STRENGTHENING FOREST FIRE MANAGEMENT IN INDIA JOINT REPORT BY THE MINISTRY OF ENVIRONMENT, FOREST AND CLIMATE CHANGE, GOVERNMENT OF INDIA, AND THE WORLD BANK JUNE 2018 © 2018 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Photographs from other sources excepted. Attribution—Please cite the work as follows: Dogra, Pyush, Andrew Michael Mitchell, Urvashi Narain, Christopher Sall, Ross Smith, and Shraddha Suresh. 2018. “Strengthening Forest Fire Management in India”. World Bank, Washington DC. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. M‚- g"kZ o/kZu Hkkjr ljdkj Dr. Harsh Vardhan i;kZoj.k, ou ,oa tyok;q ifjorZu ea=h GOVERNMENT OF INDIA MINISTER OF ENVIRONMENT, FOREST & CLIMATE CHANGE MESSAGE Forest fire management is part of India’s longer-term vision for sustainable forest management, especially in light of India’s international commitments for cooperation on climate change, as we face what has become an issue of global concern. Protecting forests from undesirable fires is crucial to sustaining India’s progress on meeting its global pledge to create an additional carbon sink of 2.5 to 3 billion tonnes of CO2-equivalent through additional forest and tree cover by 2030. It is heartening to note that the present study provided critical inputs for the preparation of a National Action Plan on Forest Fire in India by the Ministry of Environment, Forest and Climate Change, which was accomplished earlier this year. It gives me great pleasure to present this report on strengthening forest fire management in India. This collaboration with the World Bank to generate an improved understanding of the current status of forest fire management in India along with recommendations for the future represents an important initiative for ensuring that significant progress will be made to address the challenge for forest fires in the country more effectively, and to protect its precious forest resources, biodiversity and carbon sequestration capacity. Date: 27.09.2018 (Dr. Harsh Vardhan) Paryavaran Bhawan, Jor Bagh Road, New Delhi-110 003 Tel.: 011-24695136, 24695132, Fax: 011-24695329 laL—fr jkT; ea=h ¼Lora= çHkkj½ i;kZoj.k, ou ,oa tyok;q ifjorZu jkT; ea=h Hkkjr ljdkj MINISTER OF STATE (I/C) OF CULTURE M‚- egs'k 'kekZ MINISTER OF STATE FOR Dr. Mahesh Sharma ENVIRONMENT, FOREST AND CLlMATE CHANGE GOVERNMENT OF INDIA MESSAGE lndia remains committed to sustainable development and to strengthening its forest policies, and the National Forest Policy of 1988 is currently being revised. As part of the National Mission for Green India, under India’s National Action Plan on Climate Change, we have embarked on an ambitious path to increase forest and tree cover by 5 million hectares and to improve the quality of forest on another 5 million hectares. Achieving these targets will benefit the livelihoods of about 3 million forest dependent households. Forests fires extract a huge toll on India’s economy and society. Today, with growing populations in and around the forests, these fires are putting more lives and property at risk. Fires also lead to the loss of biodiversity, put water sources at risk, and lead to health impacts from exposure to air pollution. I thank all those involved in the preparation of this study on strengthening forest fire management in India and look forward for more such fruitful collaborations. (Dr. Mahesh Sharma) iape ry] vkdk'k foax] bafnjk i;kZoj.k Hkou, tksj ckx jksM+, ubZ fnYyh-110 003, Qksu : 011-24621921, 24621922, QSDl : 011-24695313 dSEi dk;kZy; % ,p-33, lSDVj-27, uks,Mk-201301 ¼m-ç-½ nwjHkk"k % 0120-2444444, 2466666, QSDl % 0120-2544488 5th Floor, Aakash Wing, Indira Paryavaran Bhawan, Jor Bagh Road, New Delhi-110 003, Ph. : 011-24621921. 24621922 Fax : 011-24695313 Camp Office : H-33, Sector-27, Noida-201301 (U.P.) Tel.-: 0120-2444444, 2466666, Fax : 0120-2544488 E-mail : dr.mahesh@sansad.nic.in, drmahesh3333@gmail.com lfpo Hkkjr ljdkj i;kZoj.k, ou ,oa tyok;q ifjorZu ea=ky; SECRETARY GOVERNMENT OF INDIA lh-ds- feJk MINISTRY OF ENVIRONMENT, FOREST & CLIMATE CHANGE C.K. Mishra MESSAGE With changing climate, more people living in and around forests, and expanding agriculture in many tropical forested countries, the area of forest that is burnt each year has grown, and fire seasons are growing longer. The prevention and management of forest fires is a priority for achieving the goals that we have set for a green India. It is important to highlight the role that local communities - the very people who rely on forests for their livelihoods - play in the prevention and management of forest fires. Indeed, the National Mission for Green India is based on a participatory, grassroots approach. Moreover, the increasing vulnerability of forests to fires as a result of a changing climate has already been recognized in the State Action Plans on Climate Change for some of the States in India. I am pleased to note that the recommendations from this report on strengthening forest fire management in India have already fed into the preparation of a National Action Plan on Forest Fire in India by the Ministry of Environment, Forest and Climate Change. I look forward to further progress being made in the country in terms of improving forest fire management. [C.K. Mishra] Dated: 28th September, 2018 Place: New Delhi bafnjk i;kZoj.k Hkou, tksj ckx jksM+, ubZ fnYyh-110 003, Qksu : (011) 24695262, 24695265, QSDl : (011) 24695470 INDIRA PARYAVARAN BHAWAN, JOR BAGH ROAD, NEW DELHI-110 003, Tel : (011) 24695262, 24695265 Fax : (011) 24695470 E-mail : secy-moef@nic.in, Website: moef.gov.in ou egkfuns’kd ,oa fo’ks"k lfpo Hkkjr ljdkj fl)kUr nkl i;kZoj.k, ou ,oa tyok;q ifjorZu ea=ky; SIDDHANTA DAS DIRECTOR GENERAL OF FOREST & SPL. SECY. GOVERNMENT OF INDIA MINISTRY OF ENVIRONMENT, FOREST AND CLIMATE CHANGE MESSAGE Twenty-one young persons succumbing to a devastating forest fire in Theni forests in Tamil Nadu is a testimony to the magnitude of the calamity which in turn highlights how important it is to prevent and manage forest fires. With this in view, technical support was sought by the Ministry of Environment, Forest and Climate Change (MoEFCC) from the World Bank. As a follow up this study was conceived in 2016 and it took about a year to complete the same. It is clear that capacity building and institutional coordination among the various agencies and stakeholders involved in aiding state forest departments in managing large fires are critical for effective forest fire management. I congratulate the team for preparing this report on strengthening forest fire management in India. bafnjk i;kZoj.k Hkou, tksj ckx jksM, ubZ fnYyh-110 003, Qksu : 24695278, QSDl : (011) 24695412 INDIRA PARYAVARAN BHAWAN, JOR BAGH ROAD, NEW DELHI-110 003, Ph. : 24695278, Fax : (011) 24695412 E-mail : dgfindia@nic.in 'kSoky nklxqIrk vfrfjDr ou egkfuns’kd SAIBAL DASGUPTA Hkkjr ljdkj i;kZoj.k, ou ,oa tyok;q ifjorZu ea=ky; ADDITIONAL DIRECTOR GENERAL OF FORESTS GOVERNMENT OF INDIA MINISTRY OF ENVIRONMENT, FOREST & CLIMATE CHANGE MESSAGE The sustainable management of forest assets and their growth has long been a priority in India, which aims to bring 33% of its geographical area under forest or tree cover. The total forest and tree cover in the country has grown steadily and stands at 24.39% of its geographical area as per the 2017 assessment by the Forest Survey of India (FSI). However, forest fires present a major challenge to protecting India’s forests, making it more difficult for India to maintain and increase its carbon sinks. Upgrading the technology and equipment used for FFPM and improving information on forest fires and knowledge of good practices in preventing and managing forest fires in India are equally important. Indeed, FSI is working to develop an early warning system with respect to forest fires, which will bolster current efforts to prevent and manage forest fires in the country. I am confident that this report will serve to further strengthen India’s efforts to manage forest fires efficiently. (SAIBAL DASGUPTA) J-504, ty foax, bafnjk i;kZoj.k Hkou, tksj ckx jksM+, ubZ fnYyh-110 003, Qksu : 011-24695279, QSDl : 011-24695280 J-504, Jal Wing, Indira Paryavaran Bhawan, Jor Bagh Road, New Delhi-110 003 Ph. : 011-24695279, Fax : 011-24695280 E-mail : adgfc-mef@nic.in, saibaldasgupta@hotmail.com CONTENTS FOREWORD 1 ACKNOWLEDGMENTS 2 ACRONYMS 3 EXECUTIVE SUMMARY 6 INTRODUCTION 17 Forestry Sector in India 17 Study Objectives and Methodology 23 Structure of the Report 24 CHAPTER 1: Characterizing the Forest Fire Challenge in India 26 1.1 Overall Patterns and Trends in Forest Fires 26 1.2 Causes of Forest Fires and Factors Influencing Fire Behavior 31 1.3 Impacts of Forest Fires 46 1.4 Summary 53 CHAPTER 2: Assessment of Current Policies, Plans, and Practices 55 2.1 Policies, Plans, and Funding for Forest Fire Prevention and Management 55 2.2 Forest Fire Prevention and Management (FFPM) Practices 64 2.3 Summary 102 CHAPTER 3: Institutional Coordination and Community Engagement 105 3.1 Coordinating Across Agencies 105 3.2 Engaging with Communities 109 3.3 Collaborating with Research Organizations 115 3.4 Summary 115 CHAPTER 4: Recommendations 117 4.1 Policies, Plans, Regulations, and Funding 117 4.2 Fire Prevention Practices 122 4.3 Fire Detection 124 4.4 Fire Suppression 125 4.5 Post-Fire Management 127 4.6 Engaging with Communities 131 4.7 Coordination with Other Agencies and Entities 133 4.8 Forest Fire Science, Data, Knowledge Sharing, and Training 134 4.9 Summary and Prioritization 135 Strengthening Forest Fire Management in India   viii ANNEX 1: Data and Methods for Geospatial Analysis of Forest Fires 139 ANNEX 2: Survey of State Forest Department Officers 157 ANNEX 3: Community Consultations and Case Studies 176 ANNEX 4: Classification of the Causes of Forest Fires 188 ANNEX 5: Equipment for Forest Firefighting in India 191 ANNEX 6: Summary Notes of Workshop Proceedings 194 ANNEX 7: Data Sheets Sent to State Forest Departments 203 ANNEX 8: Questionnaire Sent to State Disaster Management Agencies 207 REFERENCES 208 FIGURES Figure ES.1: Forest Cover, Active Forest Fire Detections, and Burnt Forest Area by Region 7 Figure ES.2: The Forest Fire Prevention and Management (FFPM) Cycle 9 Figure 1.1: Share of Active Forest Fire Detections by MODIS per District and Region, 2003-2016 28 Figure 1.2: Distribution of Forested Area Affected by Fire per District and Region, 2003-2016 29 Figure 1.3: The Wildfire Triangle 31 Figure 1.4: Seasonality of Forest Fires by State and Region 33 Figure B1.1: El Niño/La Niña Events and Monsoon Rainfall, 1930-2015 34 Figure B1.2: Monsoon Rainfall and Next-Year Fire Season Severity, 2003-2016 35 Figure 1.5: Distribution of KBDI Values in Districts on Days with or without Fires During Peak Forest Fire Season (February-May), 2012-2016 37 Figure 1.6: Predicted Daily Probability of Fire Detection at Different Levels of Drought, by Region 38 Figure 1.7: Spatial Distribution of Forest Fire Detections Relative to Nearest Road (2014-2016) 39 Figure 1.8: Spatial Distribution of Forest Fire Detections Relative to Nearest Built-Up Area (2014-2016) 39 Figure 1.9: Terrain Ruggedness Scores for Forests in Which Fires Were Detected (2003-2016) 40 Figure 1.10: Forest-Covered Area Affected by Fire by Count of Months the Area Was Affected, 2003-2016 42 Figure 2.1: Percentage of Forest Area Under Joint Forest Management (JFM) 60 Figure 2.2: The Forest Fire Prevention and Management (FFPM) Cycle 65 Figure 2.3: Biggest Challenges to Effective Forest Fire Prevention Identified by Responding Officers 66 Figure 2.4: Share of Forest Department Survey Respondents Who Said Fire Lines in Their Area Were All Cleared per the Working Plan 68 Figure B2.1: Fire Risk Zonation Map, Bhadrachalam South Division, Telangana 74 Figure 2.5: Grid-Based Pre-Warning Alert System Implemented by FSI Starting in 2017 78 Figure 2.6: State-Wise Number of Users Registered with FSI Forest Fire Alert System 2.0 83 Figure 2.7: Unstaffed Watchtower 85 Figure 2.8: Principal Techniques Used to Suppress Forest Fires (Frequency Count of Responses) 90 Figure 2.9: Additional Equipment Needs Mentioned by Officers 92 Figure B2.2: Leaf Blowers 93 Figure 2.10: Check Dams and Water Retention Works in Fire-Affected Chir Pine Forests of Uttarakhand 103 Figure 3.1: Burnt Forest Area Reported in Uttarakhand on State Forest Department and Non-State Forest Department Land (square kilometers) 106 Figure 3.2: Self-Rated Effectiveness of the Forest Department’s Engagement with the Local Community in Preventing Forest Fires 109 Figure 3.3: Methods of Community Engagement Used by the Forest Department 110 ix Strengthening Forest Fire Management in India Figure 3.4: How Can Engagement with the Local Community be Improved? 