total annual radiation received as percent of that at equator). 14.2 Description of Climatic Indices 149

The productivity of a forest site is done using a two-step process. First, the CVP index for that site is calculated based on climatic data; and then the productivity is estimated by applying the regression equation based on sample plot data.The regression equation established by Paterson for estimating mean annual increment was as follows: Y = 5.2 logX-7.25; where Y = Maximum mean annual increment of the plot (m3/ha/year); and X = CVP index of that site. The log is to the base 10. Example: The CVP index value for Dehradun is 924 as the parametric values are: Tv = 28.4, P = 2160, G = 6, E = 47.0, Ta = 15.6. Applying the equation, the estimated MAI for Dehradun (Y = 5.2 log 924-7.25) comes to 8.17 m3/ha.

14.2.2 Validation of Paterson Index for India

Seth et al. (1971), as the Central Silviculturist of the Forest Research Institute, India, used the existing sample plot data for different sites of India and also estimated the values of climatic indices based on the existing climatic data of those sites, where plots were located. The results of studies are given in Table 14.1.

14.2.3 Weck Productivity Index (WPI)

Weck had already carried out investigations on climate and the potential productivity in various regions of Germany. Later, he extended his work to the tropics with some modiﬁcation of the empirical relationship (Weck 1962). The main assumptions made were the following:

Table 14.1 Maximum mean annual increment (MAI) and potential productivity of forest for India Geographic location Sample Species MAI Potential plot no. productivity (m3/ha) Dehradun division, Uttarakhand 10 Shorea robusta 7.84 8.17 South Raipur division, Chhattisgarh 18 Shorea robusta 9.24 6.54 Udaipur division, Rajasthan 2 Acacia arabica 4.54 4.55 West Almora division, Uttarakhand 73 Cedrus deodara 6.16 7.54 Kurseong division, West Bengal 3 Shorea robusta 9.52 10.27 Kurseong division, West Bengal 13 Shorea robusta 9.36 10.27 (plantation) Puri division, Orissa 16 Shorea robusta 8.33 9.14 Lakhimpur division, Assam 8 Artocarpus 12.24 10.68 chaplasha South Chanda division, Maharashtra 1 Tectona grandis 4.20 6.50 Source Seth et al. 1971 150 14 Estimating Potential Productivity of Forests

• In tropical areas, respiration is less if night temperature is low (dT); the net production of organic matter is greater if the length of daylight is longer (S). • The relationship between the amount of precipitation and the net production of forests is not linear. A continual increase in precipitation above 2000 mm/year will correspond to a successively smaller increase in net production. • Net production is directly proportional to the length of the growing season (G). Months within the growing season must be free of frost and have a mean temperature of at least 10 C. A month is considered to be in the growing season if 12P/(T ? 10) [ 20, where P is the total monthly precipitation (mm) and T is the average monthly temperature. • Net production is linearly related to average humidity of the air (H). Vapor- pressure data were used to derive the average humidity. • The effect of precipitation on net production is less if the temperature of the growing season (Tm) increases. The model proposed was:

WPI = {dT * S * (P1 ? HP2) * G * H}/(100 * Tm) where dT (C) is the diurnal difference between average maximum temperature and average minimum temperature of the warmest month of the growing season; S (hours) is the mean length of daylight during the growing season; P1 (dm) is the average annual precipitation up to 20 dm; P2 is the average annual precipitation exceeding 20 dm; G (months) is the duration of the growing season; H is the annual average relative humidity of the air; and Tm (C) is the mean annual temperature of the warmest month of the growing season.

14.2.4 Validation of Weck Productivity Indices

FAO FRA 1990 used the site productivity data available in the Inventory files of the Scientific Research Program TELS AT T3/02/005 of the University of Gent, Belgium. A first screening revealed that just 12 inventory records were suitable for the analysis since the other were related to inventory of commercial species only, thus producing biased estimates of biomass. In addition, observations were added from the FAO Forestry Paper 51/1 (1984) relating to dry woodlands of Africa. The final inventory dataset included 28 observations distributed across the tropical Africa. A regression model was run and the following equation was found to be the best fitting: Biomass (tons/ha) =-0.000226 * WPI2 ? 0.559552 * WPI; The observed R2 value was 0.89, which shows such procedures have practical values. A graphical presentation of WPI and Potential Forest Yield is given in Fig. 14.1. 14.3 Recent Availability of Climatic Data for the Tropics 151

Fig. 14.1 The relation WPI and Forest Productivity (m 3/ha/year) between WPI and potential 9 yield 8 7 6 5 4 3 2 Series1 Potential Productivity 1 0 0 500 1000 1500 2000 Weck Index values

14.3 Recent Availability of Climatic Data for the Tropics

The situation regarding availability of climate data for the tropics had much improved by the end of 1990, thanks to the efforts of the FAO Agro-Met Division, which made extensive efforts between 1961 and 1990, to gather data from nearly 30,000 stations and compile them in form of a database, designed to extract, analyze, and map the climate information. Further, this entire database (labeled FAO WORLD CLIMATE DATA) was made available online (warning: it’s a 150 MB download!). The information includes among others the following variables: • Name and number of station, latitude, longitude, and elevation • Precipitation • Temperature estimates including average, mean max, mean min, mean day and night • Vapor pressure and wind speed • Sunshine % and total radiation • Evapotranspiration. All the data are on the monthly basis. In addition, there are estimates of average growing season by comparing interpolated daily amounts of rainfall with potential evapotranspiration. The growing season details include: • Type of season (with or without a dry season) • Number of dry days • Number of intermediate days before and after the humid season • Number of wet days • Indication whether there is more than one growing season. The above data lend to calculation of the saturation vapor pressure, the vapor pressure deﬁcit, and the relative humidity. The vapor pressure deﬁcit is of high relevance, since it is proportional to the evapotranspiration capacity of air. A brief discussion of some of the above variables relevant from the perspective of potential productivity estimation follows based on Nilsson (1997). 152 14 Estimating Potential Productivity of Forests

14.3.1 Temperature

Temperature is considered to be the dominant minimum factor for tree growth in the boreal zones. In Sweden, the growing season is the period when the mean temperature is above 5 C. In the tropical regions, temperature is not that decisive. At higher temperature, there is increase in the evapotranspiration: much water is consumed which may lead to a hot and dry non-growing season. Temperature may be used mainly when calculating humidity/aridity estimates such as the PET (potential evapotranspiration) or the WPD (water pressure deﬁcit). Data requirements may be monthly averages of maximum and minimum temperatures. Weck made use of the diurnal difference between average maximum and average minimum temperatures of the warmest month of the growing season in degrees celsius.

14.3.2 Growing Season

In the applications of Paterson and Weck, the growing season has been calculated by means of an aridity index introduced by Martonne in the beginning of this century. They changed Martonne’s annual index to apply for months. Included in the growing season are those months where the aridity index estimated by the expression: (12 * Precipitation)/(Temperature ? 10) is more than 20. In the agro-meteorological database, FAO has estimated the growing season for agricultural crops by calculating the number of ‘‘wet days’’ plus the number of ‘‘intermediate’’ days. Wet days are when the average precipitation is bigger than the corresponding potential evapotranspiration (PET). Intermediate days are added for part of the period when precipitation is more than 0.5 PET and for a period after the growing season corresponding towards 100 mm of potential evapotranspiration. Comparisons between the aridity index and FAOs method show quite similar averages with much less standard variation than in the case of monthly aridity indices.

14.3.3 Relative Humidity

Weck made use of the average annual relative humidity and considered that productivity was directly proportional to that estimate. Humidity is the most efﬁcient predictor of productivity and annual average as done by Weck gives good results. Author calculated the average relative humidity during the growing season under anticipation that this variable would be more relevant than the annual value. 14.3 Recent Availability of Climatic Data for the Tropics 153

14.3.4 Day Length

Day length may not be so important under tropical conditions. It is included in Weck0s index. The author calibrated the vegetation period by multiplying with a factor day length/12 with the assumption that it might be possible to develop a rough index with some comparability between tropical and temperate regions.

14.3.5 Precipitation

Precipitation data are of course quite important. Weck made use of annual precipitation and considered that the impact of precipitation decreased above the level of 2000 mm per year. Above 2000 mm he added the square root of additional precipitation to the amount of 2000 mm. It is a major problem to find out to which extent and under which conditions ‘‘excess’’ precipitation should be disregarded. Without any doubt, the relevance of precipitation is overestimated in the Weck index. It is essential to find some ‘‘robust’’ estimators. The question is: How much of monthly precipitation is superfluous under which conditions and how to reduce in order to get an efficient estimator of tree growth? One choice is to express precipitation in PET- units. Provisionally, he chose to disregard precipitation above three PET. The ultimate but more expensive solution would be to make use of the ‘‘run off model’’ developed by MET Department in Sweden. One of the problems with this model is that it makes assumptions as to the historical levels of evapotranspiration, if one wants to go back in time with the studies. Evapotranspiration might have changed over time due the changes in the vegetative cover. This, however, requires undertaking research that may become more and more important in connection with the climate change problem.

14.4 The Areas of Further Research on Forest–Climate Relation

The following proposal for research on forest and climate is based on Nilsson (1997) and reproduced as such without elaborating the ideas further.

14.4.1 Productivity Indices

Especially, in connection with planning of afforestation programmes, there is a need for a broad classification of sites (regions) with respect to potential productivity of forest stands. In such cases, a productivity index aims at indicating, in broad terms, 154 14 Estimating Potential Productivity of Forests the potential site quality for tree growth. It needs to be complemented with information on soil properties, hydrology, etc. A productivity index may contain components that indicate the influence of specific factors or the combined effects of several factors. Paterson CVP and Weck WPI are examples of efforts to develop approximate indications of potential productivity combining several factors. In a study for the Africa region, precipitation alone was found to be the single most important factor to predict wood increment. A potential productivity index should be transparent enough to say something about the basic site factors that would guide the selection of species. However, one cannot possibly develop one climatic model for each species especially not since basic data from sample plots are almost non-existing or non-available. One would rather develop a robust index system by means of which it is possible to build-up experiences on the species productivity level by post classification of existing data on growth and yield of concerned species.

14.4.2 Annual Growth Indices

Annual growth indices aim at indicating the annual variation of tree growth caused by the weather conditions. The variation of the year ring widths at breast-height is being used since long ago in countries with temperate forests like Sweden. In fact any detailed growth and yield study would require that the basic data are calibrated to represent average climatic conditions. This is especially true for short rotation tree crops. In Sweden, initially, year-ring series based on fairly short periods (up to 30 years) were used, but it was soon realized that longer series were needed. Hence it was decided to collect bore cone samples from trees that could give us series of at least 60 years. One of the difficulties involved is that after an early maximum there is an average annual decrease of year-ring widths which must be corrected for. It is difficult to distinguish between this normal ‘‘age’’ trend and trends which are caused by fertilization effects of air pollution that may change over time. Some 40 years ago Dr. Bengt Jonson, in Sweden, made a serious effort to substitute the year-ring method by a model based on climatic factors. He did not succeed for the reason that he could not manage to estimate availability of soil water, especially not in the beginning of the growing season, when the winter snow conditions and the snow melting conditions are decisive. He concluded that it was better, for the time being, to continue using year-ring indices and he laid down considerable and successful efforts to improve the technology in this respect. One is much better off now, thanks to considerable advancement of science in general and improvements of computing technology and data handling. One could for instance try to apply the run-off model which has been developed by the Swedish Meteorological and Hydrological Institute (SMHI) and which can provide daily estimates of the soil water deficit (defined by the millimeters of precipitation which are needed in order to start water run off within the watershed). 14.4 The Areas of Further Research on Forest–Climate Relation 155

Fig. 14.2 Annual forest growth ﬂuctuations on account of annual climate variability

Figure 14.2 shows the annual variation of diameter growth in a region of South Sweden year by year from 1910 to 1970, together with some climatic variables (number of days with a soil water deficit bigger than 150 mm, temperature sums and precipitation sums during the growing season and length of the growing season in days).The growing season is defined as the number of days with a mean temperature above 5 C. These data are matched with true indices of tree growth based on sectioning and measurement of the annual year rings at each level of measurement. Although there are no statistical studies undertaken on the data we can observe that the cyclic variation of ‘‘dry’’ and ‘‘wet’’ years as far as ground water deficit is concerned corresponds towards a similar but delayed variation of tree growth. Under Swedish conditions the variation of annual growth can be of the order of 60–70 % from 1 year to the other and the difference between two consecutive 5-year periods can be up to 20 % in extreme conditions (1934–1939 and 1940–44 in a sub-region of South Sweden). A test that was done in 1970 by Nilsson and Singh on bore cones from coniferous forests of northern India, showed that the magnitude of variation of diameter growth was of the same order as observed in Sweden. The annual variation of growth is of a much higher relevance in the tropical zone than in the temperate region for the reason that the crop rotation periods are much shorter. It is a fact that results from permanent sample plots would be fairly valueless unless the variation in growth is not calibrated for climatic fluctuations. There seem to be good prospects for developing a climatic index which could be used as a rough growth index and at the same time form a basis for a productivity index when based on average data for a long period.

14.4.3 Hardness Index

A hardness index is related to the risk for extreme weather episodes and should serve as a guide when considering the risks for damage which might endanger the 156 14 Estimating Potential Productivity of Forests survival or the undisturbed production of speciﬁc tree species. Analysis of annual climatic data from a number of stations might give some guidance for further thinking in this respect. There are two main approaches. One is to classify the tree species with respect to hardness or robustness. The other is to develop tools for describing weather hazards of different kinds depending on deviation from the average climatic conditions. In both cases one would beneﬁt from documentation of historic forest damage caused by extreme climatic conditions or weather episodes such as very sudden drop of temperature in the beginning of the vegetation period. ‘‘Hardness index’’ might consist of several components and be used in connection with some sort of risk analysis. It is likely that we can obtain knowledge and ideas from the agriculture sector in this respect.

14.4.4 Suggestions for a Climatic Index

Based on the foregoing considerations, Nilsson recommends to work with just a few elements that one may consider relevant and express them in percents of some optimal or maximal value. Based on data of 3258 stations included in the FAO agro-climatological database, he suggests to make use of the following parameters:

(i) Vegetation period calculated as an approximation of FAO’s method making use of monthly precipitation and estimated monthly evapotranspiration, both expressed in millimeters. In addition, periods with an average temperature less than 5 C are to be excluded from the growing season. Such low temperatures may occur on high altitudes even within the tropical region. Example is Sri- nagar, India. (ii) Relative Humidity (RHUM) during the vegetation period (corrected for day length): The relative humidity is already expressed in percent in the FAO dataset. (iii) Relative Radiation Sum (RRAD): Sum of daily radiation (expressed in calories per cm2 per day) during the vegetation period (corrected for day length) expressed in percent of 168,000 which is nearly equal to the value for Bogor in Indonesia which appears to have the highest radiation amongst the stations in the database. Technically the sum has been calculated as the sum of monthly averages during the vegetation period to be multiplied by 30. (iv) Relative Precipitation Sum (RPREP): Sum of monthly precipitation during the vegetation period expressed in PET units up to 3 PET during a month. This value has been expressed in percents of 36 which is the highest possible value that can be achieved. A provisional index (NEN-Index): Lacking the knowledge about the interdependence of these three variables, Nilsson recommends to calculate a simple geometric mean value according to the formula (RHUM * RRAD * RPREP)1/3. 14.4 The Areas of Further Research on Forest–Climate Relation 157

By scanning through the data it appears that RPREP has the biggest ﬂuctuations and that further work should concentrate on ﬁnding cheap estimators of the water regimes.

14.5 South–South Cooperation in Climate–Change Research

As there is a general paucity of sample plot observations in the tropics, pooling of such information on an international basis for important species would provide a cost-effective strategy to promote sustainable forest management planning. For this purpose, it is essential that a protocol is established for laying sample plots, gathering data and sharing of database. An example will be presented of a study on estimation of potential productivity of teak conducted by Pandey (1996) using pooled sample plot observations of ﬁve countries, where the species grows naturally or has been introduced. Two values of MAI were investigated, viz., biological rotation age (maximum MAI of teak is between 10–20 years); and the technical rotation age, which was kept at 50 years. Pandey (1996) made use of 86 sample plot observations originating from 24 stations of 6 countries. The number of sample plots per station varied from 1 to 14. Four climatic variables were used as independent variables, namely, the annual rainfall, the annual relative humidity, the temperature of the warmest month of the growing season, and the growing season in months; and two indicators of productivity as dependent variable, viz., (i) mean annual increment at 50 years of age, when the species gives high value return termed as the potential yield by the author; and (ii) the maximum MAI estimated using published yield tables for the site. The value of R2 in the former case was 0.61 and in the latter case it was 0.72. Results of an analysis of variance (between and within stations) gave the following results for the potential yield (see Table 14.2).

