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 sov- ereign 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 presen- tation 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 interna- tional 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 sys- tematic 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 man- agement 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 establish- ment (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 esti- mating 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 ade- quate. 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 sta- tistics 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 consis- tency 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 esti- mating 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, communica- tion 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 develop- ments 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 infor- mation demanded. The methods of assessments have not grown in a corresponding manner. Country capacity has the tendency to stagnate. The conventional inven- tory 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 com- prehensive 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 classi- fication 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 respec- tively 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 protec- tive 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-indus- trialized 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 pro- duction 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 func- tions, 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 dis- tinguished 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 com- modities. 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 pro- duction 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 environ- mental 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 docu- mented, 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 equa- tions 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 meth- ods 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 infor- mation, 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 environ- mental 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 rela- tively 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 con- sistent 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 strength- ening 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 assess- ment of forest cover state and change at global and regional levels; • Develop and disseminate a simple and robust monitoring technique for pro- ducing 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 interpre- tation) 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, frag- mentation, 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 moni- toring 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 pre- sentation 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 bio- mass 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 com- munity. 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 inter- national 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 defini- tions, 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 infor- mation in planning) • Technology applications: identify, evaluate, and apply new technology appro- priate 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 regula- tions 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 discus- sion 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 envi- ronment, 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 devel- opment 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 envi- ronmental, and other non-wood aspects of the forest under management. Besides data collected from forest inventories, long-term experimental obser- vations 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 impli- cations 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 con- sumption, 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 anal- ysis. (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 fol- lowed by regeneration (Fig. 5.1). In the NTFP context, new knowledge has to be developed to determine sus- tainable 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, expe- rience and reliable inventories of forest ecosystems, silviculture and land manage- ment. 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 interna- tional 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 difficul- ties, 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 preci- sion 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 accessi- bility. 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 ran- domly 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 con- ceptual 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 fol- lowing 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):
VSTðÞ¼VSRðÞ fR2=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 mea- surements. 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 pro- cedures 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 avail- ability 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 pro- duction 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 sam- pling 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 implemen- tation. 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, respec- tively, 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 planta- tion, 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 discontin- uous 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 useful- ness 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-effec- tiveness 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 ref- erence 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 develop- ments, 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 equa- tion 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 com- mon 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 stem- wood 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 characteris- tics 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 knowl- edge). 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, indi- cating 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 deter- mined 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 experi- ence 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 likeli- hood 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 cor- rected, 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 cal- culation 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 1Þ x2 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 defi- nitions, 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, inte- grate, 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 pro- cessing 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 envi- ronment, 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 oper- ations 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 commu- nity, 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 fore- going 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; house- hold 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 agri- cultural 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 for- ests 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 rela- tively 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 sus- tainable 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 con- sideration 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 arrange- ment. 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 var- iance 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 indi- cates 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 semi- evergreen 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 dis- tribution 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 rela- tion, 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 scle- roxylon, 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 conclu- sion 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 distri- bution. 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 pos- sibilities 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 coef- ficient 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 reso- lution 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 tech- niques 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, par- ticularly 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 obser- vation 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 measure- ments 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 investi- gations 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 miscella- neous. These differences highlight the problem associated with definition of spe- cies 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 mea- surement, 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 mea- surement 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 corre- lation 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 repro- duced 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 rep- resentative 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 1 R2 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 pro- ducer 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 compar- ing ‘‘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 envi- ronment. 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 for- mation 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 dis- tribution, 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 domi- nant 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 mea- surements 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 atten- tion 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 con- ditions. 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 b R=100 ÀÁc 1 dYðÞ=dRðÞ¼0:01 a c lnðÞ b b R=100 1 b R=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