111 Figure A1.1: Fires Detected per District from January to May, 2003-2016 142 Figure A1.2: Distribution of Forested Areas with and without Fires by Distance to Nearest Road, Using MODIS Detections for 2014-2016 152 Figure A1.3: Distribution of Forested Areas with and without Fires by Distance to Nearest Road, Using VIIRS Detections for 2016 152 Figure A1.4: Distribution of Forested Areas with and without Fires by Distance to Nearest Built-Up Settlement, Using MODIS Detections for 2014-2016 153 Figure A1.5: Distribution of Forested Areas with and without Fires by Distance to Nearest Built-Up Settlement, Using VIIRS Detections for 2016 153 Figure A3.1: Districts Visited in Meghalaya for the Community Appraisal 176 Figure A3.2 Mawphlang Sacred Groves 177 Figure A3.3: REDD+ Project Area in Mawphlang 177 Figure A3.4 View of the REDD+ Project Area under Mawphlang Block 177 Figure A3.5: Charcoal Making as One of the Livelihood Activities in Mawphlang 178 Figure A3.6: Community Members Creating a Fire Line in the REDD+ Project Area 178 Figure A3.7: Fire Lines in Project Area 179 Figure A3.8: Improved Fuel Wood Stoves to Reduce Consumption 179 Figure A3.9: Distribution of LPG Connection to Reduce Fuel-Wood Collection from Forests 180 Figure A3.10: Awareness Programme on Forest Fire Protection Measures 180 Figure A3.11: Area Affected by Forest Fires in ha. 180 Figure A3.12: Jirang Forest in Rainy Season 181 Figure A3.13: Forest Fire is Rampant in Jirang due to the Practice of Jhum Cultivation in Forest Areas 181 Figure A3.14: Dependence of People on Forest for Fuel Wood in Jirang Region 182 Figure A3.15: Village Fire Control Committee Members in the Jirang Area 182 Figure A3.16: A Fire Line Covered in Thick Foliage Which Grew Over the Monsoon 182 Figure A3.17: A View of the Khoo Blai Sein Raij Tuber Community Reserved Forest 183 Figure A3.18: Signboard with Rules and Regulations Installed at Entrance of the Community Reserve 183 Figure A3.19: Prominent and Well-Maintained Fire Lines Along the Entire Periphery of the Forest 183 Figure A3.20: Tapioca is Commonly Used as an Effective Fire Barrier Around Jhum Areas 184 Figure A3.21: Jhum Cultivation is a Common Practice in the Garo Hills 185 Figure A3.22: Forests Adjoining Agricultural Fields Declared as Community Forests - Managed and Protected by the Community 186 Figure A3.23: A Fire Line Created Around the Jhum Areas 187 Figure A3.24: A Panoramic View of a Community Reserve at Rangwal Village Supporting Paddy Field 187 Figure A3.25: A Fire Line Around the Community Reserve at Sanchagre Village 187 Figure A3.26: A Water Conservation Pond Recharged from Community Forest Reserve in Rangwal Village 187 Figure A5.1: Fire Rake with Bamboo Handle in Odisha 191 Figure A5.2: Fire Beater 191 Figure A5.3: Pulaski Tool 192 Figure A5.4: Knapsack Sprayer 192 Figure A5.5: Firebug Torch 192 MAPS Map 1: Forest Types and Distribution in South India 18 Map 2: Forest Types and Distribution in Central India 19 Strengthening Forest Fire Management in India   x Map 3: Forest Types and Distribution in Northeast India 20 Map 4: Forest Types and Distribution in North and West India 21 Map 1.1: Forested Areas Affected by Widespread and Frequent Burning in Central India 29 TABLES Table ES.1: Summary of Recommendations and Priorities 13 Table 1.1a: Top 20 Districts by Total Number of Fire Detections, 2003-2016 27 Table 1.1b: Top 20 Districts by Total Area Affected by Fire, 2003-2016 27 Table 1.2: Year-On-Year Trend in Fire Detections by State, 2003-2016 (Annual Percent Change) 30 Table 1.3: Implied Average Fire Recurrence Interval by Forest Type, 2003-2016 42 Table 1.4: Causes of Forest Fires According to Surveyed Forest Department Officers, by State (Index of Importance, 0-100) 44 Table 1.5: Poverty Rates and Fire Density in Rural Forested Districts, Grouped by Quartile According to the Poverty Headcount Ratio in 2011 46 Table 1.6: Previous Nation-Wide Assessments of Carbon Emissions from Forest Biomass Burning 52 Table 1.7: Reported Monetary Losses due to Forest Fires in Select States 54 Table 2.1: Agencies Involved in FFPM in India 56 Table 2.2: Types of Community-Held Forests in Meghalaya, Mizoram, and Nagaland 61 Table 2.3: Funding for Forest Management Under Centrally Sponsored Schemes and Other Programs, 2011-2016 64 Table 2.4: Ranking of Challenges to Effective Forest Fire Prevention Identified by Respondents (1st = Most-Mentioned) 67 Table 2.5: Information on Fire Lines Provided by State Forest Departments 67 Table 2.6: Field Reporting on Satellite-Based Fire Alerts in Eight States 82 Table 2.7: Forest Fires Reported by Ignition Source, Kerala (Number of Incidents) 96 Table 2.8: Remote Sensing-Based Versus Field-Reported Estimates of Burnt Forest Area in 2014 (km2) 96 Table 2.9: Schedule of Rates for the Calculation of Damages from Forest Fire in Uttarakhand 97 Table 2.10: Guidelines for Assessing the Costs of Forest Fires in Rampur Division, Himachal Pradesh 98 Table B2.1: Damages from 2015 Forest Fires in Indonesia 99 Table 3.1: Forest Areas Managed by Different Entities 106 Table B4.1: Interventions for Improving FFPM, Impacts, and Potential Indicators Based on the World Bank Kazakhstan Forest Protection and Reforestation Project 128 Table 4.1: Summary of Recommendations and Priorities 136 Table A1.1: Regional Definitions 140 Table A1.2: Descriptive Statistics for District-Level Analysis of Monsoon Rainfall and Forest Fires 141 Table A1.3: Negative Binomial Regression Results for Monsoon Rainfall and the Number of Fires Observed January-May the Following Year 142 Table A1.4: Zero-Inflated Negative Binomial Regression Results for Monsoon Rainfall and the Number of Fires Observed January-May the Following Year 143 Table A1.5: Spearman’s Rank-Order Correlation Coefficients, ENSO Index Values for June- December and the Number of Fires Detected January-May the Next Year, 2003-2016 145 Table A1.6: Mean Monthly Weather Conditions and Fire Detections by District, 2003-2015 146 Table A1.7: District-Level Regression for Monthly Precipitation (mm) and Odds of Fire Detection 147 Table A1.8: District-Level Regression for Monthly Wet-Day Frequency (Days per Month xi Strengthening Forest Fire Management in India with > .01 mm Precipitation) and Odds of Fire Detection 148 Table A1.9: Share of Daily Observations with Fires Detected, by Stage of Drought as Measured by the Keetch-Byram Drought Index (KBDI) and by Region, February 1 to May 31 (2012-2016) 149 Table A1.10: Regression Results for Analysis of KBDI and Odds of Fire Detection, Treating KBDI as a Continuous Variable 150 Table A1.11: Regression Results for Analysis of KBDI and Odds of Fire Detection, Treating KBDI as a Categorical Variable Representing the Stage of Drought 151 Table A1.12: Descriptive Statistics for District-Level Forest Fire Density and Poverty Rates, 2009-13 155 Table A1.13: Regression Results for Forest Fire Density and Poverty Rates, 2009-2013 156 Table A2.1: Biggest Challenges to Effective Forest Fire Prevention Identified by Responding Officers in Each State 170 Table A2.2: Principal Techniques Used to Suppress Forest Fires 171 Table A2.3: Additional Equipment Needs Mentioned by Officers 172 Table A2.4: Methods of Community Engagement Used by the Forest Department 174 Table A2.5: How Can Engagement with the Local Community be Improved? 175 Table A4.1: Categorization of Causes of Forest Fires 188 Table A4.2: Non-Timber Forest Products (NTFPs) Collected Using Fire, According to Surveyed Forest Officers 190 BOXES Box 1.1: The El Niño/Southern Oscillation (ENSO) and Forest Fire Season Severity 34 Box 1.2: Indigenous Knowledge, Early-Season Controlled Burning, and Stemming the Lantana Invasion 51 Box 2.1: Court-Ordered Green Felling Bans and Fire Prevention 71 Box 2.2: Fire Risk Zonation in Telangana, India 74 Box 2.3: Assessing Forest Fire Hazards and Risk in Tamil Nadu 75 Box 2.4: South Africa’s Lowveld Fire Danger Index 76 Box 2.5: Fire Danger Rating Systems in Canada and Indonesia 77 Box 2.6: Satellite Fire Detection in Madhya Pradesh 80 Box 2.7: Detecting Forest Fires in Belarus 85 Box 2.8: Forest Fire Suppression Techniques 86 Box 2.9: The Use of Leaf Blowers in Odisha, India 93 Box 2.10: The Economic Costs of Forest Fires in Indonesia 99 Box 2.11: Estimating Carbon Emissions from Forest Fires 101 Box 3.1: Coordinating to Control Forest Fires in Rajasthan 107 Box 3.2: Partnering with Communities in Madhya Pradesh to Prevent Burning for Tendu Leaves 113 Box 3.3: Scientific Research Organizations Studying Forest Fires in India 114 Box 4.1: Transforming Forest Fire Policy and Practices in Mexico 118 Box 4.2: Using Fire to Prevent Conflagrations in Australian Forests 121 Box 4.3: International Cooperation on Forest Fire Management in Mexico 127 Box 4.4: Indicators for Monitoring Progress and Measuring Results on Forest Fires 128 Box 4.5: Testing Community Incentives for Preventing Forest Fires in Indonesia 131 Strengthening Forest Fire Management in India   xii FOREWORD Forest fires are a challenge across many countries. more vulnerable than others. Engaging with the They lead to the loss of lives and livelihoods for communities that use forests is therefore vital, as is people directly dependent on forest produce, apart improving coordination with the other agencies that from destroying wildlife habitat, causing soil erosion are involved in managing forests and responding to and damaging water supply. According to the Fifth forest fires. Assessment Report of the Intergovernmental Panel on Climate Change, exposure to smoke from landscape The report also discusses policies on forest fire fires (including forest fires) is estimated to cause prevention and management (FFPM) at the national, 260,000 to 600,000 premature deaths annually. The state and local levels, underscoring the need for a report also finds that annual carbon emissions from comprehensive national policy and guidelines. While forest fires range between 2.5 billion to 4.0 billion tons India has made great strides in the use of technology of CO2, adding large volumes of greenhouse gases to for detecting forest fires, there is still a need to the atmosphere. strengthen fire prevention practices and to develop a well-equipped and trained workforce to fight fires. In India, one estimate shows that nearly 49,000 square This report provides detailed recommendations on kilometers of forests – an area larger than the size of five broad themes (Policy, Institutions and Capacity, Haryana – were burned in 2014 alone (a mild year Community Engagement, Technology, and Data and compared to others in the recent past). Apart from the Information) and takes into consideration national damage, forest fires pose a serious threat to India’s and international best practices in FFPM. ability to expand its forest and tree cover by 2030 to create an additional carbon sink of 2.5 to 3 billion We at the World Bank are delighted to have this tons of CO2 equivalent, in keeping with the country’s opportunity to work with the Ministry of Environment, Nationally Determined Contribution (INDC). Indeed, Forest, and Climate Change on this important agenda, India’s Ministry of Environment, Forest and Climate and to contribute towards informing a National Action Change (MoEFCC) has identified forest fires as a Plan on FFPM in India. We look forward to continuing major driver of forest degradation, and noted that this partnership to secure and enhance India’s forest the lack of a comprehensive assessment of what drives wealth. forest fires, and the best way to manage them, hinders effective action. This report analyses patterns and trends of forest fires in India. While the findings of this study indicate Junaid Kamal Ahmad that forest fires occur every year in almost every Country Director for India state in India, some districts have been found to be World Bank 1 Strengthening Forest Fire Management in India ACKNOWLEDGMENTS This report was prepared by the World Bank in Shukla (Madhya Pradesh Forest Department), Mr. response to a request for technical assistance from the C. Budnah (Meghalaya Forest Department), Mr. C. Ministry of Environment, Forest and Climate Change P. Marak (Meghalaya Forest Department), Mr. S. S. (MoEFCC) of the Government of India to help Srivastava (Odisha Forest Department), Dr. P. K. Jha strengthen forest fire prevention and management in (Telangana Forest Department), Mr. Rajinder Kumar the country. The report was prepared by a core team Mahajan (Uttarakhand Forest Department), Mr. comprising of Pyush Dogra, Andrew Mitchell, Urvashi Monish Mullick (Uttarakhand Forest Department), Narain, Christopher Sall, Ross Smith, and Shraddha Mr. Gambhir Singh (Uttarakhand Forest Department) Suresh. Contributions were also received from Swati and Mr. Digvijay Singh Khati (Uttarakhand Forest Saluja, Dinesh C. Khanduri, Suresh Khanduri, Department) who were Principal Chief Conservators Narendra Rathore, and Juan Jose Miranda Montero. of Forests (PCCFs) of these states at the time, as well as Mr. Saibal Dasgupta (Additional Director General The team is very grateful for the support of Dr. Harsh of Forests, MoEFCC; Director General, Forest Survey Vardhan (Honorable Minister for Environment, of India), for their valuable support during field visits. Forest and Climate Change) and Dr. Mahesh Sharma (Minister of State for Environment, Forest and Climate Constructive comments on the report were received Change). The team thanks Mr. C. K. Mishra (Secretary, from the following World Bank peer reviewers: Deepak MoEFCC), Mr. Siddhanta Das (Director General of Singh, Ann Jeannette Glauber, and Grant Milne. Forests, MoEFCC), Dr. Sharad S. Negi (former Director The report also benefited greatly from