Table 14.2 Analysis of variance of potential yield of teak Component Sum of squares d. f. MS Between stations 289.10 23 12.57 Within stations 89.93 62 1.45 Total 379.03 85 4.46 Source Pandey (1996)

Since the climate is uniform on plots located at the same station, the component within station reﬂects soil and other local factors; whereas the component of variance between stations is mainly due to climate, climate-induced site variations, and plantation management. The mean square within stations was 1.45 compared to between stations variance of 12.57; and the total at 4.46. European countries, in spite of the long tradition in growth studies, consider cross-country cooperation a major beneﬁt for research. Tropical countries need to 158 14 Estimating Potential Productivity of Forests strengthen studies on plantation development and also promote pooling of sample plot data and launching of integrated G&Y Research Projects, which may bear fruits say 10–20 years hence.

Recommended Further Reading

Nilsson Nils-Erik (1997) Estimating productivity of tropical natural forests and plantations by climatic factors (Personal communication) Pandey D (1996) Estimating productivity of tropical forest plantations by climatic factors. Swedish university of Agricultural Sciences, Umea Patterson SS (1956) The forest area of the world and its potential productivity. Royal University of Goteborg, Sweden Seth SK, Kaul ON, Sharma DC (1971) Potential productivity of Indian forests. In: Symposium on tropical forest ecology emphasizing organic production, New Delhi Weck (1955) Forstliche Zuwachs und Ertragslehre, Neumann Verlag, Germany Weck J (1961) Ueberpruefung der Tauglichkeit eines klimaindex zur Bestimmung forstwirts- chaftlichen Producktionspotentials. IUFRO Meeting, Vienna Weck J (1962) Climate and forest potential productivity, Germany: the magazine of the Federal Republic Ref. No: 26/1962 Chapter 15 Land Evaluation Techniques for Forestry Planning

15.1 The Purpose of Land Evaluation

Land evaluation is concerned with identifying and evaluating potential options for land use and / or forestry, so as to meet national or local needs, and in estimating the consequences of these changes. It does not make a decision on behalf of the user, but only provides options with full scientific data to aid decision. Land evaluation techniques have a potential value in helping rational land use / forestry planning and in resolving the multi-disciplinary land-use conflicts as exemplified below: (i) Which kind of land use? Frequently of critical importance in land-use planning are decisions between forestry and other major kinds of land use. Such decisions include the conversions of forests to agriculture; the conversion of other uses of land to forestry, as in the establishment of forest plantations; and the allocation of land between different uses in multipurpose land settlement, development, or watershed management projects. (ii) Which kind of forestry? Before management can be planned, decisions need to be taken on the relative importance of the different multiple uses of forests. A basic choice is between forests intended first and foremost for timber production, and those in which conservation is the primary aim. (iii) Which kind of forest management? Given the land that is to be under forest, and the kinds of products and other benefits expected from it, there are decisions on methods of forest management designed to achieve these aims.

15.2 Description of Land Evaluation Techniques

FAO work on land evaluation techniques began in the early 1970s. Contributions were brought together in a series of meetings starting in 1973, culminating in the publication of the Framework for Land Evaluation in 1976.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 159 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_15, Springer-Verlag Berlin Heidelberg 2013 160 15 Land Evaluation Techniques for Forestry Planning

Experience in using the Framework in field projects and training courses showed that, although it provided a comprehensive coverage of the basic concepts, substantially more detail was required for specific evaluations. This led to the preparation of more specific guidelines, for example, one on rain fed and irrigated agriculture and one on forestry which was published in 1984 in form of a FAO Forestry Paper No.48: Land evaluation for forestry. Figure 15.1 gives a synoptic view of land evaluation procedures. These commence with planning the evaluation, based on discussions between those who requested the evaluation and those responsible for carrying it out. The first need is

Fig. 15.1 A synoptic view of land evaluation procedures 15.2 Description of Land Evaluation Techniques 161 to clarify the objectives of the evaluation: the human need that it is directed to, policy aspects, and the kinds of land-use which are relevant for consideration of timber plantations, recreation forestry, forestry for watershed management, etc. At the same time any constraints to land-use changes are ascertained, e.g., legal aspects or acquired—rights of local population or the advisability of resettling existing users. The organization responsible for the survey then conducts a rapid review of the physical, economic, and social conditions of the area, and the kinds of data available, e.g., soil surveys, silvicultural studies, and socio-economic evaluations. This leads to the planning of the programme of work, covering specifications of the results, information needed, surveys and studies required, staff, timing, and an estimate of cost. Final approval for carrying out the evaluation may be delayed until this point. Three sets of field activities are then set in motion: studies of land utilization types, surveys of land units, and collection of economic and social data. The studies of land utilization types commence with basic descriptions of each alternative use (e.g. plantation or natural forest management or complete protection) and first estimates of required inputs and expected outputs. As the survey progresses, these descriptions are amplified and refined to provide information on the results from the evaluation. Following the description of land utilization types, comes determination of their land-use requirements. These are the sets of conditions needed for the successful operation of the land utilization type, examples are the growth requirements of possible tree species, or terrain conditions favorable to alternative methods of harvesting. They include also the conservation requirements, the conditions necessary to avoid soil erosion, adverse effects on water flow regime, or other undesirable environmental changes. From this review of land-use requirements we learn what kinds of land will be best suited to each possible use. Parallel with the above, surveys and specialized studies of land units are started. The purpose of these is to provide a mapped basis of relatively homogeneous areas of land, to which suitability for use will be related. The kind of land units chosen will vary with the objectives, ranging from forest types or land systems in a survey at national level, to site classes, soil series, land facets, or other more specific mapping units in detailed surveys for forest management planning. The third set of basic studies is the collection of economic and social data needed for analysis of alternatives. Like the basic surveys of physical data this is a time-consuming activity and so needs to be started early on. Besides its use later on in the procedures, there is an early need for such information in working out details of the land utilization types. As the land units are surveyed, some of their land qualities and characteristics will have been determined: for example, slope angle may be employed as a basis for defining land units and also be one of the properties of those units. Other land characteristics may need to be surveyed after the initial mapping, particularly if the land units have been taken from an existing survey. Next comes the critical stage of land evaluation: the comparison of land-use with land. At this point the requirements of each of the defined land utilization 162 15 Land Evaluation Techniques for Forestry Planning types are known, and so are the relevant features of climate, slope, soils, etc., for each of the land units. The first stage of comparison is called matching, in which these two sets of data are combined. This is initially done by assigning factor ratings which indicate partial suitabilities based on each land quality considered: moisture requirements compared with moisture availability, requirements for erosion avoidance as compared with erosion susceptibility of the land, and similarly. The separate ratings are then combined to give a first approximation to the land suitability classes. These provisional suitabilities are then analyzed with respect to their environmental impact, economic consequences and social consequences. Environ- mental impact will already have been taken into account in landuse requirements and matching, but an overall review is conducted at this stage as an added pre- caution. The nature and amount of economic analysis will vary considerably with the purpose of the survey, ranging from a strategic review in resource Inventories for national planning purposes to detailed analysis of costs and returns wherever investment is intended to follow. Social analysis of proposed changes should always be conducted: it covers such aspects as whether the changes will fulfill needs, whether they will prove acceptable, and if any indirect economic and social consequences will occur, e.g., reduction in the area of industrial forest land. After the provisional suitabilities have been analyzed by these further means, the final land suitability classification is drawn up. This shows suitability of each alternative kind of land use on each land unit. The final stage is that of presentation of results. These include details of the land utilization types together with land suitability maps (cf. Fig. 15.2). Com- parison of requirements of land-use with properties of the land will have led to management specifications being drawn up; for example, which establishment methods are best for each type of land, for each use on each relevant land unit. An analysis of the environmental impact of the possible changes is given, together with the results of the economic and social analysis. Finally, there will usually be data from basic surveys and specialized studies, e.g., of climate, vegetation, and economic information, which is placed on record for future use (Fig. 15.2). The end product of a land evaluation is not a land-use or management plan. Rather, it provides information on the consequences of several alternative kinds of land use, as applied to different mapped parts of the study area. For example, in a survey for the siting of plantations, the evaluation will indicate which are the best, and which are the relatively favorable sites from a land suitability viewpoint, as well as areas on which plantations should certainly not be sited. Except in broad reconnaissance surveys, estimates of costs and timber yields will be given. But the planning decision on whether to go ahead with a plantation project is based on considerations beyond those of land evaluation, whilst even the matter of location may be influenced by external considerations. The same applies in a survey intended for a forest management plan; evaluation provides considerable information on which to base such a plan, consisting of alternatives and their predicted consequences, but the actual management planning forms a later stage in the process of land development. 15.3 Applications of Land Evaluation Techniques in Forestry 163

(a)

Land Units A. Timber Production Mechanized Harvesting

B. Fuelwood Production Manual Harvesting C. Protection Forestry (b) Land Utilization Types Land Unit A. Timber B. Fuelwood C. Protection 1N2eN2eS1 2N1qS2eS3o 3S3rS2rNR S1 S1 NR

Fig. 15.2 Examples of land suitability maps: (a) suitability maps for individual uses; (b) combined tabular legend

15.3 Applications of Land Evaluation Techniques in Forestry

Table 15.1 gives examples of forestry land-use planning, combining type of decision and level of planning. In conducting a land evaluation it may be helpful to identify where the objectives lie in relation to this decision matrix. There is an interaction between planning at different levels. On the one hand, strategic guidelines at the national and provincial levels clearly provide a framework for district or project level planning, and similarly as from the district to the local levels. There is also a converse effect. Every decision taken at a local level, 164 15 Land Evaluation Techniques for Forestry Planning

Table 15.1 Application of land evaluation techniques in forestry Levels of Types of decision planning Kinds of land use Kinds of forest use Kinds of forest management National/ Conversion of forest to Forestry for production, Forest management policy provincial agriculture non-forest protection, recreation, uses to forest, National etc.; Catchment parks protection zones District/ Conversion of forest to Implementation of forest Preparation of forest project agriculture or non- land allocation for management plans. forest uses to forest. production, Silvicultural systems. Establishment of forest protection, recreation, Provisions for local plantations. Land etc. users, for amenities and allocation in nature conservation multipurpose projects Watershed management Local Adjust boundaries of forest Local needs and their Implementation of forest reserve. Extend village incorporation in forest management plans forest area management plans e.g., to expand or reduce the area of production forest, has a cumulative impact on the overall position. Proposed land-use changes at any level of planning therefore need to be appraised in the light of their effects on the wider situation.

15.4 Land Evaluation for Forestry Planning at the National Level

The following discussions will mainly focus on the national level forestry planning and elaborate the role of land evaluation at this planning level. According to FAO (1986) the main objectives of forestry planning at the national level are: (i) To optimally satisfy the current and future national needs of major and minor forest products. (ii) To ensure ecological balance in the development program by paying adequate consideration to the preservation of flora and fauna aid to the conservation of soil and water; and to integrate environmental concerns into development planning at all levels. (iii) To bestow special considerations on any particular aspects of forestry development for the country as a whole or for any specific geographic areas or target population group guided by national considerations of functional, social, and/or spatial equity or efficiency. (iv) To harmonize forestry sectoral goals with that of other sectors in order to maximize national growth. 15.4 Land Evaluation for Forestry Planning at the National Level 165

Nilsson (1985) recommends the following logical sequence of activities for formulating the national forestry plan: (i) Analysis of broad national objectives. (ii) Analysis of (or formulation of) broad forestry objectives. (iii) Formulation of provisional production objectives for forestry subsectors. (subsistence forestry, industrial forestry, environmenta1 forestry) (iv) Implementation of provisional production and consumption studies for subsectors. (v) Analysis of main options with respect to—utilization of existing forests— creation of new forests—land utilization. (vi) Identiﬁcation of knowledge gaps, formulation, and implementations of inventory and surveys. (vii) Analysis and evaluation of alternatives. (viii) Final decisions making. Nilsson (1985) considers production consumptions studies mentioned above as important prerequisites for the national forestry planning process and outlines procedures for this purpose by main forestry sub-sectors, viz., Subsistence For- estry, Environmental Forestry, Farm Forestry, and Industrial Forestry. The author describes a computer-based model (known as Area Production Model) for simu- lation of long-term land use and forest resources development; the calculations are done in three phases: Phase I estimates future consumptions; Phase II gives future forest production under alternative land and forest-use assumptions, and Phase III provides production consumption balances. Use of system approach is expected to improve the planning process and make it more reliable and effective by answering to some basic questions such as the following: (1) Which product should be produced in what quantities? (choice of output). (2) Which of several feasible production processes should be utilized (choice of process). As mentioned earlier it is not only a question about choice of technology, but possibility of modifying social, political, and economic environment (land reforms, taxation policy, etc.) related to production. (3) What should be the input mix? There is a certain interchangeability among basic inputs—land, labor and capital, more especially between labor and capital. What is the appropriate choice for a country at a particular stage of development? From the above questions and problem formulations it may be obvious that land evaluation constitutes an important basis of national level forestry planning by providing it good scientiﬁc knowledge about national land and forest resources and their productivity under alternative types and intensities of land use. 166 15 Land Evaluation Techniques for Forestry Planning

15.5 Land Evaluation for Forestry Planning at the District Level

An Expert Consultation lasting over a week, including foresters, academics, researchers, economists, and planners, was held at the FAO Regional Office at Bangkok in December 1985 marking the closing phase of the Forestry Departments Sub-Programme on land evaluation for forestry. The most important realization of the Consultation was that while there is everywhere a common approach using shared techniques, the problems are always location-specific. The general approach to planning is the same in all types of districts, and the techniques employed are similar; but the differing problems from one type of district to another require different emphasis, as do the opportunities available for their solution. Therefore, the prerequisite to any planning is a carefully designed information needs assessment defining the problems of the area and postulating questions to be explored during the planning process and to be answered by the results. The experts emphasized that forestry planning at the district level must take place within the larger context of integrated land-use planning, considering objectives, and constraints at higher levels (regional and national) in a multi-level planning framework. Land evaluation, with its basic approach of identifying and analyzing possible land-use alternatives, is central to such planning. One of the papers presented described land-use planning as ‘‘…the whole range of activities needed in a systematic process of deciding what is needed to meet desired objectives and how to bring about the required changes.’’ One discussion group report summarized the basic stages in land-use planning as: (1) Analyze the present situation and its problems. (2) Identify possible improvements through land utilization types. (3) Collect the required data. (4) Analyze the consequences of alternative courses of action. (5) Select an optimal set of developments. (6) Plan in detail for their implementation. A basic need in all types of districts is a wider role for forestry; in addition to its traditional role in wood production, forestry must strive to solve the problems of the local people. Consequently, there is increased importance for social forestry and agroforestry (as well as grazing and other land-use alternatives) in addition to the conventional priorities of production forestry and environmental forestry. Modern problems call for a wider range of forestry planning. Forestry development no longer takes place wholly within what has commonly been called forest land, but also in areas of mixed land use, including agriculture and grazing. The practice of land evaluation—in which environmental, economic, and social criteria are used to analyze changes in land use—enables comparison of competing claims for uses of land. Land evaluation thus links forestry development planning with general district development planning. 15.5 Land Evaluation for Forestry Planning at the District Level 167

Such planning requires interdisciplinary, and thus inter departmental cooperation to a greater degree than is presently found in many countries. Foresters can cooperate effectively with agricultural planners and other specialists only if there is a recognized institutional framework for doing so. Finally, all discussion groups stressed the importance of securing the participation of the local people in planning and decision making. Local participation occurs mainly at four stages of planning: analysis of present problems, suggestions for improvements, selection of responses, and the determination of the role of local institutions in planning.