comments General of Forests, MoEFCC), Mr. Saibal Dasgupta received by reviewers in MoEFCC and the state forest (Additional Director General of Forests, MoEFCC; departments, including: Mr. A.K. Mohanty; Mr. E. Director General, Forest Survey of India), Dr. Rekha Pai Vikram (Deputy Director, Forest Survey of India); (former Inspector General of Forests, MoEFCC), and Mr. C.P. Marak; Mr. Alok Nagar (Chief Conservator Mr. Anjan K. Mohanty (Inspector General of Forests, of Forests, Himachal Pradesh Forest Department); Mr. MoEFCC) for their overall support and guidance. Murray (Chief Conservator of Forests, Mizoram Forest The team would also like to thank Mr. Kamal Kishore Department); Dr. Thomas Chandy (Principal Chief (Member, National Disaster Management Authority of Conservator of Forests, Sikkim); Dr. P.K. Gupta (Chief India) and Ms. Pratima Gupta (Director, NITI Aayog) Conservator of Forests, Bihar Forest Department); Dr. for their valuable support. The team also thanks Jayanthi (Chief Conservator of Forests, Tamil Nadu); Annette Dixon (Regional Vice President, World Bank), and Mr. R.K. Sood (retired IFS officer). The team Junaid K. Ahmad (Country Director, World Bank), would also like to acknowledge several other World Karin E. Kemper (Senior Director, World Bank) and Bank colleagues for their suggestions, including Tapas Kseniya Lvovsky (Practice Manager, World Bank) for Paul, Anup Karanth and Peeyush Sekhsaria. their encouragement and support. The authors would like to thank Sandhya Krishnan The report benefited greatly from discussions with and Latha Sridhar for production management. The participants at an international workshop on forest manuscript was edited by Simi Mishra and designed by fire prevention and management co-organized by Roots Advertising. Sketches in the report were created MoEFCC and the World Bank in November 2017. The by Dinesh Francis at the workshop. Any remaining workshop brought together practitioners and experts errors or omissions are the authors’ own. from Australia, Belarus, Canada, Mexico, Nepal, the United States, and the Food and Agriculture The team gratefully acknowledges the financial Organization of the United Nations, as well as support provided for the report by the Program on government officials, researchers, and experts in India. Forests (PROFOR) of the World Bank. The team would like to thank Mr. Ramesh Hembrom (Jharkhand Forest Department), Dr. Animesh Cover Photo: Vikas Gusain Strengthening Forest Fire Management in India   2 ACRONYMS ADB Asian Development Bank ADC Autonomous District Council AFAC Australian Fire and Emergency Service Authorities Council AIC Akaike Information Criterion ANOVA Analysis of Variance ATREE Ashoka Trust for Research in Ecology and the Environment AWiFS Advanced Wide Field Sensor BA Burned Area BAPPENAS Indonesian Ministry of National Development Planning BIC Bayesian Information Criterion BRT Bilgiri Rangaswamy Temple BSNL Bharat Sanchar Nigam Limited C+NVC Cost-plus-Net-Value-Change CAG Comptroller and Auditor General (India) CAMPA Compensatory Afforestation Fund Management and Planning Authority CC Combustion Completeness CCF Chief Conservator of Forests CF Conservator of Forests CFFDRS Canadian Forest Fire Danger Rating System CIFOR Center for International Forestry Research CO2e Carbon Dioxide Equivalent CONAFOR National Forestry Commission (Mexico) CRC Cooperative Research Centre (Australia) CRIS-IMD Customized Rainfall Information System-India Meteorological Department CRPF Central Reserve Police Force CRU Climate Research Unit (University of East Anglia) DAFF Department of Agriculture, Forestry and Fisheries (South Africa) DC Drought Code DDMA District Disaster Management Authority DFO Divisional Forest Officer DRDA District Rural Development Agency DRO Deputy Range Officer DSC Decision Support Centre DWAF Department of Water Affairs and Forestry (South Africa) EC JRC European Commission Joint Research Centre ECLAC Economic Commission for Latin America and the Caribbean (United Nations) EF Emission Factor EFFIS European Forest Fire Information System ENSO El Niño/Southern Oscillation ENVIS Environmental Information System (Government of India) EU European Union FAMS Fire Alert Messaging System FAO Food and Agriculture Organization of the United Nations FBP Fire Behaviour Prediction 3 Strengthening Forest Fire Management in India FDRS Fire Danger Rating System FFMC Fine Fuel Moisture Code FFMG Forest Fire Management Group (Australia) FFPM Forest Fire Prevention and Management FIR First Information Report FIRMS Fire Information for Resource Management System (NASA) FL above-ground Fuel Load FOP Fire Occurrence Prediction FRA Forest Rights Act FRI Forest Research Institute (India) FRI Fire Return Interval FRP Fire Radiative Power FSI Forest Survey of India FWI Fire Weather Index GDP Gross Domestic Product GHG Greenhouse Gas GIS Geographic Information System GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit GPS Global Positioning System HP Himachal Pradesh IAF Indian Air Force ICS Incident Command System IIRS Indian Institute of Remote Sensing IISc Indian Institute of Science IITM Indian Institute of Tropical Meteorology INR Indian Rupee IPCC Intergovernmental Panel on Climate Change IRI Columbia University’s International Research Institute for Climate and Society ISI Initial Spread Index ISO International Organization for Standardization ISRO Indian Space Research Organisation IT Information Technology ITGC Information Technology and Geoinformatics Centre (Uttarakhand, India) JAS July, August, and September JASOND July, August, September, October, November, December JFM Joint Forest Management JFMC Joint Forest Management Committee JJA June, July, and August JJAS June-September JJASOND June-December KBDI Keetch-Byram Drought Index KML Keyhole Markup Language LDEO Lamont-Doherty Earth Observatory (Columbia University) LIDAR Light Detection and Ranging System LPG Liquefied Petroleum Gas MODIS Moderate Resolution Imaging Spectroradiometer MoEF Ministry of Environment and Forests (now MoEFCC) MoEFCC Ministry of Environment, Forest and Climate Change (Government of India) MOU Memorandum of Understanding NASA National Aeronautics and Space Administration (of the United States of America) Strengthening Forest Fire Management in India   4 NBR Negative Binomial Regression NCA National Commission on Agriculture (India) NDC Nationally Determined Contribution NDMA National Disaster Management Authority NDRF National Disaster Response Force NE Northeast NER-DRR North Eastern Regional Node for Disaster Risk Reduction NESAC North Eastern Space Applications Centre NFC National Forest Commission NFDRS National Fire Danger Rating System NGO Non-Governmental Organization NGT National Green Tribunal NPP Net Primary Productivity NRSC National Remote Sensing Centre NTFP Non-Timber Forest Product OGD Open Government Data OLI Operational Land Imager OND October, November, December ONI Oceanic Niño Index OSDMA Odisha State Disaster Management Authority PCCF Principal Chief Conservator of Forests PM Particulate Matter POR Preliminary Offense Report REDD+ Reducing Emissions from Deforestation and Forest Degradation in Developing Countries RO Range Officer SDMA State Disaster Management Authority SDRF State Disaster Response Force SEIAA State Level Environment Impact Assessment Authority SFD State Forest Department SMS Short Message Service SOP Standard Operating Procedure SPI Standardized Precipitation Index SRES Special Report on Emissions Scenarios (of the IPCC) SRTM Shuttle Radar Topography Mission SST Sea Surface Temperature Suomi-NPP Suomi National Polar-Orbiting Partnership TNI Trans-Niño Index TRI Terrain Ruggedness Index UN United Nations US NOAA National Oceanic and Atmospheric Administration (of the United States of America) USA United States of America USD United States Dollar UT Union Territory VFCC Village Fire Control Committee VIIRS Visible Infrared Imaging Radiometer Suite VSS Vana Samrakshana Samithi WII Wildlife Institute of India WSN Wireless Sensor Network YLA Young Lai Association YMA Young Mizo Association 5 Strengthening Forest Fire Management in India EXECUTIVE SUMMARY Fire has been a part of India’s landscape since time forest fires and other landscape fires (Johnston et al. immemorial and can play a vital role in healthy 2012). forests, recycling nutrients, helping tree species regenerate, removing invasive weeds and pathogens, Tackling forest fires is even more imperative in and maintaining habitat for some wildlife. Occasional India as the country has set ambitious policy goals fires can also keep down fuel loads that feed larger, for improving the sustainability of its forests. As more destructive conflagrations, but as populations part of the National Mission for Green India under and demands on forest resources have grown, the India’s National Action Plan on Climate Change, the cycle of fire has spun out of balance. Large areas of government has committed to increase forest and tree degraded forest are now subject to burning on an cover by 5 million hectares and to improve the quality annual or semi-annual basis. As these fires are no of forest on another 5 million hectares. Relatedly, longer beneficial to forest health, India is increasingly under its NDC, India has committed to bringing 33 wrestling with how to improve the prevention and percent of its geographical area under forest cover management of unwanted forest fires. and to create additional sinks of 2.5 billion to 3 billion tons worth of CO2 stored in its forests by 2030. Yet, India is not alone in facing this challenge. Forest fires it is unclear whether India can achieve these goals have become an issue of global concern. In many other if the prevention and management of forest fires is countries, wildfires are burning larger areas, and fire not improved. Field-verified data on the extent and seasons are growing longer due to a warming climate severity of fires are lacking and understanding of the (Jolly et al. 2015). With growing populations in and longer-term impacts of forest fires on the health of around the edges of forests, more lives and property India’s forests remains weak. are now at risk from fire. About 670,000 km2 of forest land are burned each year on average (about 2 percent The objective of this assessment is to strengthen of the world’s forested areas [van Lierop et al. 2015]), knowledge on forest fires by documenting releasing billions of tons of CO2 into the atmosphere,1 current management systems, identifying gaps in while hundreds of thousands of people are believed to implementation, and making recommendations on die due to illnesses caused by exposure to smoke from how these systems can be improved. According to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, annual carbon emissions from forest fires 1 are in the range of 2.5 billion to 4.0 billion tons of CO2 (Smith et al. 2014). Strengthening Forest Fire Management in India   6 WHAT DID THE ASSESSMENT areas present priorities for intervention and should be the focus of improving FFPM, as should areas of REVEAL? significant ecological, cultural, or economic value. 1. Forest fires in India are both widespread and Data from 2014, for example, showed that about 10 concentrated percent of forest cover in protected areas was affected by fire (Reddy et al. 2017b). Every year, forest fires occur in around half of the country’s 647 districts and in nearly all the states.2 2. Fire potential and behavior is shaped by a Furthermore, by one estimate, in 2014 alone, nearly combination of natural and social factors 49,000 km2 of forests – an area larger than the size of Haryana – were burned (Reddy et al. 2017b). Yet, In India’s seasonally dry forests, most forest fires though fires are spread throughout the country, they are characterized by low-intensity surface fires. The occur much more frequently and affect forest more potential for more intense and difficult-to-control in some districts than in others. Just 20 districts, fires is shaped by a complex dynamic involving the representing 3 percent of the India’s land area and monsoon rains, weather during the winter and early 16 percent of the country’s forest cover in 2000, part of the dry season, and fuel accumulation. Also, accounted for 44 percent of all forest fire detections although India’s forests are densely populated—and from 2003 to 2016. Twenty districts (not necessarily most fires occur within a few kilometers of the nearest the same ones) also accounted for 48 percent of the road or settlement—each year there is a long tail of total fire-affected area between 2003 to 2016, despite fires in more remote and inaccessible areas, where having just 12 percent of the nation’s forest cover in response is slower and the potential for fires to grow 2000 and 7 percent of its land area. While states in beyond control is greater. the Northeast account for the greatest share of fire detections, the largest area affected by fire is in the Weather, fuels, and topography may influence fire Central region (figure ES.1). Districts with the highest potential and behavior, but virtually all forest fires in frequency of fire and largest extent of fire-affected India, as in other parts of the world, are caused by FIGURE ES.1:  FOREST COVER, ACTIVE FOREST FIRE DETECTIONS, AND BURNT FOREST AREA BY REGION 60 56 55 50 Percent of national total 40 36 Burnt forest area, 2003-2016 30 27 28 Active fire detections, 2003-2016 21 22 Forest cover, 2015 20 16 10 11 10 44 232 1 1 0 0 Central North Northeast South West Western Himalaya Source: Forest cover data from FSI (2015) and Hansen et al. (2013); MODIS monthly data product for active fires (MCD14ML), provided by Forest Survey of India; MODIS monthly data product for burnt area (MCD45A1) The cited number of districts is as of 2012. 2. 7 Strengthening Forest Fire Management in India people. Roughly 92-103 million people live in areas 4. A vacuum exists at the level of national policy of forest cover, and many depend on forests for their livelihoods. Many of the important goods and services A cohesive policy framework with a clear strategic that people obtain from forests, such as fodder for direction provides the foundation for successful their livestock, are generated or gathered through FFPM. A vacuum currently exists at the policy the aid of fire. Unwanted forest fires may also occur level. The National Green Tribunal issued a due to human negligence, for example, from casually ruling in August 2017 calling for MoEFCC to discarded cigarettes or from poor control of burning formulate a national policy or guidelines for FFPM on adjacent croplands. Shifting societal and cultural in consultation with the states. The need for a practices also play a role, as with the use of fire in cohesive national policy on FFPM was also voiced traditional shifting cultivation (jhum). In some parts in a Parliamentary Committee report presented of the country, the erosion of traditional community to the Rajya Sabha in December 2016 (Standing institutions for managing forest lands has also Committee 2016). Though MoEFCC had issued contributed to more unwanted forest fires. national guidelines on FFPM in 2000, they are no longer being implemented. Without guidance and 3. The longer-term impacts and wider costs of forest standard setting from above, there is significant fires are still poorly understood variation from state to state and district to district in terms of the detail and substance on FFPM found The longer-term impacts of the current pattern of in local policies and working plans. forest fires on India’s forest ecology and the wider economy are still poorly understood; however, the Policies and prescriptions for FFPM should be available scientific evidence supports that fires are supported by adequate and predictable financing. having a degrading effect. Repeated fires in short A shortage of dedicated funding for FFPM at the succession are reducing species richness and harming central and state level has been a perennial issue, natural regeneration, in combination with other which has been documented by the Comptroller and pressures such as intense grazing and browsing. Auditor General in various states. Along with a lack Reductions in biomass, species diversity, and natural of public engagement, forest officers surveyed for regeneration due to fire may pose a risk to policy goals the assessment cited insufficient equipment, labor, for enhancing India’s forest carbon sinks. Not all fires and financial resources as one of the main challenges are bad, though. The key is to maximize the ecological for effective FFPM. Revamping the Intensification of benefits of fire while minimizing the adverse impacts, Forest Management Scheme to focus exclusively on recognizing that the controlled use of fire may play FFPM represents a positive development. Directing a positive role in the management of fire-adapted more resources specifically for FFPM will need to forests. happen at the state level too. Current estimates of the economic costs of forest 5. Forest fire prevention is not being implemented fires in India, at around INR 1,101 crore (US$ 164 consistently million, 2016 prices) per year, are almost certainly underestimates. Monetary damages due to forest fires Prevention is the most crucial link in the FFPM chain are generally assessed only for the loss of standing and should receive the greatest support. Prevention trees (natural or planted) in terms of their timber activities have included primarily the creation and value, which are usually minimal in the event of low- maintenance of fire lines and controlled area burning. intensity surface fires such as those that commonly Only half of the forest officers surveyed in 11 states occur in India. Estimates could be improved by said that all the fire lines in their area were being including the direct and indirect impacts on other cleared as required per the forest working plans; two- sectors including e.g. transportation, infrastructure, thirds said controlled burning was not being regularly loss of environmental services, etc. Without credible, performed. Other than fire lines and controlled empirically-based estimates of the costs of forest fires, burning, less emphasis has been given to silvicultural it is unlikely that FFPM will be made more of a policy practices, such as selective thinning and planting priority. fire-adapted species. Officers commonly cited a need Strengthening Forest Fire Management in India   8 FIGURE ES.2: THE FOREST FIRE 7. Well-equipped and well-trained people on the PREVENTION AND ground are essential to forest firefighting MANAGEMENT (FFPM) CYCLE Forest fire suppression in India mainly involves dryland firefighting. Although the tools used in India may differ from those used in other countries, the PREVENTION principle of effective suppression remains the same: having a competent, well-trained, and adequately- equipped workforce on the ground, ready to respond and take immediate action. This workforce includes DETECTION field staff from the forest department as well as seasonally-employed fire watchers and volunteers from the local community. Only a handful of forest department officers surveyed and interviewed agreed SUPPRESSION that the equipment currently used in their area is adequate. Most cited a lack of basic safety gear and clothing, and a need for more training, especially for firewatchers and community volunteers. POST-FIRE MANAGEMENT 8. Post-fire management is not being treated as part of the FFPM process Source: Authors Post-fire management is not being treated as part of for greater participation by local forest-dependent the FFPM process and is probably the weakest link. communities in fire prevention. Post-fire data collection is an essential part of the fire management process and crucial to producing 6. India has developed robust detection systems for informed FFPM plans and policies. However, this forest fires part of the management process is given little priority and is often performed solely for the sake of fulfilling Over the past decade, India has emerged as a leading administrative requirements. Field reporting and example of how satellite technologies can be utilized the investigation of fire causes may be hindered by for the detection and monitoring of forest fires. Using insufficient field staff, difficult terrain, and a lack satellite data, Madhya Pradesh was the first state to of communications infrastructure in more remote develop an SMS-based system to alert field staff of areas. A lack of standard protocols for collecting and active fires burning in their area. Since then, Forest reporting information on fires, including their causes, Survey of India (FSI) has rolled out a nationwide has made it impossible to aggregate data across system. Satellite-based detection has helped fill a states. The greater issue, though, are the institutional gap left by under-resourced ground detection. As disincentives for accurate and complete reporting. these satellite systems continue to be upgraded, they Fires larger than a few hectares trigger extra work would benefit from greater integration, including for field staff to report and investigate offenses, the increased collection of field-based reporting and the department and its officers may be held for verifying satellite-derived fire alerts, as well as responsible for reported monetary damages due to improved data sharing between the states and FSI. fires. The states will need help from MoEFCC and the Only through systematic ground verification and research community in developing standard methods evaluation can the existing techniques for satellite and protocols for assessing ecological impacts and detection be improved. economic damages from fire. 9 Strengthening Forest Fire Management in India 9. More effective engagement of forest-using trivial, and fires started outside state forests may spread communities is essential… to state forests. Better coordination with communities and other forest land managers and more clearly More effective engagement of local communities—the defined responsibilities (including for the provision of primary forest users in India—is essential. Strategies funds) are needed. for FFPM should be founded on a clear recognition of how local communities depend on forests for Though large fires such as those observed in important goods and services and aim to ensure Uttarakhand in 2016 and Karnataka in 2017 do the delivery of these goods and services while also occur, forest fires are not typically treated as disasters, reducing damaging and unmanaged fires. Although and the disaster management authorities have so far all forest fires are treated as an offense under existing played a minor role in FFPM. A survey of the state laws, completely excluding the use of fires in forests disaster management agencies (SDMAs) revealed by local people is an unattainable goal. Thus, the SFDs a wide variation in how forest fires are treated in must strike a fine balance, working with communities disaster planning and how institutional mechanisms to make sure fire is used responsibly in a way that have been set up for organizing the response to large promotes forest health, while avoid damaging and or destructive fires. Thus, the point at which other out-of-control fires. agencies should be mobilized to assist the SFDs with forest fire suppression remains unclear, and the Forest officers interviewed and surveyed for this authority of the forest department to call on other study agreed that more effective engagement with assets in responding to forest fires is also limited. communities will hinge on better incentives. Existing incentives have included monetary rewards, the Researchers have been an underutilized part of the provision of jobs to community members, and access FFPM community. Stronger collaboration between the to harvest NTFPs from state forests. The Joint SFDs and research entities would enable states to better Forest Management Committees (JFMCs) have been monitor the ecological and economic impacts of fires, the primary avenue through which the SFDs have to develop robust protocols for gathering fire data, and offered such incentives. Monetary payments have innovate new science-based management approaches not typically been enough to cover the costs of fire for preventing fires and rehabilitating fire-affected areas. prevention work by the JFMCs but rather have served as a behavioral nudge. Seasonal firewatchers and In all the states visited and surveyed, forest departments community volunteers are rarely provided equipment have developed innovative ways to improve FFPM. and training for FFPM. From forest firefighting squads in Odisha, to fire risk zonation mapping in Telangana, to SMS-based fire 10. …as is coordination with other agencies and alerts in Madhya Pradesh, to community reserves in entities Meghalaya, to awareness-raising street performers in Uttarakhand, and so on, the examples abound. The SFDs manage about 654,137 km2 of forest lands However, states are often unaware of what their contained in reserved and protected forests, plus neighbors are doing, data and statistics are difficult much of the 113,881 km2 of unclassed forest. Together, to aggregate across states, and there is no formal these lands comprise about 23 percent of India’s mechanism for sharing knowledge about FFPM. geographical area (FSI 2018). Not all these areas are forest covered, and additional areas of forest cover exist outside the jurisdiction of the departments. In WHAT CAN BE DONE? practice, the SFDs often assume sole responsibility for forest fires on these non-department lands, often Detailed recommendations for improving FFPM managed by communities. National data on the forest based on the findings of this assessment are presented fires on non-department lands is lacking, though data in chapter 4 of the report. Table ES.1 presents a from Uttarakhand show that these lands accounted summary of recommendations organized by level of for about 35 percent of state-wide burnt forest area priority. The recommendations fall into five general in 2016. The threat of fire on non-SFD lands is non- thematic categories: Strengthening Forest Fire Management in India   10 1. Policy construction of watchtowers and crew stations and for frontline officers and seasonal firewatchers to spot At the national level, a cohesive