Recommended Further Reading

FAO (1984) Land evaluation for forestry. Forestry paper 48, FAO Rome FAO (1985) Report on FAO consultation on district level forest and land use planning, Bangkok, 1–7 Dec 1985, FAO Rome FAO (1986) Land evaluation for afforestation planning by R. B. Lal, FAO Rome FAO (1996) Guidelines for land use planning, FAO Development Series 1, FAO Rome Kassam AH, van Velthuizen HT, Singh KD (1992) Land evaluation for forest resources development planning at national level, FAO Rome, 70 p Part IV International Dimensions of Forest Resources Assessments

Abstract This part discusses capacity building from an international perspective. The first world forest inventory report was published in 1948 and the last in 2010. The situation regarding country capacity to provide reliable information for national and international use does not seem to have changed much. ‘‘Many developing countries have difficulty in reporting because their monitoring systems are inadequate both for domestic as well as international needs (FRA2010)’’. Further, reporting responsibilities have been added to countries by the Convention on Biological Diversity (CBD) and the UN Framework Convention on Climate Change (UNFCCC). To achieve the capacity building objectives in a timely and effective manner, elements of international capacity building strategies, including strengthening of organizations’ own capacity to provide technical support to countries, is presented, following the UNCED Agenda 21 Recommendations. Chapter 16 Identiﬁcation and Evaluation of Environmental Functions of Forests

16.1 The Problem Formulation

In the opening chapter, attention was drawn towards emergence of new genera of environmental problems like climate change, biodiversity loss, and land degradation, which have a global dimension. Though technology provides means for establishing a global level database using Remote Sensing and GIS, the scientific community is yet struggling to define the nature of problems and to set up a sound analytical basis. A related problem is lack of capacity in many tropical countries to provide reliable information on forests and still less to undertake sustainable forest management measures. In this chapter, we will present a conceptual basis and also identify main variables associated with the environment benefits and then proceed with the evaluation process. Before presenting a theoretical framework, we introduce a fairly abstract concept of a ‘‘system’’. De Chardin (1959) writes: ‘‘Existence of ‘system’ in the world is at once obvious to every observer of nature, no matter whom. The arrangement of the parts of the universe has always been a source of amazement to men. This disposition proves itself more and more astonishing as, every day, our science is able to make a more precise and penetrating study of the facts. The farther and more deeply we penetrate into matter, by means of increasingly powerful methods, the more we are confounded by the interdependence of its parts. Each element of the cosmos is positively woven from all the others. From beneath itself by the mysterious phenomenon of ‘composition’, which makes it subsistent throughout the apex of an organized whole. From above through the influence of units of a higher order which incorporate and dominate it for their own ends. It is impossible to cut into this network, to isolate a portion without it becoming frayed and unraveled at all its edges.’’ The biosphere is the highest level of ecosystem organization. It stretches a few kilometers above and below the earth’s surface and includes Land, Ocean, and Atmosphere (Fig. 16.1).

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 171 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_16, Ó Springer-Verlag Berlin Heidelberg 2013 172 16 Identiﬁcation and Evaluation

Fig. 16.1 A model of the biospheric system

Land could be further classified according its dominant use into cultural and natural. The urban and rural areas form a part of what could be called cultural ecosystems. They are characterized by a concentration of cultural activities. The forest ecosystem could be considered as an example of a natural ecosystem dominated by natural (primary production) processes. The reason we have chosen to distinguish natural from cultural ecosystems is based on the observation that most of the services (and disservices) in which we are interested are related to processes and interactions within and between them. One could also expect to find important interactions between these systems and the biosphere. Long-term changes in climate and vegetation as side effects of urban industrial processes are for instance seriously suspected. Table 16.1 shows the average annual budget of CO2 attributed to Land Use, Land-Use Change, and Forestry during the years 1980–2000. In the reciprocal manner, atmospheric system affects the land use/land cover through the climate regime. From the forestry perspective, we consider the problem of identification and evaluation of forest-related contributions to environmental change as an extension of the national forest inventory problem. A national forest inventory thus could be defined as a continuous activity which aims at providing a relevant basis for long- term forestry planning and control on the national level including the commodity and environmental functions of forests. An efficient planning and control of timber production is itself a challenge and certainly demands much more extensive inventory knowledge than has been commonly recognized hitherto. The majority of developing countries do not know the extent of their timber resources nor the removels nor the actual and potential production. Planning and control of 16.1 The Problem Formulation 173

-1 Table 16.1 Average annual CO2 budget in Gt C yr for 1980–1989 and 1989–1998 (90 % conﬁdence interval) 1980–1989 1989–1998 1) Emissions from fossil fuel combustion and cement production 5.5 Æ 0.5 6.3 Æ 0.6 2) Storage in the atmosphere 3.3 Æ 0.2 3.3 Æ 0.2 3) Ocean uptake 2.0 Æ 1.0 0.7 Æ 1.0 4) Net terrestrial uptake = (1)-[(2) ? (3)] 0.2 Æ 1.0 0.7 Æ 1.0 5) Emissions from land-use change 1.7 Æ 0.8 1.6 Æ 0.8 6) Residual terrestrial uptake = (4) ? (5) 1.9 Æ 1.3 2.3 Æ 1.3 Source IPCC (2000) environmental forestry and its integration with the corresponding activities for wood production, is by far a completely untouched and unstructured problem area. To form a basis for ecosystem modeling, we start with presenting a list of variables deﬁning the state of the two ecosystems.

16.2 Components of Cultural and Natural Ecosystems

Figure 16.2 shows a model of the Cultural Ecosystem. It seems logical to make a distinction between components (or aspects) related to ‘‘man’’ himself and to environment. A list of variables deﬁning the Cultural (Human) Ecosystem follows. I Environmental component 1. Abiotic (e.g. climate, topography, soil, and water) 2. Biotic (e.g. plants, animals, microbes) II Human component 1. Biological (i.e. physical and psychological) 2. Social and political

Fig. 16.2 Model of a cultural ecosystem 174 16 Identiﬁcation and Evaluation

3. Scientific and educational 4. Technological 5. Economical 6. Aesthetics 7. Ideological (beliefs and values) Of course such a classification must be considered subjective and facultative. The order of the components must not be considered as a ranking list. On the contrary, the various components form an organized whole, where change in one of the components implies simultaneous changes in all the others. For economic development to occur, technological, social, and political changes are crucial. Investigations also show that the modern man retains many biological and mental characteristics of his remote ancestors. His physiological needs and drives, his responses to environmental stimuli have remained almost unchanged. These findings naturally have great importance for rural and urban planning. Develop- ment of environmental forestry (forestry for people) therefore should involve not only analysis of how to adapt management regimes to serve multiple-use purposes but also studies of people’s real needs and demands and of their behavior and attitudes. Let us now consider classification of components of a forest ecosystem (see Fig. 16.3). I. Abiotic 1. Atmospheric (including climate, solar energy, air etc.) 2. Hydrological 3. Edaphic (including rocks and minerals) II. Biotic 1. Plants 2. Animals 3. Microbes

Fig. 16.3 Model of a forest ecosystem 16.2 Components of Cultural and Natural Ecosystems 175

III. Cultural influences 1. Harvesting and utilization 2. Cultivation of land 3. Sowing and planting 4. Irrigation, fertilization 5. Exploitation, devastation, etc. 6. Urbanization The elements of abiotic and biotic components of the forest ecosystem are the same as was stated for the cultural ecosystem where they were called environmental. In the ‘‘area production approach’’ (refer Chap. 2) they were referred to as natural production factors. It should be noted that the term environment implies that some sort of center is defined. In the area production approach we have put man in the center and hence the use of the concept environment may be acceptable. In a natural ecosystem it is more doubtful if the concept has a precise meaning. As can be seen above we have used the concept environment when describing components of a cultural ecosystem but in the case of a natural system we only refer to biotic and abiotic components. Primary production in a forest ecosystem results from interactions among all of the components. Like in other systems, change in one component involves changes in all others. Influences of forests on soil, water, and climate regimes will depend on all factors biotic and abiotic and cultural components. Any general statement about the environmental influences on the forests without specification of all factors, environmental and cultural has little meaning and may even be misleading. Similarly, valuation of a particular forest production is dependent on the state of the human ecosystem which takes advantage of the forest production. Timber production, recreation, anti-pollution influences have different implications in rural and urban ecosystems. It is thus obvious that no general valuation system can exist. Valuation is therefore meaningful only in a specified environmental and cultural context, which needs to be stated.

16.3 The Ecosystem Dynamics

The Human Ecosystem (C) interaction with the Forest Ecosystem (N) over time (T) is pictorially shown in Fig. 16.4. The sub-scripts 1, 2, 3, etc., refer to values of the system over time. The terms input and output are used in a wider sense and include not only planned direct inputs and effects, but also side effects associated with forest change such as land degradation, biodiversity loss, etc. The cultural aspects are being highlighted as they are the most important determinants of the forest area and ecosystem processes of a country at any time. In the course of time, human influence on the natural ecosystems has been continuously increasing and spreading. This has been reflected in a higher yield per unit of area, which was possible only through drastic modification of the natural ecosystems—change of gene pool, change of soil properties by cultivation, 176 16 Identification and Evaluation irrigation and fertilization and use of insecticides, pesticides etc. We have reached a stage where the cultural influences on the natural ecosystems may be described as ‘‘ecosystem engineering’’. For that reason it might be rational to consider natural and cultural ecosystem as forming an interdependent system representing a higher level of organization. If we also introduce time as an important dimension, the idea can be pictured by the model in Fig. 16.4. Planning obviously aims at guiding the transition of the cultural ecosystem from the state C1 at time Tl to another state C2 at the time T2 and so on. For this purpose we make use of the ecosystem N1 to produce an output O1 that requires the input I1. As a result of the production process the state of the natural ecosystem will change from N1 to N2. The terms input and output have been used in a wide sense and include not only planned effects but also such side effects of the process that arise due to ecosystem interactions. Thus the term ‘‘input’’ stands for: input of matter and energy to the production process; other influences of the cultural system on the natural system such as pollution, radiation etc. Similarly ‘‘output’’ stands for: planned products and services; and other influences of the natural system on the cultural system.

16.4 The Ecosystem Variables and Change Model

The ecosystem model shown graphically in Fig. 16.4 is presented in Table 16.2 in a matrix form with a hypothetical example where three alternative uses of a forest ecosystem are compared with each other with respect to their effect on the selected state variables of the integrated system.

Fig. 16.4 An integrated natural and cultural system 16.4 The Ecosystem Variables and Change Model 177

Table 16.2 Change matrix showing state variables, production alternative, and changes Speciﬁcation of system variables Current status Production alternatives 123 Associated changes in the ecosystem

Natural ecosystem N DN1 DN2 DN3

Cultural ecosystem C DC1 DC2 DC3

As was argued earlier no relevant discussion about changes can take place without a complete speciﬁcation of the environmental and cultural state. Further elaboration of state variables of the natural and cultural ecosystems are given in Table 16.3.

Table 16.3 Detailed list of state variables of an integrated natural–cultural ecosystem Main State variables components Natural Ecosystem: Abiotic Air and climate Temperature, precipitation, humidity, wind, radiation, air Topography Slope, aspect Soil Physical and chemical properties Water Availability, movement, primary water production

Natural ecosystem: biotic Flora Vegetation cover: Composition, diversity, yield Fauna Composition: Abundance, yield Cultural Harvesting/utilisation, cultivation, seeding, planting, irrigation, fertilization, exploitation, devastation, urbanization

Cultural ecosystem: Environmental Air and climate Temperature, precipitation, humidity, wind, radiation, air Geology Rocks and minerals, earthquakes Soil Physical and chemical properties, avalanches, run-off Water Availability and distribution of water with respect to cultural use, run-off, physical and chemical properties, water pollution Biotic Insects and pests

Cultural ecosystem: Human Biological Basic biological needs Demographic Population, urbanization Economic Production and distribution, of goods and services, income employment, ownership, proﬁtability Educational Self-preservation, social adjustment, literacy Social/political Law and order, ownership, income distribution, employment Ideological Religion, doctrines, science, art, aesthetics, beliefs and value 178 16 Identiﬁcation and Evaluation

16.5 Example of a Study Using Ecosystem Approach

Ecosystem approach permits building of a multivariate, hierarchic, and dynamic database, which permits study of environmental problems associated with forest changes from local to global levels. Though not explicitly mentioned, the chapters included in Parts III and IV of the book, directly or indirectly are based on ecosystem concepts. The following example is intended to provide an illustration of applying the ecosystem concepts for evaluating the impact of current land-use practices on soil and water run-off to plan an effective soil and water conservation program. Table 16.4 gives a list of system variables, Fig. 16.5 shows the analytical model in form of a flowchart, and Table 16.5 presents the findings. The first step in the assessment consists in classifying the land according to landform on a physiographic basis such as: mountain top, mountain side slope, hummock/hillocks, valley, piedmont plain, alluvial plain, and levees. Next, landform units are sub-divided into different run-off potential classes based on variables such as, geology, physiography, slope, vegetation, land use, land management, soil, and erosion hazard. The traversing is combined with field observations, which are transferred to 1:50,000 scale Survey of India topographic sheets. This is followed by delineation and codification of run-off potential mapping units at a micro- watershed level by hierarchical stream order. An effort is made to maintain a viable size of the micro-watershed for use in the subsequent detailed soil survey for preparation of forest working plans and soil conservation programmes. The statistics on run-off potential for two watersheds are presented in Table 16.5. The Dikrang sub-catchment of the NE Himalayas has a total area 161,731 ha. Of this, 81,858 ha (50.61 %) is under high to very high risk of erosion and has ‘‘to be rehabilitated as soon as possible under an integrated watershed management programme to check the devastating floods in the lower sub-catchments’’; 17,190 ha (10.63 %) is under shifting cultivation and needs to be brought under a sustainable land-use system; 46,019 ha (29.07 %) is under severe to moderate erosion needing sound soil–water conservation measures. The Ravi Sub-catchment in NW Himalayas has an area 89, 985 ha. As expected, 76.05 % of the area belongs to very high potential run-off class. Very

Table 16.4 Identification of state variables determining soil run-off Main State variables Further specification components Climate Rainfall Amount, distribution Topography/Soil Slope, length of slope & aspect Texture, structure Infiltration capacity Vegetation Land cover, humus Vegetation types, density Hindrance to water flow classes Cultural factors Cultivation practices, harvesting Soil working, land practices, grazing, and fire management practices 16.5 Example of a Study Using Ecosystem Approach 179

Fig. 16.5 Methodology for run-off potential mapping: AIS & LUS (1991) important fact to note is that in both the sub-catchments, land is mostly unmanaged and forest degraded. The survey data and maps provide valuable information and form the basis of developing effective plans for greening of land in the prevailing socio-economic and environmental condition. A deforestation model (refer Chap. 3), based on FAO FORIS database, provides a basis for estimating likely forest area loss due to expansion of subsistence agriculture associated with uncontrolled population growth in the forested regions. Chapter 17 presents a model to estimate risks of species-richness loss in different 180 16 Identiﬁcation and Evaluation

Table 16.5 Soil and water conservation priority areas in Dikrang and Ravi watersheds SI. No. Priority category Dikrang, Arunachal Middle Ravi, J&K (NE Himalaya) (NW Himalaya) (Score Max: 100) Ha % Ha % 1 Very high (Above 75) 26,716 16.52 56,485 62.77 2 High (71–75) 55,142 34.09 11,950 13.28 3 Medium (66–70) 32,420 20.05 5,250 5.85 4 Low (61–65) 11,932 7.38 8,400 9.34 5 Very low (Below 61) 35,521 21.96 7,900 8.78 Grand total 161,731 100.00 89,985 100.00 Source All India soil and land-use survey reports of the Dikrang and Ravi Watersheds (AIS & LUS 2000) ecological zones associated with deforestation and Chap. 18 describes approaches for climate change reporting. A prerequisite for application of these models/ approaches is availability of time series of reliable forest inventory and demographic data and, of course, the need for capacity development to maintain and use existing data for solving emerging problems, national and international.

Recommended Further Reading

AIS & LUS (1991) Methodology for run-off potential mapping, New Delhi, India AIS & LUS (2000) Survey of the Dikrang and Ravi watersheds, ministry of agriculture, New Delhi, India De Chardin PT (1959) The phenomenon of man, Harper, New York, pp 43–44 IPCC (2000) Land use, land use change and forestry, Cambridge University Press, Cambridge Chapter 17 Ecological Zoning and Assessments of Biological Diversity in the Tropics

17.1 The Need for Ecological Zoning

Ecological zones classify land/vegetation formations in an area based on similarities and differences of their ecological attributes. Forests belonging to the same zone have comparable structure, phsiognomy, and respond to management treatments in a common manner. The land units too falling in the same zone have comparable physical and chemical properties. A striking similarity in the characteristics of rain forest formations across the continents was presented earlier in Chap. 11. This is true for the other forest formations too. An ecological zone map would enable countries in the same zone to exchange the experiences about forest inventory, growth, and yield. With the above in mind, FAO started a programme during 1981–1986 on classification and mapping of tropical vegetation types on an ecological basis in collaboration with the International Institute for Vegetation Mapping (ICIV) at Toulouse in France. The objective was to provide a common system for classification and mapping of vegetation cover in the tropics, which could be used as a basis for assessment and monitoring of changes in the extent of forests by ecological zone, country, region, and the world. The assumption was that ecological zones constituted the basis for development offorests and such maps will have many uses and promote exchange of scientific and technical know-how related to forest management and forest ecology among countries. The idea was first successfully tested in selected countries of tropical Asia including India, Philippines, and Thailand in cooperation with the national institutions and later extended on an operational basis to all the tropical regions.