first-order policy or fires is needed, as most of the areas surveyed reported action plan can set forth the guiding principles and shortfalls and field officers reported frequent delays framework for FFPM, beginning with a clear statement in making payments to seasonal firewatchers. People of goals and priorities. A national action plan would on the ground are the key to effective fire suppression also provide MoEFCC an opportunity to consolidate using dry techniques. In spite of the availability of its existing guidelines and the standing instructions it hi-tech equipment globally, the principal need is has issued over the years, and to issue comprehensive always to have a competent, trained, and equipped guidelines for a range of topics, including for the workforce on the ground, ready to respond and development of standard operating procedures take immediate action. Therefore, a top priority for (SOPs) by the states for various aspects of FFPM, for SFDs is to fill vacancies for field staff and community siting and maintenance of fire lines and controlled firewatchers. burning, standard protocols for post-fire reporting, and standard methods for assessment of damages. Additionally, their staff need to be trained, and this The national policy should also draw on climate activity too should begin immediately. The need for change policies given the clear overlap. A national greater training was almost unanimously mentioned level policy should also clearly delineate the respective among the officers surveyed and interviewed. roles and responsibilities of the MoEFCC, state forest Training should be provided to field officers, seasonal departments, and disaster agencies, and establish a firewatchers, and community volunteers involved mechanism for the provision of regular funding for in firefighting. The type of training provided to FFPM to the states. firefighters should be tailored according to their level of responsibility and role in the command structure The process of formulating the national policy or in responding to fires. Provision of training should action plan on FFPM would be just as important as extend beyond state-managed forests to community the policy or plan itself. The process should be open, institutions in regions such as the Northeast, where consultative, clearly defined, and time-bound. A core communities are responsible for managing most of the group with the Director General of Forests, MoEFCC, forest estate. and representatives from the SFDs, disaster agencies, NGOs, and research institutes should be established Lastly, forest fire prevention and management immediately to initiate the process for the development practices used by state forest departments also of the national policy and action plan over the course need to be strengthened. Only a few states have of one to two years. Guidelines to help establish the developed SOPs or manuals on standardized forest basic requirements for different aspects of FFPM can fire response systems. Such SOPs can cover a range of be drafted immediately, finalized in consultation with management practices including the more systematic relevant stakeholders, and later incorporated into the use of silvicultural practices, for example. There is a national policy. Similarly, coordination mechanisms at need for more systematic use of silvicultural practices the national, state, and district level, between forest such as selective thinning, pruning, and early- departments and disaster management agencies, could season controlled burning to reduce fuel loads, in be defined and established alongside the development areas managed by the forest department and those of the policy, and eventually brought under its scope. managed by other entities. SOPs can highlight where they should be applied, how local communities should 2. Staffing, capacity, and management practices be involved, and what measures should be put in place to ensure that they are conducted safely. Similarly, Inadequate resources and lack of sufficient staff on underreporting post-fires of causes, extent of burnt the ground have been cited repeatedly as reasons area, and economic damages needs to be addressed. for ineffective prevention, detection, suppression, One of the reasons for such underreporting is and post-fire practices. Even with the advent of new institutional disincentives (field officers who report remote sensing technologies, ground-based detection large fires may create additional work for themselves will continue to be essential. Greater funding for and their superiors in filing and prosecuting a 11 Strengthening Forest Fire Management in India forest offense, and the department may receive less 4. Community engagement financing). Management practices that hold officers accountable for the fulfillment of required prevention The total exclusion of fires from forests is not an and control activities, say, by including performance attainable or desirable goal for FFPM. Some fires of fire control duties in the annual evaluations of field can be beneficial, both from an ecological and social staff, can help remove the disincentives. point of view. There exists a fundamental tension between the total prohibition on fire under current 3. Technology law in India and the reality on the ground, as fire continues to be used as a landscape management tool Technologies available for improving FFPM range by communities of forest users across the country. A from the very high-tech to the very low-tech, from more effective policy for FFPM may begin with the new satellite and wireless sensor technologies for recognition that people will continue to use fire, that detecting forest fires, to self-fashioned jhapas for some fire is desired, and that the goal of FFPM should beating out fires. FSI has begun the development be to minimize the ecological, social, and economic of systems for early warning and fire danger rating, impacts of fire while ensuring that the benefits reaped and these efforts should be continued. Similarly, fire from fire may continue. From this starting point, fire alert systems developed by FSI and the states should managers may then work with communities to ensure be strengthened further. For one, the digitization that fire is used responsibly in a way that promotes of management boundaries by the state forest forest health, while seeking to avoid damaging and departments should be completed so that FSI can out-of-control fires. more accurately determine which fires to report and to whom. Additionally, ground verification data on If effective community involvement is to be garnered, satellite-based alerts should be collected by field staff, it is essential to work with communities and give them shared with FSI by the state forest departments, and a voice in the decision-making process. If they have analyzed, to determine the accuracy of satellite-based that, they will more likely feel included and be an alerts and thereby help improve the system. Fire alert effective part of the partnership. Forest officers who systems can also be improved by integrating ground- were interviewed and surveyed pointed to the need based detection with the satellite-based alert systems. for greater incentives as the most important way for Finally, the satellite-based detection systems should the forest department to increase the effectiveness of be expanded to include other forest areas beyond its engagement with communities on FFPM. Many department jurisdiction. Only a handful of field noted that the department already provides incentives officers surveyed agreed that firefighting equipment is to communities in their area. These incentives have adequate and sufficiently available in their area. Many taken a variety of forms, including wage labor, small pointed to the need for basic safety equipment and cash rewards, and public recognition for outstanding clothing. Some called for additional hand tools and performance. However, in many parts of the country, transport vehicles for field staff. current incentives have not been enough to mobilize communities as partners in FFPM. Stronger incentives Whether high-tech or low-tech, effective tools and may include securing forest tenure, resource rights, technologies must satisfy local financial, social, and and sharing revenues from commercial products such environmental constraints. Rather than prescribing as teak, sal, and bamboo, where allowable. specific fire-suppression tools to use in all the states, MoEFCC can promote the use of new technologies for 5. Data and information FFPM by supporting local research, encouraging states to experiment, and scaling up best practices, where Lastly, there is a need to support forest fire appropriate. International experience has shown management through improved data, research to that early warning and fire danger rating systems fill critical knowledge gaps, and regular knowledge developed with inputs from local fire managers and exchange. tailored to local conditions are more likely to be successful than systems that are imported directly Currently, nationwide information on forest fires from other contexts. in India is limited to satellite-based remote sensing Strengthening Forest Fire Management in India   12 data. The creation of a common classification fires on forest degradation in India, and methods for scheme for the causes of fire, standard reporting assessing such impacts signals the need for greater protocols, and standard methods for assessing involvement of the country’s research community burnt area would facilitate the creation of a national on FFPM. This would include public institutes and forest fire information database incorporating field- agencies, universities, and NGOs. The definition of a reported data. The database should also capture national research agenda for forest fires and provision information on fire lines, controlled burning, watch of funding opportunities for scientific research would towers, firefighting assets (and their locations), and be instrumental in bringing these entities together. communications infrastructure. Such a database would be instrumental for assessing longer-term India could, however, benefit from the development trends across states and regions and for planning fire of a mechanism to allow useful exchange between prevention and response. states. There is real need for a suitable forum where state representatives can regularly meet and swap India’s research community represents an invaluable ideas and information. Presently, each state forest asset for improving FFPM, though little formal department seems to operate in isolation from others. cooperation currently exists between members of the There are excellent initiatives developed by individual research community and the forest department. The states that could easily be transferred to and adopted still-limited knowledge about fire ecology in different by other states. A formal mechanism for knowledge forest types and climates, the longer-term impacts of sharing between states should be established. TABLE ES.1: SUMMARY OF RECOMMENDATIONS AND PRIORITIES Recommendation Lead Implementer Priorities and Timing FFPM Guidelines to cover: MoEFCC (in consultation with MoEFCC to begin drafting relevant stakeholders) these immediately, and to • Revised Working Plan Code finalize them in consultation • Development of Standard with relevant stakeholders. Operating Procedures (SOPs) by the SFDs (see below) • Fire lines, siting and maintenance. Controlled burning • Silvicultural practices (prevention and post-fire restoration or rehabilitation) • Common classification scheme for the causes of forest fires • Standard protocols for post- fire reporting, the investigation of fire causes, and standard methods for assessment of damages • Incentivizing accurate reporting by field staff on fires occurrence, burnt area, and damages 13 Strengthening Forest Fire Management in India Recommendation Lead Implementer Priorities and Timing Ensuring adequate funding and SFDs In the near term, states field staffing should examine existing budget resources to determine if enough is being allocated for FFPM. CAMPA offers a potential source of funding. In the longer term, states should seek to increase funding by increasing productivity of forests and thereby, the revenue generated from the sector. A top priority is for SFDs to fill vacancies for field staff and community firewatchers in fire-prone areas. Boots on the ground are essential for all aspects of FFPM, including prevention, detection, and timely response to fires. Training in fire suppression DFE (training curriculum) to be There is a real need for (prevention, detection, and post- rolled out in coordination with this, and this activity must fire reporting) for field staff SFDs begin immediately with the development of a curriculum for all forest guards and other field-level officers in the SFD. Provision of equipment for field SFDs in coordination with FRI There is a real need for this, staff and this activity must begin immediately. The focus should be on basic hand tools, safety gear and other equipment for ground crews that are appropriate and suited to local needs and conditions. Establishment of coordination MoEFCC at the national level, and This process should also mechanism, at national, state, SFDs at the state and district level, begin immediately, both and district levels, between working with relevant disaster to define the coordination forest departments and disaster management agencies mechanism and also to management agencies establish it. MOEFCC and NDMA should take the lead and provide guidance for the state-level mechanisms. Strengthening Forest Fire Management in India   14 Recommendation Lead Implementer Priorities and Timing Development and deployment of FSI with SFDs FSI to continue the Fire Danger Rating System (FDRS) development of FDRS in collaboration with SFDs, with the recognition that this is a long-term process. The immediate priority is to formalize this process and create a mechanism for SFDs to provide input to the FDRS and field data/ feedback for testing the FDRS. Continued improvement of FSI with SFDs FSI has a well-functioning satellite-based fire detection system nationwide satellite-based fire detection system in place. This system can be refined as new technologies and detection algorithms become available, and both FSI and SFDs should work toward this. The immediate priority is to improve two-way communication between FSI and SFDs and strengthen the process by which field-level forest officers provide feedback to both SFDs and FSI on the accuracy of the alerts. National Policy or Action Plan MoEFCC Core group with Director (which would also clarify role of General of Forest and other agencies) representatives from SFDs, NDMA, NGOs, and Research Institutes to be established immediately to initiate a consultative process for the development of the national policy and action plan over the course of one to two years. Incentivizing communities SFDs working with communities There is a real need for this, and local NGOs and this activity must begin immediately, although it will entail a longer-term process. 15 Strengthening Forest Fire Management in India Recommendation Lead Implementer Priorities and Timing Standard Operating Procedures SFDs in consultation with relevant SFDs to begin development agencies once MoEFCC issues guidelines. Defining a national research ICFRE ICFRE, as part of its agenda (with funding) mandate, has developed a National Forestry Research Plan for 2000-2020. FFPM research needs can be defined as part of this on- going process. Formal mechanism for knowledge MoEFCC MoEFCC organizes annual sharing between states meetings of PCCFs and one of these meetings can focus on forest fires. National Forest Fire Information FSI While such a database will Database serve many needs, it can be developed over the coming years once the underlying processes to collect the necessary data have been established. National Center of Excellence ICFRE in coordination with FSI While there is need for such a Center of Excellence, this too can be developed over the coming years, once the underlying processes have been established. Strengthening Forest Fire Management in India   16 INTRODUCTION FORESTRY SECTOR IN INDIA no taller than 20 meters, interspersed with shrubs, grasses, and other herbaceous vegetation. Tropical Forests cover 708,273 km2 or 21.54 percent of India’s moist deciduous forest accounts for the second largest land area (FSI 2018).3 Deforestation has gradually share of forest by area and occurs in all regions except slowed from an average annual rate of 4,795 km2 the Himalayas and the drier parts of the north and during 1930-1975 (Reddy et al. 2016) and has recently west (Ibid). begun to reverse, thanks to large-scale afforestation and reforestation efforts and a reorientation of Administratively, most of the country’s forests are national forest policy toward conservation (Nayak, contained on state-owned lands. Public lands classified Kohli, and Sharma 2013). Data published by Forest as forests in government records totaled 767,419 km2 Survey of India (FSI) in 2018 shows that forest cover in 2017, including 654,137 km2 of government lands grew by 6,778 km2 from 2015 to 2017 (FSI 2018). designated as reserved forest or protected forest per the India Forest Act of 1927 or the state forest acts (FSI India’s forests consist of a diverse range of forest types, 2018).4 Not all of these lands are actually covered by as depicted in maps 1 to 4—from the rainforests of forests, and additional areas of forest cover may exist the Western Ghats and northeastern states, to the outside them.5 Besides reserved forests and protected coniferous hill forests of the Himalayas, to the desert forests, there are also 113,881 km2 of unclassed forest scrub and thorn forests of Rajasthan. Tropical dry lands, which are documented in revenue records or deciduous forests comprise the largest share of forest various other state and local acts but are not necessarily by area and are spread across large parts of the Central managed by the state forest departments (FSI 2018). Highlands and Deccan Plateau in central and southern The Forest Survey of India (FSI) has estimated that India (Reddy et al. 2015). Much of this forest land about 14 percent of forest cover is held on private is characterized by open canopy, with trees typically lands, and that communities have management rights 3. Per the India State of Forest Report 2017 from which these estimates are drawn, Forest Survey of India (FSI) defines forest cover as “all lands more than one hectare in area with a tree copy of more than 10%, irrespective of land use, ownership and legal status” (FSI 2018: 5). Forest cover differs from recorded forest area, the latter of which refers to “all the geographic areas recorded as ‘Forests’ in government records” (Ibid). 4. Hunting, grazing, felling, fuelwood collection, and other extractive activities are prohibited in reserved forests unless specific permission is otherwise granted by the state to rights holders to conduct such activities. Protected forests include any forest lands owned and managed by the government that are not notified as reserved forests, including demarcated and un-demarcated protected forests. In practice, restrictions are generally less strict in these forests areas, and forest-dwelling and forest-fringe communities can hunt, graze their animals, and collect non-timber forest products, so long as these activities do not degrade the forest. 5. For the 16 states that have mapped and digitized their recorded forest area, FSI estimates that forest cover on government-classified forest lands was about 68 percent in 2017 (FSI 2018). 17 Strengthening Forest Fire Management in India to about 37 percent of public forest lands (FAO 2014). percent of all forests by area, and about 95 percent Community and privately managed forest is most of all forest plots inventoried show some signs of top common in the Northeast, where the state forest drying, girdling, illicit felling, blazing, lopping for departments only control a small portion of the total fodder, or other injuries to trees (FSI 2015). Only forested area. about 5 percent of natural forest remains intact (Reddy et al. 2017b). The degradation of forests Though India has succeeded in curbing large- leads to irreversible erosion, reduced soil fertility, scale deforestation, forest health across much diminished water catchment function, downstream of the country continues to show signs of strain. flooding, diminished biodiversity and additional Regeneration is inadequate or absent in about 45 rural poverty (Matta 2009). MAP 1: FOREST TYPES AND DISTRIBUTION IN SOUTH INDIA Source: Forest type data from Reddy et al (2015) Strengthening Forest Fire Management in India   18 MAP 2: FOREST TYPES AND DISTRIBUTION IN CENTRAL INDIA 19 Strengthening Forest Fire Management in India Source: Forest type data from Reddy et al (2015) MAP 3: FOREST TYPES AND DISTRIBUTION IN NORTHEAST INDIA Strengthening Forest Fire Management in India   20 Source: Forest type data from Reddy et al (2015) MAP 4: FOREST TYPES AND DISTRIBUTION IN NORTH AND WEST INDIA 21 Strengthening Forest Fire Management in India Source: Forest type data from Reddy et al (2015) Widespread forest degradation points to the immense to GVA understates the importance of forests because pressure on forests from people and the demand for many of the goods and services provided by forests are forest resources. The numbers are striking. As of 2015, not bought and sold in the formal economy and are about 92-103 million of India’s 1.31 billion people missing from India’s national accounts. People living lived in areas of forest cover.6 About one-quarter of in forest-fringe areas gather a wide variety of foods, all people in India rely on forests for at least part of materials, and medicines from forests—much of which their livelihoods.7 Nearly 200 million livestock are is used for subsistence by the household (World Bank dependent on forests for at least part of their diet, 2006). They may also sell these goods for cash income. either through stall feeding or grazing, and about 200 A survey of forest-fringe villages in Jharkhand found million people use fuelwood gathered from forests to that subsistence and cash income from forest goods satisfy their household energy needs (FSI 2011). This accounted for 12-42 percent of total household income huge demand for forest resources exists in the face of on average (Belcher, Achdiawan, and Dewi 2015). limited supply—the area of forest per capita in India Surveys of forest-fringe areas in other states have was less than one-twelfth the world average in 2015,8 found forest income shares as high as 40-60 percent and forest productivity is low, with stocking rates (Nayak, Kohli, and Sharma 2013). Other non-market at one-third the world average (MoEF 2009). Low services provided by forests may be reflected in the forest productivity is driven by a number of factors contribution of other sectors to the economy. For including lack of adequate resources and staffing, example, soil retention services from forests reduce lack of scientific management practices, and lack of the build-up of silt in hydropower facilities, improving engagement of forest-dependent communities operating efficiency, extending asset lifetimes, and increasing revenue from power generation. Vital to rural livelihoods, India’s forests are an undervalued pillar of the economy. The contribution The economic importance of forests is perhaps of the forestry sector to Gross Value Added (GVA) greatest for India’s rural poor. Subsistence and cash averaged only about 1.48 percent from 2011-2016, income from forest goods often account for a larger compared to 15.74 percent from farming and share of total income for the poorest households livestock.9 Most of the value added by the sector compared to better-off ones (Angelsen et al. 2014; came from industrial timber (60 percent), including Belcher, Achdiawan, and Dewi 2015).10 Forests also act from natural forest areas and plantations, followed by as a safety net, providing a source of supplementary firewood (37 percent). Yet, contribution of the sector employment, income, and nutrition during lean 6. Estimates are by the authors, using forest cover data for 2000 from Hansen et al. (2013), including areas of forest gain from 2000 to 2015. Areas of forest cover are at least 1 hectare in size and have a minimum of 10-percent canopy cover, as defined by FSI. The 30m x 30m forest cover data from Hansen et al. have been resampled to a resolution of 100 m x 100 m to be consistent with this. Population within forests is estimated in two ways, using two different source of gridded population data. First, high-resolution (100m x 100m) population data from WorldPop are overlaid on the forest cover extent, and the population within areas is summed. This produces an estimate of 92 million people. Second, estimates are also produced using the Gridded Population of the World v4 (GPW) data from the NASA Socioeconomic Data and Applications Center (SEDAC), which have a resolution of 1 km x 1 km. For this analysis, 1 km x 1 km plots in which at least half of the area has at least 10 percent forest cover are classified as forest. This produces an estimate of 103 million. The GPW population data are based primarily on district-level census data and assume that population within census areas is distributed evenly, thus may result in overestimates where population within forests is presumably less dense than in non-forests. See WorldPop, http://www.worldpop.org, and SEDAC, “Gridded Population of the World (GPW), v4,” http://sedac.ciesin.columbia.edu/data/collection/ gpw-v4. 7. Lynch (1992) first estimated in 1992 that 275 million people were partly or fully dependent on forests for their livelihoods. World Bank (2006) repeated this finding. MoEF (2009) has put the number of people who are partly or fully reliant on forests at 350-400 million. The basis for these figures is unclear, and these estimates should be taken as indicative at best. 8. Per capita forest cover in India in 2015 was 0.05 hectares versus 0.65 hectares for the rest of the world. Estimates for India are using FSI (2018) data; estimates for rest of the world are using UN Food and Agricultural Organization (FAO) data, available from World Bank World Development Indicators database at https://data.worldbank.org/. 