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 181 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_17, Springer-Verlag Berlin Heidelberg 2013 182 17 Ecological Zoning and Assessments

17.2 The Approach for Ecological Zoning

The methodology for the preparation of the map was ﬁnalized after a number of consultations between concerned ofﬁcials of FAO (mainly Dr. K.D. Singh) and ICIV (Dr. F. Blasco) and consultation with countries under the overall direction of Prof. P. Legris. The methodology followed for the preparation of the map consisted of the following steps (Sharma 1986):

17.2.1 The Choice of Parameters

For identiﬁcation and mapping of boundaries of ecoﬂoristic zones (EFZ) the following parameters were used: P = mean annual precipitations in millimeters ms = mean number of dry months in a year calculated according to Gaussen’s method R = average number of rainy days per year T = mean annual temperature tm = mean temperature of coldest month Hm = mean annual air humidity t = annual thermal amplitude

17.2.2 The Classiﬁcation and Mapping of EFZ

The classification and mapping of ecofloristic zones was done as a two-step process taking into account ecological and floristic parameters of the area under mapping. The first level of classification was based on the ecological characteristics including climate, physiography, major soil formation (as in FAO Soil Map of the world). The bio-climatic limits were determined on the criteria of mean annual rainfall, length of the dry season, relative humidity, and temperature. The physiographic contours and soils help in refining the bio-climatic limits. The second level of classification was mainly based on floristics. For the determination of eco-floristic zones, the dominant or characteristic forest species of indigenous flora was taken into account as described in the published national and international forest type and vegetation maps. Maps of the ecofloristic zones were supplemented by tables (as an inset) of the corresponding principal vegetation types (syn. formations). For each ecofloristic zone it is possible to give a complete classification of the vegetation types that it contains. This correspondence should permit the photointerpreter to recognize the types described. For each zone tables may be established between the vegetation types and possibilities of their recognition at different levels of remote sensing. 17.2 The Approach for Ecological Zoning 183

The formations were further subdivided on the basis of density, ranging from the woody types that are the most closed, to the types that are the most open. This expresses the different stages of the regressive series or, in very rare cases, the evolutionary tendency of the vegetation within a zone. The denomination of the formations was based on the classification of Yangambi (1956) and of UNESCO (1973). In order to avoid very long comments and repetitions, main bibliographic references to the fundamental publications and articles, providing complementary details of the phonological characteristics, floristic composition, dynamics, etc., of each formation, were given. This presentation at two levels of generalization, completed for each of the formations by the text and bibliographic references, constitutes a sufficiently complete tool for the denomination and interpretation of the formations located by remote sensing. More complete data may be introduced when needed for the study of large-scale aerial photographs of the sensitive zones selected for continuous monitoring. The classification used has essentially a cartographic aim because it is to be utilized for the interpretation of the documents of remote sensing for continuous monitoring of the tropical forest vegetation. The classification is flexible enough to be applicable at the four essential levels: • Global level (bioclimate, soils, formation class which is the first level of UNESCO’s classification) • Regional level (bioclimate, soils, formation groups, UNESCO’s third level) • Country level or national level (which includes floristic classes) • Management level or district level, which includes stand types and site classes. At the latter two levels it is obvious that detailed ecological, floristic, and physiognomic data are absolutely necessary. It is also recalled that the testing of the classification on aerial photographs, radar and satellite imageries wherever available on the various vegetation types, and their comparison with existing maps of recent forest inventories, might also be partly done by national institutions. The EFZ maps were prepared to begin with for Asia, followed by Africa, and finally South America. The three maps were prepared using similar criteria but obviously the number of primary and secondary level classes differ between the three continents depending on the specific ecological and floristic characters. In particular there are 19 ecological and 36 eco-floristic classes in South America and 10 Ecological and 30 eco-floristic classes in Tropical Asia and 16 main ecological zones and 59 eco-floristic zones in Africa. The EFZ map for Central America was made by the FRA Team at FAO HQ based on available data and maps. Since a main goal of EFZ classification was to assess the state and changes in the tropical forests by main ecological types at pan- tropical level, it was necessary to harmonize the classification of the three continental maps. Both tasks were carried out by the FRA90 Team. A common set of 11 ecological zones were defined for the three continents which was discussed and agreed by the Toulouse group. 184 17 Ecological Zoning and Assessments

17.2.3 The Validation Phase

The draft maps were validated using data from different sources which included Landsat imageries, published vegetation maps for the region, and information on bioclimate and ﬂoristics. The draft maps were sent to selected specialized institutions in the tropical region for validation and improvement. The initial maps were prepared on 1:1,000,000 scale which were later photographically reduced to 1:5,000,000 scale and the information was transferred on the base map ‘‘FAO soil map of the world.’’

17.3 The EFZ Map and the Database

Two sets of maps constituted the final output in the form of GIS overlays: ecofloristic zone (EFZ) and vegetation cover, both produced on 1:5,000,000 scale for all the three tropical regions. These overlays were transformed into a digital form and integrated with subnational boundaries as already described in Chap. 3. These spatial datasets have served a most useful purpose in many of FAO and other international studies such as modeling of forest cover changes presented earlier and evaluation of protected areas adequacy in the tropics described in FAO Forestry paper 112. The other potential uses cited are as under: • For estimating vegetation types area in the historic periods refer (refer Sect. 11.5). This information is useful for conservation and rehabilitation planning. • For identifying analogous areas having comparable EFZ characteristics to guide introduction of plant species in the analogous areas from one region to another. • For developing national level EFZ maps with further refinement, as and when countries take up the task of updating their maps using improved satellite data and more ground control. The FAO ecological zone and forest cover maps will be useful as reference documents.

17.4 The Tropical Forest Ecosystems Report 1992

The ﬁrst operational use of the EFZ dataset was for preparing the UNCED 1992 Report on the state of tropical forest ecosystems. For this purpose, the EFZ were grouped into four broad forest formations, one located in the upland and three in the lowland. Table 17.1 and Fig. 17.1 provide an overview of their distribution: the 1990 areas and rates of changes by main categories. The country level distributions are available in Annex 1 of the Forestry paper 112. By end 1990, the tropical lowland formation covered 1,544 million ha (88 % of the total tropical forest area) and upland (hill and montane) forests 204 million ha (12 %). In the lowlands, the tropical rain forests formed the largest part, namely 17.4 The Tropical Forest Ecosystems Report 1992 185

Table 17.1 Estimates of forest cover area and rate of deforestation by ecological zones Population Forest cover and deforestation Land Ecological zone area Density Change Areas Deforestation (1990) 1981–1990a 1990 1981–1990a million inh/ % per million % million % per ha km2 annum ha land ha annum Forest zone 4,189.7 57 2.4 1,748.2 42 15.3 0.8 Lowland formations 3,476.6 57 2.3 1,543.9 44 12.8 0.8 Rain forest 937.1 41 2.2 718.3 76 4.6 0.6 Moist deciduous 1,298.6 55 2.4 587.3 46 6.1 1.0 Dry and very dry 1,241.0 70 2.3 238.3 19 2.2 0.9 Upland formations 713.1 56 2.6 204.3 29 2.5 1.1 (Hills and mountains) Non-forest zone (Alpine 588.6 15 3.1 8.1 1 0.1 1.0 areas, deserts) Total tropicsb 4,778.3 52 2.4 1,756.3 37 15.4 0.8 Notes a Compound interest formulae were used for population growth and deforestation rate calculations b Totals may not tally due to rounding

Fig. 17.1 Distribution of main tropical forest formations at end 1990

718 million ha (or 41 %), the moist deciduous forests 587 million ha (33 %), and the dry and very dry zone forests 238 million ha (14 %). The rest of the forests, nearly 8 million ha, were scattered in the non-forest zone. The annual losses of forest by ecological zone were as follows: the tropical rain forests 4.6 million ha (0.6 %), the moist deciduous forests 6.1 million ha (1.0 %), the dry and very dry zone forests 2.2 million ha (0.9 %), and the upland formations 2.5 million ha (1.1 %) and the total 15.4 million ha. 186 17 Ecological Zoning and Assessments

17.5 Biodiversity Loss Associated with Tropical Deforestation

17.5.1 Problem Formulation

Biological diversity assessment is a very new and unstructured area of forest inventory. There are a number of outstanding problems still to be solved, before operational inventories on the subject can be undertaken. In the interim period, use could be made of the existing forest inventory data, in particular, those with reliable species identification. For the future, it is very important that countries undertake inventories with complete tree species identification above a certain diameter class, generally 10 cm DBH. It is important to mention some important differences in the survey procedures for assessing biodiversity with those used in the normal forest inventories. The concern of biodiversity studies is the whole of nature; complete pattern of spatial variations (and connectivity) and not only with spatial averages. For this purpose, identification of all elements of flora and fauna; and even soil organisms may be needed. An ITTO Research Report of the author (Singh 2005) with the title ‘‘Forest Biological Diversity: Assessment and Conservation Planning’’ introduces the subject from a forest inventory perspective. Recognizing the practical difficulties in reporting on biological diversity as a part of Global Forest Resources Assessment 2000, the specialists meeting in 1998 at Kotka (Finland) agreed that considerable progress in understanding the situation and trends for biological diversity in the world’s forests could be made by including the following parameters in the national and global forest assessments (FAO 1996): • ‘‘Naturalness’’ (breakdown of forest area into natural forest, semi-natural managed and unmanaged forests and plantations); • Protection status (according to six IUCN categories) • Fragmentation (using remote sensing) • Ecofloristic zones (breakdown of forest area and its protected status by ecological zones). The experts were of the view that natural forests and forest plantations need to be monitored and reported separately on a continuing basis. The two types of changes have different biological diversity implications and by adding them together, there may be a tendency to underestimate the gravity of the biological diversity loss associated with tropical deforestation. Modeling of the natural forest area loss associated with expansion of culture (in particular agriculture) could provide a reasonable estimate of relative loss of biological diversity with time. Such time series information is possible to be obtained with the help of satellite remote sensing by countries on a cost-effective basis. 17.5 Biodiversity Loss Associated with Tropical Deforestation 187

17.5.2 Modeling of Biological Diversity Richness Loss: FRA1990 Approach

The FRA 1990 approach consisted in developing species area relation for different ecological formations of the tropics based on the observed number of species reported by countries (i.e. species richness). In a generic form the model looks like: S = c. Az, where S is the number of species, A the area, c is a constant depending on the ecological zone, and z a constant depending on the plant group under inventory and the size of the study area. With a logarithmic transformation, the equation has a linear form: Log S = Log c ? z log A as shown in Fig. 17.2. It is important to stress the point that the model is not predicting species extinction, but the aim is only to indicate likely loss of species richness with a reduction in forest area. In fact, there is a subtle relation between different forms of biological diversity, viz., the ecosystem, species richness, and genetic diversity; and between ﬂoral and faunal diversity (see Fig. 17.3). The steps in estimating the risk of species richness loss (RSRL) were the following: (i) Species richness data from the existing archives of IUCN was compiled in the form of a database; (ii) Species—area relation was developed by main ecological region.

Fig. 17.2 Species-area relation. (Points shown in scatter refer to individual countries) 188 17 Ecological Zoning and Assessments

Fig. 17.3 Similarity in plant and animal diversity by ecological zones in Sri Lanka

(iii) Estimation of the current forest area by ecological region and deforestation by ecological region. (iv) Calculation of an index value expressing likely species richness loss as a function of the current rate of deforestation and the remaining forest area.

17.5.3 Species Area Relation by Ecological Zone

The species-area relation by ecological zone was considered as an essential input for assessing the risk of biological diversity loss. Basic data for species-area relations were taken from the World Conservation Monitoring Centre (WCMC) report ‘‘Global Biodiversity’’. In Fig. 17.2 the total number of Angiosperm species, recorded in 86 tropical countries, are plotted against corresponding country land area, both on a logarithmic scale to the base 10 (WCMC 1992). Guided by the scatter points and other data collected during the literature review, species-area relations by ecological region were established empirically as shown in Fig. 17.3. Different trends can be observed in the species-area relation according to the dominant ecological region, in respect of the slopes of the lines and their intercepts on the Y-axis. It may be noted that the assessment made of the species richness loss on account of ongoing deforestation is intended to raise awareness of the problem.

17.5.4 Rate of Deforestation by Ecological Zone

Estimates of forest cover at 1980 and the deforestation during 1980–1990 by ecological zone constituted another important input to calculation, which were taken from the FAO Report Forest Resources Assessment 1990: Tropical Countries (FAO 1993). 17.5 Biodiversity Loss Associated with Tropical Deforestation 189

Table 17.2 Loss of species richness associated with deforestation Forest area Species Risk of Forest cover area loss richness species Ecological zone richness loss 1980 1990 1981–1990 1990 1981–1990 million ha million ha million ha number (%) Africa Tropical rainforest 91 87 4.7 30,700 2.0 Moist deciduous forest 274 251 22.5 13,000 2.5 Dry and very dry zone 162 151 10.8 1,900 1.0 Upland formations 38 35 2.9 12,300 2.5 Asia & Paciﬁc Tropical rain forest 199 177 21.6 40,400 4.3 Moist deciduous forest 48 42 6.6 7,600 4.3 Dry and very dry zone 46 41 4.7 1,600 1.6 Upland formations 53 47 5.8 13,500 3.8 L. America & Caribbean Tropical rain forest 474 454 19.4 57,900 1.6 Moist deciduous forest 326 294 31.8 13,700 3.0 Dry and very dry zone 52 46 6.2 1,600 1.9 Upland formations 138 122 16.2 18,400 4.0 Sources FAO FRA (1993)

17.5.5 Risk of Species Richness Loss

Using the species area model, the number of species was estimated at two time periods for the current forest area and also 10 years earlier at a region level taking the rate of regional deforestation into account. The computed values are given in Table 17.2.

Example: The Species richness loss is computed by: (S1980 –S1990)/(S1990). Substituting areas in place of species, the expression could be rewritten as: z z z z z z (A1980 –A1990)/(A1980); or 1 - (A1990/A1980)or1- (A1990/A1980) . Substituting 1980 and 1990 forest areas of a region and the appropriate z values in the expression, we get the richness loss values given in Table 17.2 It is important to clarify that one is not justiﬁed in inferring species extinction from the risk of species richness loss, as the latter relates to the total number of species in a geographic area and not to a particular species. It is important to highlight that the forest area loss, which is different in different ecological regions, has also, due to the curvilinear shape of the species-area graph, a different impact on the impoverishment of species richness. Thus, deforestation in the tropical rain forests of Asia seems to have strong effects in terms of likely impact on species richness of angiosperms. The estimated risk of loss is twice, though the area deforested is about half, that of Latin America and the Caribbean. 190 17 Ecological Zoning and Assessments

The tropical rain forests of Asia seem to be most endangered in terms of the likely impact of deforestation on species richness loss. The estimated loss is twice, though the area deforested is about half, that of Latin America and the Caribbean. This fact may be taken as the implication arising from the model used. A good archive of historic inventory data would serve as a useful guide in developing the model. Another important issue is the forest degradation, which has also assumed a critical dimension, as presented in Chap. 11. The forest degradation process can take many forms: reduction in number of stems per ha compared to a well-stocked (un-degraded) forest, less number of species, etc. Again the process is to be described using time series data and not using cross-sectional data at one moment of time. The degradation may also affect the ground ﬂora, soil humus, etc. The entire process is gradual and somewhat hidden, but needs to be monitored carefully as it may lead to gradual forest productivity deterioration and other adverse environmental consequences. The forest area loss is given with a decimal digit to enable improved calculations of species richness. The model for estimating the likelihood of loss of species richness in an area on account of forest loss aroused a lot of discussion. There is nothing sacrosanct about it. The fact is that species diversity decreases with reduction in forest area. The exact values may differ from area to area and from forest type to forest type.