9. Data are for fiscal years 2011-12 to 2015-16. Ministry of Statistics and Programme Implementation, Government of India, “Statement 1.6: Gross Value Added by economic activity” and “Statement 8.3: Output & Value Added from Forestry & Logging,” National Accounts Statistics 2017, http://www.mospi.gov.in/publication/national-accounts-statistics-2017-1. 10. However, as Angelsen et al. (2014) and Belcher, Achdiawan, and Dewi (2014) point out, forest income is typically much higher in absolute terms for households in the highest income groups. Strengthening Forest Fire Management in India   22 times, such as the slack period between agricultural The assessment of FFPM practices in India took harvests, and in response to shocks, such as a drought place from December 2016 to November 2017. A or an ill family member (see Wunder, Angelsen, and variety of quantitative and qualitative data on the Belcher 2014). The social and economic importance FFPM situation in India from a variety of primary of forests for the rural poor, including those living in and secondary sources were gathered. Primary data tribal areas, is explicitly recognized by India’s National collection focused on a sample of 11 states chosen Forest Policy of 1988.11 through consultation with MoEFCC for more in- depth field research: Andhra Pradesh, Assam, India has ambitious goals for improving forest cover Chhattisgarh, Himachal Pradesh, Jharkhand, Madhya and forest health. Under the National Action Plan Pradesh, Meghalaya, Odisha, Telangana, Tripura, on Climate Change (NAPCC), the government has and Uttarakhand. The selection of states aimed to committed to increase forest and tree cover by 50,000 represent different forest types, climates, geographies, km2 and to improve the quality of forest on another causes and patterns of forest fires, forest fire impacts, 50,000 km2 (MoEF 2008). In its Intended Nationally institutional arrangements for FFPM, and levels of Determined Contribution submitted to the UN technical capacity to ensure the broader applicability Framework on Climate Change in 2015, India has of findings at the national level. Logistical feasibility, committed to bringing 33 percent of its geographical the willingness of the states to participate, and existing area under forest cover and to creating additional contacts with a network of relevant stakeholders were sinks of 2.5 billion to 3 billion tons worth of CO2 also considered. stored in its forests by 2030 (GoI 2015). India has also set goals for improving the economic productivity of An initial scoping mission was held in two parts, forests, seeking to increase the forest-based livelihood between December 12 and 16, 2016 and again from income of about 3 million households (MoEF 2008). January 23 and February 3, 2017. During this first mission, the World Bank team visited the Forest Survey of India (FSI), and the states of Madhya STUDY OBJECTIVES AND Pradesh, Meghalaya, Telangana and Uttarakhand. METHODOLOGY In each of the states, the team interviewed forest department staff and community representatives It is unclear whether India can achieve its policy goals on the implementation of FFPM in their area. The for expanding forest cover and improving forest health purpose of the initial mission was to clarify the study if the prevention and management of forest fires is not objectives, identify potential challenges to FFPM that improved. India’s Ministry of Environment, Forest could be assessed through subsequent field work and and Climate Change (MoEFCC) has cited forest fire other primary data collection, and determine the as a “one of the major degenerating factors which not sample of states. only directly damage the forest cover, but also results in adverse ecological, economic and social impacts.”12 A second mission took place from May 11-19, MoEFCC has identified strengthening forest fire 2017, during which the World Bank team held prevention and management (FFPM) as a priority. technical discussions with FSI in Uttarakhand on the To inform Government of India’s efforts to improve development of a fire danger rating system and then forest fire prevention and management, this study conducted further site visits to fire-affected areas in documents current management systems, identifies Odisha and Jharkhand to meet with forest department gaps in implementation, and makes recommendations staff and community members and collect data for the on how systems can be improved. The assessment is assessment. intended to inform a national action plan on FFPM currently under preparation. See sections 3.5, 4.3.4.3, 4.3.4.4, 4.6, and 4.9 (MoEF 1988). 11. MoEFCC, preliminary project proposal to World Bank, August 2016. 12. 23 Strengthening Forest Fire Management in India An online survey with senior officers and field-level database on forest fires in India, satellite data are staff in the forest departments of the 11 states in the currently the best resource for the large-scale analysis sample was conducted between April and August 2017. of fires across different states and regions. Details on The survey gathered information on forest fire causes, the methods of analysis can be found in Annex 1. incidence, prevention, community engagement, and suppression in each of these states. More than 100 Finally, an international workshop was organized responses were received and were analyzed. Annex 2 in New Delhi in November 2017, bringing together provides the details of the survey. policymakers, foresters, scientists, and fire managers from eight countries, including Australia, Belarus, Data requests were also sent by MoEFCC to nodal Canada, India, Italy (FAO), Mexico, Nepal, and officers in the state forest departments in March 2017 the United States. The workshop aimed to identify to collect basic information about forest area, fire lines, relevant lessons from experience in other countries controlled burning, causes of fire, reporting of fire that could be applied to improve policies and incidents, and burnt area in each of the sampled states. practices for FFPM in India; share the initial results Data sheets were received from 7 states (Chhattisgarh, of the assessment; and build consensus among Indian Himachal Pradesh, Kerala, Meghalaya, Telangana, stakeholders as to needed areas for improvement in Tripura, and Uttarakhand). Details of the data request FFPM and recommendations for how these areas are given in Annex 7. could potentially be addressed. Additionally, interviews were conducted with Initial findings of the assessment were presented at the stakeholders in the disaster management authorities workshop to stakeholders in MoEFCC and the state and state governments at various points from forest departments. A draft of the assessment report February 2017 to August 2017. Responses from State was also circulated to the state forest departments and Disaster Management Authority (SDMA) officials other concerned agencies by MoEFCC for written were received from 5 states (Kerala, Madhya Pradesh, comment in January 2018. Odisha, Tripura, and Uttarakhand). Sample interview questions to SDMA officials are provided in Annex 8. Data collection for the assessment faced certain limitations. The study team has endeavored to Field-based community assessments were completed incorporate findings from the scientific literature, in two states, Meghalaya and Uttarakhand, from official statistics, and forest department reports from August to September 2017. In Meghalaya, 41 other states outside the sample of 11; however, the respondents from 5 districts spread across the state findings of the assessment may not reflect the specific were consulted. In Uttarakhand, respondents from circumstances in other states not included in the sample. 10 villages were consulted. Frequency and cause of The assessment is also limited by the availability and forest fires were discussed with the forest dependent quality of existing data on forest fires in India. Data on communities vis-à-vis various factors such as forest burnt area and damages caused by forest fires were not type, ownership pattern, availability and control provided by all the states in the sample.13 over resources by community, and so on. Structured questionnaires and focus group discussions were used to collect community perceptions on protecting forests STRUCTURE OF THE REPORT from fires. A description of the appraisals and findings are presented in Annex 3. Chapter 1 presents an analysis of forest fire characteristics in India, including spatial patterns, Geospatial analysis of forest fire characteristics, temporal trends, and factors influencing fire potential patterns, and trends across India (including in states and behavior. The chapter also discusses the central other than the 11 in the sample for fieldwork) was role played by people in shaping the forest fire regime performed in August and September 2017 using in India, the impacts of forest fires on forest ecology, satellite remote sensing data. For lack of a national and the economic costs of fire. Per MoEFCC’s request, the assessment also did not include an independent evaluation of the impacts and costs of forest fires. 13. Strengthening Forest Fire Management in India   24 Chapter 2 delves into an assessment of FFPM in India department and the disaster management authorities. today, beginning with policies at the national, state, Local communities of forest users represent the main and local level. The chapter then evaluates the on- pillar of FFPM, and the chapter discusses the various the-ground implementation of FFPM at each stage of ways the forest departments have reached out to them. the prevention, detection, suppression, and post-fire As the chapter concludes, the FFPM community also management cycle. includes researchers, who have been underutilized so far but have much to offer in improving knowledge of Chapter 3 discusses working with other agencies and forest fires and FFPM. communities on FFPM. Other agencies with roles in FFPM include public land managers outside the forest Chapter 4 presents conclusions and recommendations.   25 Strengthening Forest Fire Management in India CHAPTER ONE CHARACTERIZING THE FOREST FIRE CHALLENGE IN INDIA 1.1 OVERALL PATTERNS AND Comparing the number of active fire detections versus TRENDS IN FOREST FIRES14 the total area of forest affected by fire in figures 1.1 and 1.2 below, some distinct regional patterns emerge. Each year, fires affect forests across much of India. In the figures, the size of each rectangle represents the According to satellite detections of forest fires by the total number of active fire detections or burnt area per Moderate Resolution Imaging Spectroradiometer district from 2003 to 2016. Colors represent regions. (MODIS), from 2003 to 2016, as few as 380 and as The figures show that while the Northeast experiences many as 445 of the country’s 647 districts experienced the most frequent fires, fires tend to be concentrated fires each year (i.e. at least in 59 percent, but as many in a smaller area that is subject to repeat burning. This as 69 percent of districts).15 cyclical pattern of burning on small plots of forest is consistent with the practice of shifting cultivation Yet, some areas exhibit a much higher incidence of fire (jhum) that is seen throughout the Northeast. By than others. Just 20 districts, representing 3 percent of contrast, fires in other regions, particularly districts the India’s land area and 16 percent of the country’s in Central and Southern India, are more expansive. forest cover in 2000, accounted for 44 percent of all Districts experiencing widespread and frequent forest forest fire detections from 2003 to 2016. Similarly, the fires include areas of dry and moist deciduous forest top-20 districts in terms of area affected by fire from 2003 to 2016 account for 48 percent of the total fire- in the borderlands of Chhattisgarh, Maharashtra, and affected area,16 despite having just 12 percent of the Telangana that are affected by fire on a nearly annual nation’s forest cover in 2000 and 7 percent of its land basis (map 1.1, below). The Western Himalayas, which area (tables 1.1a and 1.1b, below). The top-20 districts experienced an especially severe fire season in 2016, in terms of fire frequency are mainly located in the account for a relatively small share of total burnt area Northeast, while the top-20 districts in terms of burnt and forest fire detections over the longer timeframe area are mainly in Central India.17 analyzed. 14. The analysis of forest fire characteristics and trends presented in this chapter draws primarily on detections of active fires by the Moderate Resolution Imaging Spectroradiometers (MODIS) aboard the Aqua and Terra satellites. Analysis was performed for 2003- 2016, the years for which complete MODIS data from both the Aqua and Terra satellites were available at the time of writing. MODIS data on active fires were processed and provided by Forest Survey of India (FSI). For a detailed explanation of data and methods, see Annex 1. The analysis was performed by Christopher Sall. 15. The administrative boundaries and number of districts used for the analysis are as of 2012. 