Recommended Further Reading

FAO (1993) Forest Resources Assessment 1990: Tropical Countries, FAO Forestry Paper 112, FAO Rome, Italy FAO (1996) Forest Resources Assessment 1990: survey of tropical forest cover and study of change processes, Forestry paper 130, FAO Rome Sharma MK (1986) A vegetation map of tropical continental Asia on 1:5 Million scale, Institute Of International Map of Vegetation Toulouse (France) Singh KD (2005) Forest Biological Diversity: assessment and conservation planning, WWF India and ITTO, Yokohama, Japan UNESCO (1973) International classiﬁcation and mapping of vegetation, Paris, France WCMC (1992) States of plant and animal inventories for protected areas in the tropics, Cambridge Yangambi (1956) Report of specialists meeting on phytogeography, Scientiﬁc Council for Africa, South of Sahara, Publication No. 22 Chapter 18 Forest Assessments for Climate Change Reporting

18.1 Reporting Requirements Under Kyoto Protocol

The Kyoto Protocol (KP) is linked to the United Nations Framework Convention on Climate Change (UNFCCC) signed at UNCED 1992. The major distinction between the Convention and the Protocol is that the Convention encourages industrialized countries to stabilize GHG emissions, the Protocol commits them to do so. KP was negotiated in 1997 and ratified in 2005 setting a legally binding target for a minimum of 5 % reduction in the 1990 level of emissions by industrialized nations (known as Annex I Parties) before the end of a first commitment period 2008-2012. The KP specifies, not only the information to be submitted by Parties, but also the way such information should be collected, analyzed, and reported. The following statements, dispersed in various Articles of Kyoto Protocol, are relevant for the present chapter:

The Parties included in Annex I shall, individually or jointly, ensure that their aggregate anthropogenic carbon dioxide equivalent emissions do not exceed their assigned amounts….(and take measures) to reduce their overall emissions of such gases by at least 5 per cent below 1990 levels in the commitment period 2008–2012 (Article 3.1).

Emissions resulting from direct human-induced land use change and forestry activities, limited to afforestation, reforestation and deforestation since 1990, measured as veriﬁable changes in each commitment period,…….. shall be reported in a transparent and veriﬁable manner and (peer) reviewed in accordance with Article 7 and 8 (Article 3.3).

Those Parties included in Annex I for whom land-use change and forestry constituted a net source of greenhouse gas emissions in 1990 shall include in their 1990 emissions base year or period the aggregate anthropogenic carbon dioxide equivalent emissions by sources minus removals by sinks in 1990 from land-use change for the purposes of calculating their assigned amount (Article 3.7).

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 191 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_18, Springer-Verlag Berlin Heidelberg 2013 192 18 Forest Assessments for Climate Change Reporting

The Kyoto Protocol further speciﬁes that ‘‘eachParty shall have in place a national system for the estimation of anthropogenic emissions by sources and removals by sinks’’ (Article 5.1).

Each Party shall incorporate in its annual inventory of anthropogenic emissions the sup- plementary information (on the method used) for the purpose of ensuring compliance with Article 3 (Article 7.1).

Formulate, where relevant and to the extent possible, cost-effective national, and where appropriate regional programmes to improve the quality of local emission factors, activity data and/or models for the preparation and periodic updating of national inventories of anthropogenic emissions etc (Article 10a). The reporting requirements of the Kyoto protocol look tough and stringent even for Annex I Parties!

18.2 The Approaches for Reporting Used by Annex I Parties

The Annex I Parties (viz. the Industrialized Countries) have a long tradition of management plan inventories, growth, and yield assessments using research plots and forestry account keeping. They have been supplying comprehensive forest resources statistics to the UNECE/FAO since FRA1948 on a regular basis. Their input data, broadly speaking, is derived from two sources: • Management plan inventories; and • National forest inventories The two inventory approaches are not competitive but complementary. Esti- mation of forest area and area changes, growing stock, increment and drain, which are important for carbon stock calculations, are done in a different manner in the two systems. They also have varying uncertainty associated with the size of reporting unit, time frame, and statistical characteristics under consideration. They are intended, respectively, for local and national level accounting. It is feasible to think of a combined approach for Kyoto Protocol purposes, in which detailed information needs may be met from management plan inventories and large area assessment from national forest inventories. In some countries (e.g. Poland), where forest management plan inventories are conducted on a centralized and national basis, quite reliable data on forest area and growing stock at the national level can be calculated. The assessment of changes between 1990 and 2012, however, will require application of approximation procedures yielding estimates based on assumptions or auxiliary information about the continuation or alteration of the trend. 18.2 The Approaces for Reporting Used by Annex I Parties 193

Relative Merits of Management Plan versus National Forest Inventories The forest management plan inventories collect detailed data, for relatively small management units, on variables like forest area, species composition and growing stock, which are updated usually on a ten yearly basis. Yield tables made from permanent research plots, combined with field and remote sensed data from aerial photographs, provide the basis for thinning, final felling and yield regulation. Forest inventories are, however, not intended to yield annual data and uncertainties in annual estimates based on the inventory data may be high. As the first commitment period will last for 5 years it is likely that the current forest inventories will yield data with high uncertainty. In case forest management plan inventories are considered an important source of data for the Kyoto Protocol purposes, the length of future commitment period should be extended to 10 years, or the inventory methods would need to be modified to provide estimates for shorter periods also. National Forest Inventories, on the other hand, are concerned with the reliable assessment of the state and changes in forest area and growing stock of a country. The permanent sample plot (PSP) techniques provide a cost-effective basis for estimating status of regeneration, increment, drain and trends. Estimations could be done on an annual basis using trend lines. A forest balance could be constructed from inventories using permanent sample plot data. This makes the process of volume assessment very transparent and also statistically sound

Table 18.1 presents results from the Swedish National Forest inventory, 1985–2008. Running 5-year averages are given referring to productive forestland outside protected areas. This illustrates how an ongoing national inventory provides annual data for carbon stock calculations required by Protocol over large areas. However, the precision of estimates for small area unit is low compared to management plan inventories. The recent trend in combining purely ﬁeld-based national forest inventories with remote sensing data may help in getting precise estimates on a sub-national basis (see Fig. 18.1).

Table 18.1 Results of Swedish national forest inventory (an example) Forest area Standing volume Annual cut Annual increment Years 3 3 3 1,000 ha Million m Million m Million m 1985 22,871 2,493 56.5 92.8 1990 22,590 2,582 60.2 96.4 1995 22,394 2,719 66.9 96.2 2000 22,588 2,810 64.2 98.0 2005 22,681 2,916 85.4 111.4 2010 22,477 2,943 68.4 111.6 Source National Forest Inventory of Sweden. Note that NFI-figures on annual cutting NOT are official statistics of Sweden since the Swedish Forest Agency calculates the official cutting based on consumption, stocks, import, export, etc. 194 18 Forest Assessments for Climate Change Reporting

Standing Volume, Standing Volume, Change of ---- End of Period – Beginning of Period = Standing Volume ---- Equation 1 3131 (2005) 2837 (1995) 293

Increment DRAIN Volume Change 1104 – 792 = 312 ---- Equation 2 (1996-2005) (1996-2005) (1996-2005)

Fig. 18.1 Forest balance calculations illustrated with volume and growth data from Swedish NFI

An equation, called forest balance, can be constructed from national forest inventories using permanent sample plot data. This makes the process of adjusting volume estimates for different dates very transparent: Growing Stock(at time 2) - Growing Stock(at time 1) = Growth - Harvest - Mortality Two types of calculations can be carried out using NFI data, using information shown on the left and right hand-sides of the equation (see Fig. 18.1). Equation 1 is based on standing volume data; and Eq. 2 is based on the growth and harvest data during the period. The unit is million m3. There is a high level of agreement between the two methods of calculating forest balances. With regard to change in total volume, the difference between them is approximately 20 million cubic meters standing volume, which corresponds to approximately 7 % of the entire change. The removal of standing volume during the 10-year period was over 25 % of the initial volume. Growth during the 10-year period was approx 35 % of the initial volume. During 1995–2005, the total standing volume increased by over 28 million cubic meters annually (Swedish Statistical Yearbook of Forestry 2010). Annual or periodic observations of growth and yield of forest crops are influenced by weather condition in the year or the period under observation. The Swedish national Forest inventory has developed a system for relating annual observations of tree and crop growth with weather conditions and making adjustments to arrive at a normal growth on the basis of average climatic conditions over the period under study. As per these results, as much as a maximum of 60 % of fluctuation may result around the trend line in a year. Between two 5-year periods the deviation in climatic indices around the mean could be as much as 20 %. A study made in coniferous forests of India showed the same order of fluctuations as reported in Sweden. The national forest inventories in Finland, Switzerland, and the USA combine remote sensing with field sampling. Sweden is now also making use of satellite data for post-stratification of forest sample plots. The remote sensing provides a basis for stratification and geo-referencing on a small area basis. The gain of precision on a small area unit basis is shown in Table 18.2. 18.2 The Approaces for Reporting Used by Annex I Parties 195

Table 18.2 Gain of precision in swiss forest inventory by use of remote sensing Enterprise Mean volume Standard error (%) Number of plots needed (per ha) for the same precision With Without Stratified Un-stratified stratification stratification sampling sampling 1 324 8 13 353 991 2 472 13 19 197 431 3 220 3 5 838 2,074 4 286 8 14 219 657 5 267 11 13 252 436 Source Schmid-Haas (1983)

The combined approach has the following overall benefits: • The stratification of field plots into remote sensing strata results in the reduction of error in the assessment of total volume. • The combined survey enables reporting of results for smaller area units than possible through surveys based on field sampling only due to reduction of error in estimation of area of small units. It may further be noted that all Parties listed in Annex I have in place management plan inventory systems since many decades, but only some parties have ongoing national forest inventory system. Therefore, the methodology will differ from party to party according to the inventory systems in use and other sources of available information in the country.

18.3 Potential Approaches for Non-Annex I Parties

FRA2010 reports that very few tropical countries have in place forest management plan inventories and still less national forest inventories to provide a basis for estimating area, volume, and C stock changes. If A is the area under change, V the volume per ha, and K the factor to convert growing stock into Carbon (C) terms, the expression for C for a reporting unit could be stated as C = K*A*V. Inventory on such lines have to be undertaken for each area under change. The Kyoto Protocol mentions the qualifying clause that a country reporting is to include only ‘‘direct human-induced land use change and forestry activities’’ limited to ‘‘afforestation, reforestation and deforestation’’ (ARD), as shown below in the form of matrix (Table 18.3): It is obvious that remote sensing could be effectively used to identify ‘‘Kyoto Lands’’ for undertaking further estimation of volume and conversation factors. The survey procedure can be generalized to undertake annual assessment of C-stock changes by keeping track of area, volume, and the conversion factor by year. 196 18 Forest Assessments for Climate Change Reporting

Table 18.3 Forestry and land-use change process Initial land-use Final land-use Forestry Non-forestry Forestry Reforestation Deforestation Non-forestry Afforestation Non-forestry changes

18.3.1 Estimating Forest Area Changes

FRA2010 in the framework of Global Remote Sensing Survey has implemented on a sampling basis assessment of state and changes in forest areas for 1990, 2000, and 2005. The sample survey for 2010 is under publication. The overall approach seems to provide a good framework for the tropical regions, under which detailed country level surveys can be organized. These surveys have to be on a wall-to-wall basis and use detailed classification of forest area and changes appropriate for C- Reporting. FRA1990 (see Chap. 4) has developed a classification system and method of assessment, which can be used for classifying forest areas and assessing growing stock. Its basis is interdependent interpretation procedure (by the same interpreter at the same time) which ensures consistency and reliability of results. Statistical procedures are available to transform the change matrix from one pair of dates to another pair of standard dates. For example, observed values for 1988 and 1998 could be transformed into a standard matrix referring to 1990 and 2000 by making assumptions or using auxiliary data about the continuation or alteration of the trend. Similar methods could be applied to remote sensing repeated on more than two occasions to estimate changes in the trend line. Definition scenarios that require data for intermediate years between 1990 and 2008 would very likely require three or more interdependent image dates.

18.3.2 Estimation of Growing Stock and Changes

The methods of estimating growing stock and conversion factors are the same as presented for Annex 1 Parties. An important issue is the choice between temporary and/or permanent sample plot techniques. Table 18.4 shows the results of a study conducted by the Swiss National Forest Inventory on the reduction of the sampling error of growing stock changes using permanent sample plots. In order to get the same precision as 60 permanent plots, 375 temporary plots would need to be measured. Moreover, the survey based only on temporary plots fails to provide information on growth and depletion. But, permanent plots cost more. Therefore, joint consideration of precision and cost may call for a mix of temporary and permanent plots, known as sampling with partial replacement. Forest carbon estimation usually involves conversion of forest inventory data on species composition, numbers of trees by size class, and volume into carbon stocks. 18.3 Potential Approaches for Non-Annex I Parties 197

Table 18.4 Precision of volume change estimation with permanent and temporary plots Observations Number of plots Mean Standard error Permanent Plots Temporary plots Date 1 60 194 m3/ha ± 10 m3/ha ± 10 m3/ha Date 2 60 212 m3/ha ± 11 m3/ha ± 11 m3/ha Stock change 60 18 m3/ha/year ± 6m3/ha/year ± 15 m3/ha/year Harvest 60 37 m3/ha/year ± 5m3/ha/year No information Total growth 60 55 m3/ha/year ± 3m3/ha/year No information Source Schmidt-Haas (1983) Swiss Forest Inventory: Twenty years of experience

For many areas of the world there are well-deﬁned relationships between these inventory variables and aboveground biomass, which converts readily to carbon. These relationships are not without error as discussed in the next section.

18.3.3 Estimation of Aboveground Biomass and Carbon

Currently, the majority of national forest carbon estimates are based on estimates of aboveground merchantable wood (merchantable wood deﬁnitions vary by countries). Estimates expressed in volume units are transformed into aboveground biomass or carbon (expressed in mass units) using biomass (or carbon) expansion factors (e.g. Brown 1997; Schroeder et al. 1997). Aboveground biomass estimates may serve as input data into regression models or more complicated models based on physiological processes to obtain belowground biomes and carbon assessments. Most historic forest inventories were developed to efﬁciently measure merchantable wood production and not biomass or carbon. Forest inventories generally provide estimates of main stem volume in a green condition. The use of biomass-expansion factors enables estimation of biomass and carbon from volume measurements. Brown (1997) gives biomass density estimates for tropical countries by forest type. The Kyoto Protocol reporting requirements may facilitate the change of forest inventory procedures from those oriented towards woody commodities into those oriented towards ecosystem services and, thus, towards forest biomass assessment. This should include inter alia extending forest inventory data collection to include small trees and nonmerchantable wood present in trees already included in forest inventory and use of GIS modeling (see Brown 1997).

18.3.4 Estimation of Belowground Carbon, Litter, and Woody Debris

Carbon in soil, litter, and coarse woody debris, may either be monitored or estimated by establishing statistical relationships with forest and site characteristics such as stand age, land-use history, and climate, or with soil maps and land cover 198 18 Forest Assessments for Climate Change Reporting data. Carbon stored in the soil has been found to be significant. Sombroek (1992) found that the majority of the terra firme soils in tropical rain forests of Brazil contain from 130 Mg/ha (for the very clayey soils) to 80 Mg/ha (for the sandy and/ or only moderately deep soils). The amount of carbon stored in soils may be of the same magnitude or larger, as much as 80 % of the total in cold northern forest ecosystems. Fearnside and Barbosa (1998) showed that about one-third of the belowground carbon is concentrated in the upper 30 cm depth unless the soil is shallow or imperfectly drained. This part of carbon, in the form of leaf litter and humus in topsoil, is affected by deforestation and burning and should be included in the accounting of carbon due to ARD activities. Most of the organic carbon in soil and that in roots is closely correlated to forest type and the system of forest management including fire and grazing in the area as well as disturbance history.

Soil carbon stocks of the Brazilian Amazon Basin Three soil types make up 60 % of the total land area of the Legal Amazon Basin. A comparison of C densities up to 1 m depth in 1162 soil proﬁles spread across the three soil types showed that C densities within soil types varied from 2.3 to 21.7 kg m-2; whereas the mean values of the C density of the three soil types varied from 8.5 to 10.5 kg m-2 only. The relatively low variability of C content of the soil types covering the largest areas makes the estimates of total basin soil stocks relatively insensitive to small changes in measured areas of different soil types. The C stored in the upper soil horizons represents the most active pool and the pool most sensitive to changes as a result of forest conversion to other land uses (Detwiller 1986). Of the 47 Pg C that were estimated in the top 100 cm for the Brazilian Amazon basin, 21 Pg C were contained in the top 20 cm. This large pool in the most active region of the soil indicates the potential for large changes in total Amazon Basin soil stocks as a result of changes in land use (Moraes et al. 1995)

The carbon content in forest soils is reported to be highly variable (Ellert and Bettany 1995; Ellert and Gregorich 1996) which limits the use of global or default data for local carbon balance estimations. Therefore it is likely that information systems in all countries will need enhancements to monitor soil carbon if this variable is included in the carbon accounting for ARD activities. Results of the Finnish National Forest Inventory indicate that the relative contribution of stumps and roots in Scots pine total dry mass is not lower than 14 %. Litter and coarse woody debris may comprise less than 10 % of carbon in forest ecosystems. Although relatively small, these pools are impacted strongly during forest operations and land-use changes. Carbon in soils, litter, and coarse woody debris may be easy but expensive to measure during ﬁeld inventories, and almost impossible to detect using remote sensing techniques. The use of regression equations enables estimation of belowground carbon pool from aboveground biomass measurements (Kurz et al. 1996). 18.4 Uncertainties in ARD Reporting 199

18.4 Uncertainties in ARD Reporting

18.4.1 Uncertainty Arising from Deﬁnitions of Forest

This will be illustrated with results of an FRA1990 study to highlight the relativity of concepts of forest and deforestation. The following three definition scenarios were used and three distinct forest cover and deforestation rates estimated. F-1 = closed forest F-2 = closed ? open forest ? 2/3 fragmented forest F-3 = closed ? open forest ? fragmented forest ? long fallow F-1 represents forest in the strictest sense, mostly dense, F-2 is a wider class aimed at matching the concept of forest used by FAO in the assessment based on country data. F-3 represents forest in its broadest sense, including all types and phases of degradation (but still with the connotation of forest. Average values of mean and standard error were computed from FRA1990 data over 117 sample locations spread over the entire tropical zone. There is an apparent decrease in deforestation rate with the broadening of forest definition (see Table 18.5). This is determined by the fact that the changes within the forest, when a broad forest concept is used, are classified as degradation and fragmentation and not as deforestation. It may be noted that for an in-depth analysis of forest area changes a broader definition of forest (viz. F-3) is preferable since it allows for finer differentiation among changes and a better description of their environmental impact.

18.4.2 Uncertainties Related to ‘‘Direct Human Induced’’ Activities

Article 3.3 of the Kyoto Protocol limits C accounting requirements to afforestation, reforestation, and deforestation resulting from ‘‘direct human-induced activities’’ (DHI). This DHI clause will require further clariﬁcation and deﬁnition prior to operationalizing the KP. The DHI clause establishes the need to determine, in every case in which a change in land use or forest cover has occurred, whether or

Table 18.5 Mean forest cover and deforestation rates for three definitions of forest (F-1, F-2, F- 3) with associated standard error Forest cover at year 1990 Rate of deforestation Definition of forest cover (%) (%) Mean Standard error Mean Standard error F-1 40.58 ±1.84 7.42 ±1.02 F-2 51.73 ±1.84 6.62 ±0.82 F-3 54.45 ±1.90 6.03 ±0.75 200 18 Forest Assessments for Climate Change Reporting not it resulted from direct human-induced activity. The difficulty in establishing this fact differs between definitional scenarios. The most restrictive definitional scenarios are those that only create Kyoto land through land cover change to or from forests but do not include reforestation that is part of forest management activities. In these scenarios the direct human involvement can be established where forests are replaced with a new form of land use, and where afforestation has created new forests. Difficulties arise, however, where change in forest cover results from human activities, such as exclusion of cattle grazing or fire suppression resulting in forest establishment, or where a natural disturbance, e.g., a severe large forest fire followed by regeneration failure, caused the change in forest cover. Clear guidelines will be required to establish whether or not direct human involvement has occurred in such cases. Less restrictive definitional scenarios are those in which forest management activities, such as harvest and reforestation create Kyoto land. In these cases, two events need to be assessed, the cause of the reduction in forest cover (e.g. fire, windthrow, harvest) and the cause of stand establishment. It must be clarified whether one or both events need to be human induced. For example, a wildfire caused by lightning in 2009 followed by reforestation through planting would create Kyoto land, for which the C stock change between 2008 and 2012 must be evaluated. In some definitional scenarios, the decision to plant the area deforested by natural disturbance will result in a C debit because the C stock in 2008 was greater than that in 2012. This would penalize the decision to plant even though this should be considered an activity that was beneficial for the atmosphere. In summary, definitional scenarios that create the largest amount of Kyoto land are also those that require the largest amount of spatially explicit data to establish whether direct human induced activities have occurred. In many cases it will be difficult to obtain such data, in particular for activities that occurred between 1990 and the present.

18.5 Validation and Veriﬁcation of Reports

With error or uncertainty assessment coupled with quality control procedures, validation could be unnecessary. Most forest inventories include quality control and sometimes report statistics on data quality. This procedure could be extended to carbon inventory. It would probably be necessary to establish institutional mechanism to verify that reported estimates were made using transparent methodology and included quality assurance and control procedures. The methodology could be monitored (or certified) by an independent authority. At a regional or local scale, networks of tower measurements using the eddy correlation technique are beginning to indicate regional trends in CO2 exchange with the atmosphere. These methods do not discriminate well between kinds of activities, but when coupled with a complete 18.5 Validation and Verification of Reports 201 carbon inventory (instead of the partial needs of the Kyoto Protocol), can be used to validate the inventory estimates. Use of remote sensing, as one among other data sources in establishing the land-use baseline would serve to ensure that results are verifiable. It also helps in detecting, delineating, and measuring area changes and provide some objective information for enquiry on the nature of change process whether land use and forestry activities are human-induced or not. In a similar way, use of technology like GPS helps in not only recording and delineating ARD activities, but also in relocating the area during the field survey phase. As all field coordinates are recorded on the basis of satellite observations, they are repeatable. The use of permanent sample plot techniques would also facilitate not only validation, but also re-measurements.

Recommended Further Reading

Fearnside PM, Barbosa RI (1998) Soil carbon changes from conversion of forest to pasture in Brazilian Amazonia. Forest Ecol Manage 108:147–166 (Elsevier, New York) Kurz WA, Beukema SJ, Apps MJ (1996) Estimation of root biomass and dynamics for the carbon budget model of the Canadian forest sector. Can J For Res 26(11):1973–1979 Moraes JL, Cerri CC, Melillo JM, Kicklighter D, Neill C, Skole DL, Steudler PA (1995) Soil carbon stocks of the Brazilian Amazon Basin. Soil Sci Soc Am J 39:244–247 Schmid-Haas P (1983) Swiss continuous forest inventory twenty years experience, renewable resource inventories for monitoring changes and trends. College of Forestry, Corvallis, USA Schroeder P, Brown S, Jiangming M, Birdsey R (1997) Biomass estimation for temperate broadleaf forests of the United States using inventory data. For Sci 43(3):424–434 Sombroek G (1992) Personal communication, Land and Water Development and Conservation Division, FAO Van Deusen PC (1997) Annual forest inventory statistical concepts with emphasis on multiple imputation. Can J For Res 27:379–384 Chapter 19 Global Forest Resources Assessments

19.1 The Rising Importance of Global Forests

Few themes have been a source of continuing global interest, concern, and con- troversy as forests and deforestation rate. During the early periods, the dominant concern was about the capacity of forests to meet the growing international demand of timber. The first Global Forest Resources Assessment, 1948, began with the words: ‘‘The whole world is suffering from shortage of wood’’. The Stockholm Conference, 1972, recognized forests as the largest, most complex, and self-perpetuating of all ecosystems; acknowledged their multiple roles as provider of environmental services and producer of economic benefits to humans; and expressed ‘‘alarm’’ at the large-scale ongoing deforestation (paragraph 51). UNCED 1992 spent significant time in discussing the draft of a ‘‘Forest Con- vention’’ on the lines of Climate Change and Biological Diversity, but could reach an agreement on Forestry Principles. Only in 2000, an ‘‘international arrangement on forests’’ has been concluded and given the form of an intergovernmental body: the United Nations Forum on Forests (UNFF) located at the UN HQ in New York. The following section presents the evolution of objective as well as the method for conducting global forest assessment since 1945 and lessons from the past successes and failures. The review is divided into three parts, viz., early assessments, 1948–1974; recent assessments, 1980–2010; and future scenario.

19.2 The Early Assessments: 1948–1968

The early assessments were based on questionnaires and included inquiries on forest resources (area and volume), annual growth, annual drain, and forest products. Four reports were published at 5-yearly intervals for the reference years 1948, 1953, 1958, and 1963. In WFI 1968 the questionnaire approach met with

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 203 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_19, Springer-Verlag Berlin Heidelberg 2013 204 19 Global Forest Resources Assessments great difficulties in developing countries and was abandoned. The negative experience with questionnaires is according to expectations. If one bears in mind that till 1970 not a single tropical country had carried out a nationwide inventory, replying to questionnaires in a reliable manner, must have been a nightmare, if not an impossible task to most countries. ‘‘Official statistics’’ were provided, which reflected the nominal rather than actual condition of forest resources. Such numbers aggregated to regional and global levels produced statistics of unknown reliability and meaning. Reidar Persson, who had worked at the FAO HQ for WFI 1968 as a Forestry Officer, continued to develop a better method as his doctoral thesis at the Royal College of Forestry in Sweden, in cooperation with FAO. In 1974, he made a ‘‘centralized assessment’’ through studies financed by the Royal College of For- estry, Sweden, and produced reports on Forests of the World, as a part of his thesis; and later a report on Forests of Africa. Both reports were on the basis of information collected by him during his earlier responsibility at FAO HQ, replies to questionnaire for OECD countries, replies to 1968 questionnaires received from African countries, existing reports, field visits, personal contacts, and guesstimates.

19.3 Recent Assessments: 1980–2010

19.3.1 FRA1980

The worldwide concern about the human environment led to creation of the United Nations Environmental Program (UNEP) in 1972. It was logical that UNEP and FAO collaborated to undertake the Tropical Forest Resources Assessment 1980 Project with the goal to estimate the rate of tropical deforestation using reliable method and institutional mechanism. It took about a total of 8 expert years to produce the most quoted statistics on the rate of tropical deforestation, viz., 11.3 million ha per year (FAO 1982). The Report made tremendous impact compared to all previous FAO assessments and led to creation of the Tropical Forestry Action Plan (TFAP) in 1984 to support sustainable management of tropical forests and to save them from wanton destruction.

19.3.2 FRA1990

In the wake of mounting international concern, FAO Rome started the Global Forest Resources Assessment 1990 after a gap of 20 years. The methodology of assessment was ﬁnalized at an Ad Hoc Meeting of Forest Inventory Experts held in 1987 at Kotka, Finland. FAO/ECE Geneva carried out the assessment for the industrialized (OECD) countries using a questionnaire approach. Considerable variation was observed by 19.3 Recent Assessments: 1980–2010 205 them in the quality and quantity of input data between countries and many gaps in the information supplied (FAO/EEC 1995). The real problem was the change estimation, which eluded a satisfactory solution for the industrialized regions as a whole, being based on a questionnaire approach. For the developing countries, utilizing the experience of FRA1980, FAO HQ continued the use of centralized assessment and made improvements in the methodology. The most reliable data available in the country was used as the baseline and standard values of forest cover for 1980 and 1990 were computed using a model and GIS database as presented in Chap. 3. An independent pan- tropical remote sensing survey (RS) was also organized as presented in Chap. 4. Results from the two approaches were found signiﬁcantly correlated (see Fig. 19.1: each dot represents a country).

19.3.3 FRA2000

FRA2000 beneﬁtted from three expert consultations and introduced a common forest cover deﬁnition worldwide. The scope of assessment was expanded to include; Biological diversity; Forest degradation; Climate change; and Non-wood Goods and Services. FAO/ECE Geneva made the assessment of the industrialized countries achieving remarkable improvements in reliability and comparability of results compared to 1990. Some innovative chapters were: carbon sequestration, carbon cycle and biomass, biological diversity, protective and socio-economic functions of forests, including discussion about the dependent indigenous people. Still, the

1990 forest cover by stratum - RS F2 vs. FORIS Sampling frame (land area) = 3 068.01 million ha RS F2 forest cover estimate = 1 528.13 million ha FORIS forest cover estimate = 1 522.56 million ha Correlation coefficient = 0.969 1000

RS F2

100

forest cover

million ha

1 1 10 100 1000 FORIS 1990 forest cover (million ha)

Fig. 19.1 Correlation between RS and FORIS forest cover estimates by country 206 19 Global Forest Resources Assessments forest area change assessment and quantiﬁcation of biological diversity, in other words ‘‘forest quality changes’’, posed a challenge to assessment. FAO HQ made the assessment for developing countries using a questionnaire approach, as was started by FAO in 1948. The consequences were also as expected, viz., large differences in the forest area and deforestation estimates between FRA1990 and FRA2000. The pan-tropical forest area of the latter at 1990 is higher by 237 million ha and deforestation rate down by 3.2 million ha per year (see Table 19.1)! The pan-tropical remote sensing surveys conducted by FRA1990 and FRA2000 using identical approach gave very consistent ﬁgures over 1980, 1990, and 2000 (see Table 19.2).

19.3.4 FRA2005 and FRA2010

The two assessments are closely related and were implemented like FRA2000 using a questionnaire approach and a common deﬁnition worldwide. The reporting interval was reduced to 5 years, the same practice as it was in the early periods of global assessments. The reporting framework follows the sustainable forest management concept, encompassing social, economic, and environmental dimensions of forest resources and includes 6 thematic elements and 21 related variables in 2005., and a further element added in 2010, viz., policy and institution. An effort was made to involve countries to a much higher degree in the assessment

Table 19.1 Forest areas and deforestation rates of FRA2000 and FRA1990 (in million ha) Zone/region FRA2000 FRA1990 Rates of tropical Forest areas Forest areas deforestation 2000 1990 1990 1980 FRA2000 FRA1990 Africa 634 687 528 568 53 41 Asia 324 349 311 350 25 39 South America 913 957 918 992 44 74 Tropics 1,871 1,993 1,756 1,910 122 154

Table 19.2 Comparative results of RS surveys by FRA1990 and FRA2000 (million ha) Region Forest cover in 1990 Deforestation rate 1980–1990 1990–2000 FRA1990 FRA2000 FRA1990 FRA2000 Africa 556 542 28 23 Lat America 824 824 54 44 Asia 290 297 26 25 Pan-tropical 1,671 1,663 107 92 Sources Forestry paper 130 and 140 19.3 Recent Assessments: 1980–2010 207 process than done in previous assessments, leading to a higher response rate and quality control of information at the national level. On the international side, effort was made to forge Inter-Agency collaboration, to reduce the reporting load on countries. Definitely, very positive steps! No doubt, if the questionnaire approach is to be continued, country capacity has to be enhanced and south–south cooperation established firmly to achieve harmonization of the country figures. FRA2010 also made a remote sensing survey of world forests using a systematic Sample of Landsat satellite imagery for the years 1990, 2000 and 2005 located at the intersection of each degree of longitude and latitude. The area surveyed at each sample site was 10 km 9 10 km, providing a sampling intensity of about 1 % of the global land surface (FAO/JRC 2012).

19.3.5 Some Comments on Global FRA Figures

An in-depth study of recent FRA reports by the author brings to mind a number of questions about figures reported by FAO FRA on many topics for many countries and regions. The following comments are not on an individual basis, but compiled from opinions of experts involved over a long period of time with forest inventory in general and global forest assessments in particular. • There is an international pressure that FRA shall collect information about more and more themes/variables. But many figures for simple things like forest area and plantations are evidently wrong. What is then the quality of information about issues that so far have not been collected in forest inventories? Often unqualified guesses. FAO should perhaps make clear, the quality of the information they are presenting. There are continuous changes in reported figures which makes it is difficult to discuss changes over time. • For Ethiopia, for example, there are very conflicting figures between FRA2000 and FRA2005 reports. FRA2010 reports 6 million ha of plantations in Sudan. Behind this figure lies a guess that 20 % of the so-called forest area is planted. What is then the quality about difficult information like primary forests, area affected by insects/diseases, areas designated for protection, value of NWFP, progress towards SFM etc.? • Instead of reporting very unreliable figures it would be more valuable to have a detailed discussion about the quality of the reported information. And adjust, based on studies, doubtful national figures in a sub-regional context, so that the quality of the overall information is improved. • Really credible regional and sub-regional figures are what the interested parties in the international debate may benefit from. The new RS inventory may be one way of showing trends—at least at sub-regional levels. The above observations are important to think over in relation to future global forest assessments. 208 19 Global Forest Resources Assessments

19.3.6 Questionnaire Approach

The relevance of the questionnaire approach needs to be carefully evaluated. It seems very unlikely that the method, in view of weak country capacity, can furnish consistent time series information: the collected data would lack a statistical basis (see Table 19.3). The method has been found unsuitable for making time series analysis in the industrialized countries too.

19.3.7 Remote Sensing Survey Method

The three rounds of remote sensing surveys (RSS), conducted respectively by FRA1990, FRA2000 and FRA2010, demonstrate unquestionably the relevance of the survey technique for global forest assessments and suggest the need to design them on the lines of continuous forest inventory, as was the vision during the ﬁrst application. Another interesting fact emerging from Table 19.3 is a remarkable closeness of RSS and model estimates: for the tropics and each region. This shows that the model approach could serve as an effective alternative to the questionnaire method. This would call for country capacity building to organize the existing information.

19.4 Future Global Forest Assessments

19.4.1 Problem Formulation

A major change in the global thinking is a holistic view of the world, as pointed out in Chap. 16. The biosphere is recognized as a geographically bound and ecological well-deﬁned entity, for which an accurate accounting of natural and cultural processes over time is accepted essential in order to monitor, understand,

Table 19.3 Tropical forest area reported by successive assessments Zone/region Forest area in 1990 (million ha) Questionnaire approach Model approach Remote sensing FRA2010 FRA2000 FRA1990 FRA2010 Pan-tropics 2,076 1,993 1,792 1,730 Africa 731 687 530 520 Asia 356 349 338 350 South America 989 957 924 860 19.4 Future Global Forest Assessments 209 and regulate to the extent possible the ongoing global changes. In this context, the demand for reliable global forest resources information has rapidly grown over the last 40 years, not only in terms of thematic details, but also regarding reliability, increased spatial and temporal resolutions. The auto-evaluation made in 2010 highlights that credibility of assessment deserves the highest consideration. The global forest assessments, like all serious inventory undertakings, must have precisely deﬁned objectives, make a problem-oriented information need analysis, use appropriate statistical method(s)/models, and institutional mechanism.

19.4.2 The Changing Objectives and Information Needs

In view of the growing demand for reliable and consistent global forest-related information and the need for enhanced reliability of forest area and area change information, FAO needs to work out a two-fold strategy: (i) Identify variables that satisfy the general information needs of a wide variety of users without specifying the precision goals; and (ii) Identify some key variables needed for a rigorous analysis at the national, sub-regional, regional, and global levels with speciﬁed precision targets. In particular, four variables related to extent offorest, growing stock, and the associated rates of change seem most interesting variables, of key importance, and major interest to national and global concerns. This group of information needs to be estimated by geographic (viz. by country) and ecological regions. There is need for a global database at the sub-national, national, regional, and global levels for the key variables, keeping in view the need for a system analysis (including geographic, ecological, economic, and social factors) as a part of complex ecological, social, and economic change model. During the IPCC consultations on the Kyoto Protocol, the inadequacy of global forest resources information including deﬁnitions, reliability of forest change information (in time and space), missing information like Soil Carbon, etc., was repeatedly pointed out. The FAO/ECE Team of Specialists in their 2000 report have made a good beginning in relation to Carbon Sequestration and Forest balance studies. Such thinking needs to be further developed in the tropical regions too and expanded to cover all major global forest-related issues including biological diversity; land and forest degradation, indigenous people, etc. There is urgency for suitable action on the subject as there is wide variation in estimation of total forest area of the tropics and changes that are taking place.

19.4.3 Method of Data Collection and Analysis

The future approach to global assessment based on country data as well as remote sensing sample survey needs to be carefully reviewed. The science community and policy makers (and Media), all want a consistent time series information at least 210 19 Global Forest Resources Assessments since 1980; and separating changes relating to reforestation, afforestation, and deforestation. All reliable country data, collected over a long time preferably at a sub-national level (to make use of partial inventories of countries!), need to be tapped along with other ancillary information, to build the time series and to enhance the reliability of change assessment. Specific action is required in order to: (i) Strengthen FORIS database, as recommended above, and undertake research on methods of assessment based on country data for specified variables; build capacity of countries and strengthen regional institutions to improve the quality of country-wide field assessments, and remote sensing surveys, by using interdependent assessment. The latter will not only enhance the reliability of change assessment, but also separate positive from negative changes. (ii) Integrate successive rounds of remote sensing-based tropical sample surveys into a single continuous forest inventory design. This improvement will provide a reliable time series of information on forest over the tropics and production of composite change matrices. Expansion of the Remote Sensing Sample Survey to the entire world is a very positive development in FRA2010. (iii) Improve the coverage and quality of reporting for selected global topics like Deforestation Rate, Carbon Sequestration, Biological Diversity, and Forest Extent by political as well as ecological zones. The geo-referenced data on forest cover, ecological zone are already available. These need to be updated and integrated with statistical data at a sub-national level especially for large countries. Most of the above ideas have been discussed at the Expert Consultations held earlier. The FAO/ECE Team of Specialists has also exchanged views on the above points. The final choice of key variables and the method (for monitoring the quality improvement in global assessments) will depend on the political, technical, as well as financial considerations.

19.4.4 Institutional Strategy

In view of very high expectations (in terms of number of variables and reliability of selected variables) from Global Forest Resources Assessments combined with interest in more frequent reporting (5 years and some want every 2 years), the institutional aspects of the process assume great signiﬁcance. The 60-year experience with FAO/ECE Geneva in conducting assessment of the industrialized regions demonstrates the value of regional groups, which discuss not only regional assessments but also exchange experience about national assessments and ﬁnd technical solutions. FAO/ECE Cooperation in forest assessment has grown over the years and has become ‘‘a continuous multi-level and multi-functional complex process’’ different from a periodic gathering for ‘‘one time collection and compilation of the forest resources information’’ (UNECE 2000). The results are 19.4 Future Global Forest Assessments 211 very visible, if one just compares the quality and contents of the reports of FAO/ ECE for 1980, 1990, and 2000 reference dates, not to say about its continuity since 1948!

Recommended Further Reading

Czaplewski R (1996) Evaluation of FAO remote sensing survey 1990. (unpublished report), FAO Rome FAO (1982) Tropical forest resources assessment. FAO, Rome FAO, Rome FAO (1993) Forest resources assessment 1990: tropical forest resources, Forestry Paper 112 Global Synthesis, FAO, Rome FAO (1995) Forest resources assessment 1990: global synthesis, Forestry Paper 124, FAO, Rome FAO (1996) Forest resources assessment 1990: survey of tropical forest cover and study of change processes, Forestry Paper 130 FAO (1998) Strengthening of country capacity for planning, assessment and periodic evaluation of forests: ﬁndings and recommendations, FAO/SIDA Project, Rome FAO-ECE 200: Forest Resources of Europe, CIS, North America, Australia, Japan and New Zealand, United Nations, New York and Geneva FAO (2001) Global forest resources assessment 2000: main report, FAO Rome FAO (2006) Expert consultation on global forest resources assessment TOWARDS FRA 2010. Kotka, Finland. 12–16 June 2006 (on FAO web-site) FAO and JRC (2012) Global forest land-use change 1990–2005. In: Lindquist EJ,D’Annunzio R, Gerrand A, MacDicken K, Achard FR, Beuchle A, Brink HD, EvaP, Mayaux J, San-Miguel- Ayanz, Stibig H-J. FAO Forestry Paper No. 169. Food andAgriculture Organization of the United Nations and European Commission Joint ResearchCentre. FAO, Rome. Ilvessalo L, Jalava M (1930) Forest resources of the world, communications of Ex Instituto Questionum Forestalium Finlandiae Editae 16, Helsinki Persson R (1974) World forest resource, review of world forest resource in early 1970s. Department of Forest Survey, Royal College of Forestry, Sweden UNEP (1972) United Nations conference on human environment, environmental aspects of natural resources Management Zon R, Sparhawk WN (1923) Forest resources of the world. McGraw-Hill Book Company, Inc., New York Chapter 20 International Support to Country Capacity Building

20.1 Agenda 21 Recommendations

Agenda 21 sets very high expectations regarding international support to country capacity building in forest assessment. This is evident from the following statement made in paragraph 11.33: ‘‘The international community should extend to the Governments concerned necessary technical and financial support for implementing this programme area, including consideration of the following activities: (a) Establishing conceptual framework and formulating acceptable criteria, norms and definitions for systematic observations and assessment of forest resources; (b) Establishing and strengthening national institutional coordination mechanisms for forest assessment and systematic observation activities; (c) Strengthening existing regional and global networks for the exchange of relevant information; (d) Strengthening the capacity and ability and improving the performance of existing international organizations, such as the Consultative Group on International Agricultural Research (CGIAR), FAO, ITTO, UNEP, UNESCO and UNIDO, to provide technical support and guidance in this programme area’’. This chapter will identify areas for international support to accelerate the capacity building process and the need for strengthening organizations’ ability and performance to realize the Agenda 21 objectives. This subject will be discussed in the following four sections: • development and dissemination of knowledge for assessments and systematic observations of forests relevant to tropics keeping into account multiple functions of forests; • provision of technical and financial assistance to countries for capacity building with a long-term perspective;

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 213 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1_20, Springer-Verlag Berlin Heidelberg 2013 214 20 International Support to Country Capacity Building

• regional/global networking of institutions for exchange of forest inventory information; • strengthening international organizations’ own abilities and capacities for effective participation in the process.

20.2 Development and Dissemination of Knowledge for Tropical Forest Assessments

There is hardly any R & D institution in the tropics to undertake basic and applied research on different sub-disciplines associated with forest inventory such as sample survey design, growth and yield studies, accessibility and cost-studies, remote sensing and data processing, appropriate to forests and socio-economic conditions of the regions. To ﬁll this important knowledge gap, the FAO Forestry Department since its inception has been allocating regular funds to support applied research activities such as preparation of ﬁeld manuals and guidelines, remote sensing methods, growth and yield studies, etc. Though funds are small, the impact is large, as professionals of the region and interested international experts are involved in the process. Of 140 Forestry Papers (which are in fact comprehensive reports) published by the Forestry Department since 1980, 16 papers related to Forest Resources Assessment. The decadal distribution of the papers was as follows: 1980–90: 7 papers; 1990–2000: 6 papers; 2000–2010: 3 papers. A main reason for less Forestry Papers during 2000–2010 was the competing claims of HQ resources for global forest resources assessments.

Chart 20.1: Main Constraints to Development & Transfer of Knowledge • insufﬁcient staff to cover the whole range of technical responsibilities at HQ; • for the same reason, insufﬁcient time available for technical backstopping and learning from FAO Field Projects; • outsourcing of technical works to institutions in developing countries and effective backstopping takes time. However, it is important for promoting country capacity. Source: FAO (1994)

Even though tropical forests have gained worldwide importance, there is less funding to teaching and research on quantitative aspects of tropical forests compared to say 50 years back, with the important exception of remote sensing, which has caught global fascination in the context of climate change studies. In the 1960s, when the author graduated from the World Forestry Institute, Reinbek, Germany, World Forestry formed an important topic for post-graduate Research and Education in the three important universities of the industrialized regions. Within the developing regions, there is hardly any forestry institute today, which is conducting applied or theoretical research or imparting advanced forestry education supported with long-term observation plots. In South and South East Asia, the regular measurements of permanent sample plots have been abandoned since long. 20.2 Development and Dissemination of Knowledge for Tropical Forest Assessments 215

This situation needs to be reversed, if tropical forests are to be protected and managed on a sustainable basis to produce local and global goods and environmental services. Concerted action is called for by the concerned international organizations mentioned in Agenda 21 (CGIAR, FAO, ITTO, UNEP, UNESCO, and UNIDO) and donors support to institutions in the tropical regions, to develop adequate global competence to face the challenge of scientiﬁc management of tropical forests.

20.3 Review of International Assistance to Country Capacity Building

This subject is very important in the context of FAO and Donors’ support to countries, intended to serve the capacity building goals. The following evaluation is based on a review of 40 FAO ﬁeld projects implemented during: 1980–1994 and 2000–2010.

20.3.1 Technical Assistance During 1980–1994

In 1994, following UNCED 1992, FAO organized an auto-evaluation of technical assistance programme from the perspective of capacity development. The ﬁndings were as under (Singh and Janz 1994): • ‘‘Most of projects implemented during the period, made major impacts in generating information but only a few contributed to building of long term capacity and none to establishment of an institutional frame for effective use of forest resources information. In the cases, where a lasting impact was produced, the key to success was continued FAO involvement with the concerned institutions, spread over a long period of time. In all cases, a necessary condition was Government commitment to continue development of the institution after ter- mination of FAO Project. • Another important limitation is due to fact that FAO technical assistance is provided on request only and does not form an integral part of a long term strategy for institution and human resources development and generating self- reliance as envisaged in UNCED Agenda 21. As a result though technical assistance has been provided to member countries for over 50 years, the existing capacity, as discussed earlier, is rather weak. Both countries and FAO have to work together and agree on general criteria to be applied in the selection, formulation and implementation of short term projects within the framework of national policies, strategies and plans’’. • The quality of FAO technical assistance to countries is being impaired by another reason also, viz. a reduction in the use of long term experts and an increase in the use of short term consultants. 216 20 International Support to Country Capacity Building

• Some possibility of south–south cooperation supported by FAO could provide access to very useful information and ensure that appropriate technology is supported by FAO in the countries’’.

20.3.2 Technical Assistance During 2000–2010

The NFMA (National Forest Assessment and Monitoring) Programme was launched in 2001 by the FAO Forestry Department in the support of FRA2010. Pro- jects have been already implemented in several countries including Bangladesh, Cameroon, Costa Rica, Guatemala, Honduras, Lebanon, Nicaragua, the Philip- pines. There are ongoing projects in an additional 13 countries and 33 countries are waitlisted pending availability of resources. The NFMA programme has innovative characteristics and includes: • Piloting of national forest/land inventory methods that are based on a low intensity; • sampling design able to produce national relevant information in a relatively short time and at a relative local cost while maintaining statistical soundness; • strengthening the capacity of national teams and institutions to undertake inventories and analyze data; and • disseminating and communicating knowledge at global and country project levels. The programme has mobilized resources from a wide range of sources and benefitted from advice and guidance from the FAO NFMA Advisory Panel, Expert Consultations, and Technical Meetings (2003, 2005, 2006, 2008, and 2009). The programme has benefited from SIDA extra-budgetary resources since its inception in 2001. During the period from 2001 to 2008 the total NFMA budget was US$ 16,252,000, of which 24 % was from the FAO Regular Programme, 49 % from donors, and 27 % from recipient countries. For the period 2009–2013, known additional extra-budgetary commitments have substantially increased to approximately US$ 243,539,000 largely owing to sizeable extra-budgetary support from UN-REDD (MRV), GEF (Brazil), expected funding from AfDB’s Congo Basin Forest Fund (develop REDD ? MRV in 10 countries) and Finland’s support to FAO’s forest assessment and monitoring. In 2010, FAO commissioned an impact evaluation of NFMA Programme and country projects with the following objectives: (i) assess the utility and impact of NFMA to countries where it has been implemented; and, (ii) based on lessons learned, identify obstacles and opportunities for the future (FAO 2010). What NFMA did Achieve: The FAO Review shows that projects have helped countries and international agencies alike by generating national level data on ‘a vast’ number of variables. These included variables to calculate growing stock, biomass, and carbon (stem diameter, tree height, deadwood biomass, soil carbon, and litter); vegetation type (including species composition and naturalness); extent 20.3 Review of International Assistance to Country Capacity Building 217 of land use and land cover, biodiversity status, land-use history, human distur- bances, management practices, and conservation status. What NFMA Did Not Achieve and Why? According to FAO Evaluation (2010), NFMA projects could not fulfill the following very important capacity building tasks on account of time and resource constraints: • Country capacity in data analysis was not strengthened enough. FAO staff and short-term consultants have done most of the analysis; • Accessibility to the NFMA data is very limited and the level of integration of NFMA data with other national level databases has been generally weak; • Dissemination of country project findings has been generally limited and the level of awareness raising and use of the country assessment data remains low in the countries assessed; • The format of final reports, a combination of both analysis and project management report, has not been user friendly for senior staff and key decision- makers; • Stakeholders have not always used the NFMA generated information for informing design of strategies, policies, and programmes to strengthen sustainable forest, and land-use management. • After the official conclusion of NFMA assessments in the countries (Cameroon, Philippines, Guatemala and Nicaragua except Zambia), the HQ staff could not undertake country missions to provide guidance and support in the uptake and use of assessment data and further data processing after the projects ended. This lack of post assessment support can be attributed to the demand on the time of NFMA FAO HQ staff for global programme coordination activities, backstopping of ongoing country projects, and work on country pipeline. • It was also observed that members of the core NFMA team in all countries have dispersed to other institutions or are posted at different duty stations within the lead forestry institution. It is nevertheless expected that a number of the original field team members can be reassembled to participate in a subsequent field assessment. Producing and disseminating technical reports without a concerted plan to provide follow-up with policy-makers is unlikely to result in policy or programme changes in the countries. The above reviews show that the classical technical assistance paradigm is expected not to be very successful from a CCB perspective. Except for information generated, little is left to continue project activities. A strategic reorientation is, therefore, called for to make project capacities useful for policy making and planning as foreseen in the Agenda 21 formulation. It is not merely a question of time and resources constraints, as pointed out in the evaluation report, but lack of long-term gain to countries from an international technical assistance and to some extent lack of proper appreciation of CCB needs by countries and donors. 218 20 International Support to Country Capacity Building

20.4 Lessons of Interregional Country Capacity Building Project, 1995–1998

Following UNCED 1992, FAO with SIDA Support launched an interregional project ‘‘Strengthening of Country Capacity for Planning, Assessment and Periodic Evaluation of Forests’’ with a 3-year term, 1995–1998. The project objective was to develop effective strategies for implementing the Agenda 21 Recommendations on FRA in a cost-effective manner. The thinking was that country capacity building in the Agenda 21 sense is a different process than implied in formulation of traditional projects of FAO and Donors. It was designed to answer the question: How could CAPACITY Development be achieved rapidly and maintained more permanently in developing countries than is the case presently? Chart 20.2 presents a synoptic view of the project objectives and planned outputs. A synergy with the FRA process was an important consideration in view of interdependence in the two activities.

CHART 20.2: Country Capacity Building Framework Development Objective To Institutionalize Forest Resource Assessment for Sustainable Development in the developing Countries Immediate Objective and Outputs

1. Information Gathering Provide current 2. Capacity Building Strengthen national and objective information for country policy and international institutions’ support makers, the research community and network. international needs. Outputs Outputs

1.1 Databases and database management 2.1 Institutions system 1.2 Processed results 2.2 Skills 1.3 Studies and models 2.3 Standards 1.4 Reports 2.4 Methodologies

Within the project framework a Programme Development and Coordination Unit was established at FAO HQ to advise on capacity building activities; and coordinate implementation support to FRA2000. The project started smoothly as all project staff had earlier worked as members of the FRA1990 Team and perfectly understood implications of the objectives, approach, and outputs of the project: its dual nature involving information gathering and capacity building. The project ﬁndings and recommendations are contained in the Report: FO: GCP/INT/621/SWE. The holistic architecture of the project is pictorially shown in Chart 20.3, highlighting the thrust areas at the national, regional, and global levels. 20.4 Lessons of Interregional Country Capacity Building Project, 1995–1998 219

Chart 20.3 Institutional Networks for Capacity Building in Forest Resources Assessments

Specialized Institutions in Project Headquarter Rome developing and developed countries

Global Level Activities

Regional level Activities

Region Asia Region Africa Region Latin America

TCDC Lead Center TCDC Lead Center TCDC Lead Center

Regional Level Activities

National Level Activities

National Forest Resources Assessment (support to individual countries)

20.4.1 National Forest Resources Assessments (at the Base of the Chart)

The project was actively involved in providing technical back-stopping to ongoing forest inventory projects, where it promoted a holistic view of capacity building, integrating forest resources assessments with planning and evaluation. Country level training programmes were organized, covering essential topics such as: problem formulation, capacity building to identify main developmental and environmental problems, assess information needs; plan institutional development, and establishing the necessary inter-sectorial and programme linkages; technical support was provided in designing problem-oriented forest inventories, use of Geographic Information Systems (GISs), remote sensing, computer technology, and modeling techniques; and establishment of deﬁnitions, standards, norms, and methodology for data collection and processing. These aspects of country capacity building were developed and demonstrated by the Project through a model country project in Andhra Pradesh India (see Sect. 5.6). This project has continued to grow and has now become a centre of excellence in the country. Several papers and 220 20 International Support to Country Capacity Building documents were published on the technical aspects of forest resources assessment and how databases are organized and used. Establishment of national analytical units was an important topic discussed and supported.

20.4.2 Regional/International Network and Activities (Middle of the Chart)

Capacity building, particularly on a global tropical level, is a complex task. With limited resources at its disposal, the project developed and tested cost-effective CCB through ‘‘regional cooperation’’ based on the ‘‘Technical Cooperation among Developing Countries’’ (TCDC) Concept. Some concrete tasks, assigned to the Lead Centers were the following: • Organize regional workshops, seminars, and training courses; • organize Forest Resources Information Systems (FORIS, GIS, Remote Sensing); • assist countries in the region to compile and organize forest resources data, and gather new data (i.e. forest changes using remote sensing, biomass, biodiversity), • facilitate technology transfer South–South and North–South. The staff already in place in the lead centers got further trained through workshops in the various aspects of FRA (FORIS, GIS, RS, biomass/biodiversity assessment). Limited equipments (hardware, software) were also provided to them. The experience showed that regional centers generally needed limited external funding, which could come from HQ/Regional Projects or even country sources. Why so much emphasis on Regional Cooperation? A main reason was that countries within a sub-region learnt by seeing and talking: how neighboring countries are solving their problems, which they also face. The relative cost per course was very small as only travel within a sub-region was usually involved and local facilities were often used for organizing the courses. The capacity of host- country also improved in the process. Agenda 21 mentions ‘‘Strengthening of regional and global networks for the exchange of relevant information’’ as a key capacity building task. This is correct. During the auto-evaluation of FRA2010, when national correspondents were asked ‘‘what represents the most signiﬁcant added value to participate in the FRA, most expressed the main gain to be networking and collaboration’’, rather than knowledge about global forests. During the FRA2010 Kotka Meeting, the countries of Central and South America recognized that a continent-wide monitoring system had critical importance and resolved to include this activity in their Regional Cooperation Programme, in addition to global assessment, for which they were invited. A GOFC-GOLD Study (2010) highlights that regional cooperation and South– South capacity development cooperation had ‘‘great potential for some of the non- 20.4 Lessons of Interregional Country Capacity Building Project, 1995–1998 221

Annex I countries’’, who may beneﬁt from the existing capacities of countries like India, Brazil, and Mexico, which have long experience in conducting forest inventory and monitoring. Some of these processes have already started and should be further strengthened and encouraged by the international community. Some other promising initiatives were present in Central Africa, Latin America, South Paciﬁc, Mekong river countries, and other Central Asian and South-East Asian Regions. It is obvious that the present design of networking in the framework of global forest assessments needs to be re-thought and redesigned to serve dual purposes of country capacity building and global forest assessments.

20.4.3 The Global Forest Resources Assessments (Top of the Chart)

The FAO/ECE Geneva Statistical Party was an inspiration to the project for undertaking global assessments based on regional cooperation. Noteworthy is the fact that FAO/ECE Geneva continued to carry out regional forest assessment, when global assessments were discontinued in 1968 by FAO HQ. A strategy to produce similar institutional synergy in the assessment of tropical countries was experimented and found feasible. This considerably reduced the pressure to global FRA at FAO HQ and contributed to integrate national and global forest assessments. Such arrangements were found equally interesting for realizing national obligations speciﬁed in the UNCED Conventions and other internationally agreed processes on forests and also strengthen country capacities to make continuing assessments and undertake sustainable forest management as was elaborated in the earlier chapter. Twinning arrangements were also promoted between ‘‘lead centers’’ and ‘‘specialized forest inventory institutions/universities’’ in developed and developing countries (See the Box in the top right Corner of Chart 20.3). Donors also deputed their experts to learn and contribute to training courses. In some cases, it led to direct bilateral cooperation too.

20.5 Concluding Observations

‘‘The history of international assistance to developing countries shows that donors and partner countries alike have tended to look at capacity development as mainly a technical process, or as a transfer of knowledge or models from North to South. It explains how international agencies and donors have often failed to recognize the critical importance of country ownership and leadership of projects, and how they underestimated the importance of the broader political context within which capacity development efforts take place. 222 20 International Support to Country Capacity Building

It also offers invaluable guidance about how to think systematically through the capacity development challenge. The evidence suggests that what is necessary is a fundamental change in development practice, including focusing on capacity as an endogenous process, agreeing at country level on capacity objectives and monitoring outcomes from the perspective of the beneﬁciaries. Such changes could have a substantial impact on development outcomes.’’ The above quote is taken from the OECD Report (2006): ‘‘The Challenge of Capacity Development: Working towards Good Practice’’. The ﬁndings are based on a review of over 40 country projects and are almost identical to Evaluation of FAO-assisted forestry projects presented earlier in this chapter. It shows that the problem associated with capacity building is not limited to the forestry sector or to design of FAO Technical assistance, but is deep rooted in the current paradigm of technical assistance. A triangular understanding between the Country, FAO, and Donors will go a long way to ameliorate the problem.

Recommended Further Reading

FAO (1994) A Review of FAO’s Achievements in Forest Resources Assessment and a Strategy for Future Development, An Auto-Review by K.D. SINGH and K. JANZ, FAO, Rome FAO (1998) Strengthening Country Capacity for Planning, Assessment and Periodic Evaluations of Forests: Support to Programme Development and Coordination Unit: Project Findings and Recommendations. FAO, Rome FAO (2008) NFMA approach and process: an analysis of cost and time. Background paper prepared for the national forest monitoring and assessment Expert Consultation, Rome 26-28 Nov 2008 FAO (2010) The Global Forest Resources Assessment Auto-Evaluation, prepared by Judita Jancovic, Forest Resources Assessment Working Paper 178. FAO, Rome FAO (2011) Impact Evaluation of the FAO National Forest, Monitoring and Assessment Programme Ofﬁce of Evaluation GOFC-GOLD (2009) An assessment of national forest monitoring capabilities in tropical non- Annex I countries: Recommendations for capacity building, Friedrich Schiller University Jena

On Web

UNCED Agenda 21 Index

B sustainable non-timber forest Biological diversity, 186–188 management, 45 assessing risk of biological diversity loss, the role of international/regional 188 cooperation, 47 modeling of biological diversity richness Climate change, 191–201 loss, 187 Kyoto protocol, 191 problem formulation, 186 potential approaches for non-annex I risk of species richness loss, 189 parties, 195 species area relation, 188 estimating forest area changes, 196 estimation of aboveground biomass and carbon, 197 C estimation of belowground carbon, Capacity building, international strategy, litter, and woody debris, 197 213–222 estimation of growing stock and chan- agenda 21 recommendations, 213 ges, 196 concluding observations, 221 reporting requirements for annex I parties, development and dissemination 192 of knowledge, 214 uncertainties in ARD reporting, 199 international assistance, 215 uncertainty arising from definitions of period 1980-1994, 215 forest, 199 period 2000-2010, 216 uncertainties related to definition of lessons of interregional country capacity ‘‘Direct Human Induced’’ activities, building project, 218 199 regional/international networking, 220 validation and verification of of reports, support to global forest, 221 200 support to national forest assessments, 219 Capacity building, national strategy, 41–48 D areas for capacity development, 42 Data processing, 79–88 integration of planning with forest case study of FAO forest inventory data inventory, 43 processing system (FIDAPS), 86 long term forestry planning, 43 database storage and archival routines, 84 medium and short term forestry data processing steps, 82 planning, 44 development of volume functions, 82

K. D. Singh, Capacity Building for the Planning, Assessment, and Systematic 223 Observations of Forests, Environmental Science and Engineering, DOI: 10.1007/978-3-642-32292-1, Ó Springer-Verlag Berlin Heidelberg 2013 224 Index

D (cont.) the changing demand for forest inventory, estimation of means and 14 standard errors, 83 Forest inventories, the changing emphasis, final tabulation, 84 3–10 generalized versus tailor made EDP environmental issues, 3 systems, 85 global forest resources assessments, 5 manual and computer assisted editing, 80 strengthening country capacity output and input specifications, 86 in the tropics, 6 plot level summaries for error Stockholm conference, 4 calculation, 82 roles of data processing, 79 tree volume estimation, 82 G Global forest assessments, 203–211 future assessments, 208 E institutional strategy, 210 Ecological zoning, 181–190 method of data collection and analysis, approach for ecological zoning, 182 209 choice of parameters, 182 problem formulation, 208 classification and mapping of EFZ, 182 the changing objectives and EFZ map and the database, 184 information needs, 209 forest ecosystems report, 184 recent assessments, 204 validation, 184 FRA1980 need for ecological zoning, 181 FRA1990 Environmental functions of forests, 171–180 FRA2000 definition of major components, 173 FRA2005 ecosystem dynamics, 175 FRA2010 ecosystem variables and change model, 176 the early assessments, 203 problem formulation, 171 Growth and yield studies, 131–144 studies using ecosystem approach, 178 applications in forest management, Existing inventory information, 23–29 142 FAO forest resources information system hill dipterocarp forests (FORIS), 27 of Malaysia, 143 forest dynamics plots (FDP), 26 mixed tropical forests, Indonesia, 143 role of existing information, 23 growth and yield of mixed tropical forests and trees, 138 growth and yield of single trees, 138 F increment borings, 140 Field data collection, 73–78 methods of research, 138 check-crew work, 77 stump and stem analysis, 139 classification of data sources, 73 growth and yield of temperate forests, 134 computer-assisted editing, 77 current status of knowledge, 134 data archival routines, 77 plantations development, 132 field manual and field forms, 74 methods of study, 132 field plan and logistics, 74 the current state of knowledge, 133 special studies, 76 special conditions in the tropics, 131 Forest inventories, formulation of objectives, 11–20 assessment of environmental functions of L forests, 18 Land evaluation, 159–167 identification of information needs, 16 applications in forestry, 163 linking forest inventory with the problem, district level, 166 13 national level, 164 problem-oriented classification of forest description of techniques, 159 inventories, 15 purpose of land evaluation, 159 Index 225

P the land cover and forest changes, 94 Potential productivity, 147–158 trees outside forests, 95 climatic indices, 148 survey planning, 90 current availability of climatic data, 151 special studies, 93 day length, 153 survey design, 91 growing season, 152 precipitation, 153 relative humidity, 152 S temperature, 152 Spatial variation patterns in the tropics, need for further research, 153 105–113 annual growth indices, 154 macro-variation patterns, 106 hardness index, 155 meso-variation patterns, 107 productivity indices, 153 micro-variation patterns, 109 need for south-south cooperation, 157 similarities across the continents, 105 paterson’s indices, 148 Special studies, 63–72 validation of indices, 149 development of biomass functions, 67 suggestions for a new climatic development of volume equations, 64 index, 156 planning of special studies, 63 the need for estimation, 147 scope of special studies, 63 weck indices, 149 survey of non-wood forest products, 69 validation of indices, 150 Statistical planning, 51–62 a planning check list, 61 distance method, 60 R main steps in the sample survey design, 52 Rain forest loss and change, 111 purpose of statistical planning, 51 Remote sensing applications, 115–129 role of forest statistician, 52 applications of aerial photographs, 119 some commonly used designs, 54 complete photo interpretation, 120 survey of trees outside forests, 56 point photo interpretation, 123 inventory of village forests in Bangla- cost-effectiveness of remote sensing, 126 desh, 58 FSI state of forest report, 127 national inventory of TOF in india, 60 accuracy assessment, 128 assessment method, 127 lidar potentials, 118 T recent developments, 115 Technology applications, 31–39 Report writing, 89 classification of emerging technologies, 31 problem formulation, 89 example of FAO remote sensing surveys of survey evaluation, 100 tropics, 35 survey findings, 94 strategy for adopting new technologies, 32 fuel wood gathering and sal strengthening of core competence, 32 leaf plucking, 98 the role of technology in forest livelihood and resource use pattern, 97 inventories, 31 the condition of the forest floor, 95