16. Fire-affected area in table 1.1b includes any area that was under forest cover in the year 2000 (at least 10-percent canopy cover) and which was affected at least once by fire between 2003 and 2016. Fire-affected area is estimated using the standard science-quality data product for monthly burnt area (“MCD45A1”) provided by NASA and the University of Maryland (United States), which is derived from MODIS and has a spatial resolution of 500 m. Forest cover data are from Hansen et al. (2013). See Annex 1 for details. 17. For regional definitions refer to Section 1 of Annex 1. Strengthening Forest Fire Management in India   26 TABLE 1.1A: TOP 20 DISTRICTS BY TOTAL NUMBER OF FIRE DETECTIONS, 2003-2016 No District, State, Region Fire detections, Share of fire Share of total 2003-2016 detections, 2003- forest cover, 2000 (number) 2016 (%) (%) 1 Lunglei, Mizoram, NE 13,453 3.82 0.87 2 Karbi Anglong, Assam, NE 12,238 3.48 1.71 3 Dima Hasao, Assam, NE 11,608 3.30 0.91 4 Churachandpur, Manipur, NE 11,068 3.15 0.87 5 Mamit, Mizoram, NE 9,005 2.56 0.58 6 Lawngtlai, Mizoram, NE 8,501 2.42 0.43 7 Tamenglong, Manipur, NE 8,163 2.32 0.79 8 Aizawl, Mizoram, NE 6,705 1.91 0.61 9 Gadchiroli, Maharashtra, C 6,264 1.78 1.56 10 Dhalai, Tripura, NE 6,234 1.77 0.40 11 Champhai, Mizoram, NE 5,940 1.69 0.64 12 W. Khasi Hills, Meghalaya, NE 5,220 1.48 0.88 13 Narayanpur, Chhattisgarh, C 5,098 1.45 0.78 14 Ribhoi, Meghalaya, NE 4,835 1.37 0.43 15 Kandhamal, Odisha, C 4,753 1.35 1.09 16 E. Garo Hills, Meghalaya, NE 4,687 1.33 0.50 17 Ukhrul, Manipur, NE 4,645 1.32 0.78 18 Chandel, Manipur, NE 4,628 1.32 0.56 19 Bijapur, Chhattisgarh, C 4,615 1.31 1.19 20 North Tripura, Tripura, NE 4,087 1.16 0.33 Top 20 total 141,747 40.29 15.91 Notes: C = Central; NE = Northeast; S = South Data source: MODIS monthly data product for active fires (MCD14ML), provided by Forest Survey of India; MODIS monthly data product for burnt area (MCD45A1); forest cover data for 2000 from Hansen et al. (2013); district boundaries as of 2012 TABLE 1.1B: TOP 20 DISTRICTS BY TOTAL AREA AFFECTED BY FIRE, 2003-2016 No District, State, Region Fire affected Share of burnt Share of total area, 2003-2016 area, 2003-2016 forest cover, 2000 (km2) (%) (%) 1 Gadchiroli, Maharashtra, C 4,106 8.24 1.56 2 Bijapur, Chhattisgarh, C 2,633 5.29 1.19 3 Khammam, Telangana, S 1,923 3.86 1.13 4 Narayanpur, Chhattisgarh, C 1,346 2.70 0.78 5 Warangal, Telangana, S 1,273 2.56 0.45 6 Koriya, Chhattisgarh, C 1,169 2.35 0.42 7 Adilabad, Telangana, S 995 2.00 0.39 8 Chandrapur, Maharashtra, C 970 1.95 0.31 9 Surguja, Chhattisgarh, C 948 1.90 0.79 27 Strengthening Forest Fire Management in India No District, State, Region Fire affected Share of burnt Share of total area, 2003-2016 area, 2003-2016 forest cover, 2000 (km2) (%) (%) 10 Kurnool, Andhra Pradesh, S 895 1.80 0.23 11 Amravati, Maharashtra, C 888 1.78 0.23 12 Y.S.R., Andhra Pradesh, S 854 1.71 0.32 13 Prakasam, Andhra Pradesh, S 849 1.70 0.31 14 Dakshin Bastar Dantewada, Chhattisgarh, C 803 1.61 0.73 15 Bilaspur, Chhattisgarh, C 799 1.60 0.36 16 Raipur, Chhattisgarh, C 777 1.56 0.50 17 Betul, Madhya Pradesh, C 727 1.46 0.29 18 Champhai, Mizoram, NE 707 1.42 0.64 19 Lawngtlai, Mizoram, NE 673 1.35 0.43 20 Dima Hasao, Assam, NE 665 1.34 0.91 Top 20 total 24,000 48.18 11.97 Notes: C = Central; NE = Northeast; S = South Data source: MODIS monthly data product for active fires (MCD14ML), provided by Forest Survey of India; MODIS monthly data product for burnt area (MCD45A1); forest cover data for 2000 from Hansen et al. (2013); district boundaries as of 2012 FIGURE 1.1: SHARE OF ACTIVE FOREST FIRE DETECTIONS BY MODIS PER DISTRICT AND REGION, 2003-2016 Notes: “…” = multiple districts with few observations grouped together; individual rectangles represent districts; same-colored districts are grouped into regions; size of rectangle is proportional to the number of fire occurrences from 2003-2016. Data source: MODIS monthly data product for active fires (MCD14ML), provided by Forest Survey of India Strengthening Forest Fire Management in India   28 FIGURE 1.2: DISTRIBUTION OF FORESTED AREA AFFECTED BY FIRE PER DISTRICT AND REGION, 2003-2016 Notes: “…” = multiple districts with few observations grouped together; individual rectangles represent districts; same-colored districts are grouped into regions; size of rectangle is proportional to the number of fire occurrences from 2003-2016. Data source: MODIS monthly data product for burnt area (MCD45A1); forest cover data for 2000 from Hansen et al. (2013) MAP 1.1: FORESTED AREAS AFFECTED BY WIDESPREAD AND FREQUENT BURNING IN CENTRAL INDIA Source: MCD45A1 burnt-area data from NASA and University of Maryland; forest cover from Hansen et al. (2013) 29 Strengthening Forest Fire Management in India According to FSI, southern dry mixed deciduous for forest fires is improving or worsening over time. forest, dry teak forest, and northern dry mixed At the national scale, observations by MODIS of active deciduous forest were among the forest types most fires do not show a consistent increase or decline in affected by fires in 2016 (FSI 2016). These forest types fire incidence since 2003. At the subnational level, are prevalent across Andhra Pradesh, Chhattisgarh, trends are more varied. Table 1.2 illustrates year-on- Madhya Pradesh, Maharashtra, and Telangana. Moist year changes in the number of active fire locations per deciduous forest, which is most characteristic of states state from 2003 to 2016. The coefficients in the table in the Northeast but also occurs in other regions, represent the average annual percent change by state accounted for about 33 percent of the total forest area in the number of fires during the peak 7-, 14-, and affected that year (Ibid). 30-day period during forest fire season.18 Statistically significant decreases are indicated by blue, while The NRSC scientists have also found evidence of fires orange indicates a significant increase. affecting forests in areas of significant ecological value, especially for biodiversity conservation (Reddy et al. According to the table, statistically significant decreases 2017a). Between 2006 and 2015, the authors report in fire frequency occurred in only a couple of states in that forest fires were detected in just under half (281 the Northeast: Mizoram and Tripura. The observed of 614) of the protected areas in India. In the year decline in fire frequency in the region may be due to 2014, fires burned about 8.6 percent of forest cover in a gradual shift away from traditional jhum practices.19 protected areas. Statistically significant (though small) increases have been observed in Andhra Pradesh, Assam, Odisha, Due to the limited historic record, there is little Telangana, and West Bengal. evidence to determine whether the overall situation TABLE 1.2: YEAR-ON-YEAR TREND IN FIRE DETECTIONS BY STATE, 2003-2016 (ANNUAL PERCENT CHANGE) Increase, significant at Increase, significant at Decrease, significant at Decrease, significant at 95% level 90% level 90% level 95% level State Annual 7-day max 14-day max 30-day max Andhra Pradesh 4.4 0.4 1.0 2.1 Arunachal Pradesh 0.1 0.1 0.7 1.4 Assam 2.6 0.0 0.9 1.4 Bihar 1.9 0.4 0.6 1.4 Chhattisgarh 2.7 0.2 0.3 1.3 Gujarat -0.2 -0.1 -0.3 -0.2 Haryana 0.0 0.0 0.0 0.0 Himachal Pradesh 0.6 0.3 0.4 0.4 Jammu and Kashmir 0.1 0.2 0.2 0.3 Jharkhand 1.9 0.2 0.5 1.2 Karnataka 1.3 0.1 0.2 0.4 Kerala -0.6 -0.1 -0.2 -0.7 Madhya Pradesh 0.5 0.0 -0.3 -0.6 18. See Section 1 of Annex 1. 19. Comments made by Mr. Lal Ram Thanga, Principal Chief Conservator of Forests, Government of Mizoram, at the International Workshop on Forest Fire Prevention and Management, New Delhi, India, 1 November 2017. Strengthening Forest Fire Management in India   30 State Annual 7-day max 14-day max 30-day max Maharashtra 0.8 0.1 0.1 0.3 Manipur -1.3 -0.2 0.2 -0.3 Meghalaya 1.2 0.2 0.9 0.7 Mizoram -5.2 -0.4 -0.4 -1.5 Nagaland 1.7 -0.1 0.5 0.8 Odisha 0.2 0.2 0.6 1.5 Sikkim 0.0 0.0 0.0 0.0 Tamil Nadu -0.4 -0.2 -0.2 -0.6 Telangana 3.8 0.1 0.4 1.1 Tripura -1.9 -0.7 -1.5 -1.7 Uttar Pradesh 0.6 -0.3 -0.5 -0.7 Uttarakhand -2.1 0.1 0.1 0.2 West Bengal 2.0 0.1 0.4 1.2 Notes: “Annual” = change in total number of fire detections during forest fire season from January to June; “7-day max” = change in number of fire detections during peak 7-day period during fire season; “14-day max” = change in detections during peak 14-day period; “30-day max” = change in detections during peak 30-day period Data sources: MODIS monthly data product for active fires (MCD14ML), provided by Forest Survey of India 1.2 CAUSES OF FOREST FIRES AND is crucial for developing policies and plans to enhance the resilience of forests at the landscape, regional, or FACTORS INFLUENCING FIRE national level. BEHAVIOR India’s monsoons are largely responsible for the Forest fires result from a combination of natural and seasonal nature of forest fires in the country. For most social factors. The forest fire triangle in figure 1.3 illustrates how these factors are interrelated. As shown FIGURE 1.3: THE WILDFIRE TRIANGLE by the triangle, topography, weather, and fuel—the corners of the triangle—influence the potential for intensive fire behavior and spread. At the center of the triangle are people. 1.2.1 Weather Topography Fire intensity and behavior are intricately linked to Terrain, aspect, exposure, weather and climate. Day-to-day weather influences accessibility the likelihood that fires will ignite, grow, and spread. Seasonal weather patterns influence the onset, People duration, and severity of the fire season. Over the Current and past land longer term, shifts in climate caused by anthropogenic management practices, forest resource use, global warming may further alter India’s forest fire use, policies landscape and fire regime. and regulations, Fuel Weather etc. Understanding how weather influences forest fires Living and dead Temperature, vegetation, is fundamental to developing seasonal forecasts of rainfall, humidity, organic soil fire season severity and quantifying and predicting winds material fire danger from day to day. Assessing the possible effects of longer-term climate change on fire regimes Source: Authors, adapted from Roy (2004) and Schnepf et al. (2010) 31 Strengthening Forest Fire Management in India of India, forest fires peak during the dry months of drying has been observed over much of the Indian March or April before the arrival of the monsoon (FSI subcontinent (Guhuthakurta et al. 2014). Although 2012). As can be seen in figure 1.4, the fire season part of this decrease in monsoon rainfall may be mainly occurs during the four-month period between explained by multidecadal variability, research has February 15 and May 15. The figure illustrates the also linked the drying to rapid warming of the Indian seasonality of forest fires by showing how fires in each Ocean in contrast with more subdued warming state are distributed across the months of the year. over land (Roxy et al. 2015), possibly as a result of The lengths of the blue-colored violin plots show higher emissions of aerosols from burning biomass the continuous period between September 1 and and fossil fuels (Ganguly et al. 2012; An et al. 2014). August 31 the following year in which 80 percent of This reduced contrast in land-sea temperatures has all fires are concentrated. As seen in the figure, the weakened the engine that drives the monsoon. It is peak fire season is the most concentrated (shortest) in not clear, however, how the drying of the monsoon has the Northeast and the Northern state of Bihar. The affected the intensity or frequency of forest fires. season is the longest in the Western Himalayan states of Uttarakhand, Himachal Pradesh, and Haryana, District-level analysis for 2003 to 2016 suggests that where the season exhibits a bimodal distribution, monsoon rainfall provides an early warning of the with one peak in late April and another in late May. next year’s fire season severity.21 A district in which High-altitude areas of the Western Himalayan states rainfall is one standard deviation above the long-term may also experience fires during the months of average for the months of June to August or July to October to December due to pasture clearance. In the September will typically experience 7-12 percent Northeastern states, fires observed in December and fewer fires from January to May the following year (the January could be due to the preparation of land for fire season before the arrival of the subsequent year’s jhum cultivation. The apparent peak in fires during monsoon). If rainfall continues to be one standard October and November in Punjab may be due to deviation above average over the longer period of July wheat stubble burning on adjacent farm fields and not to December, then the average district will experience actually due to forest fires.20 about 21 percent fewer fires. Though the monsoon is the primary determinant of Whereas monsoon rainfall and ENSO have been when the fire season occurs, figure 1.4 also reveals suggested as early indicators of fire season severity (see how people influence the seasonality of forest fires. box 1.1), weather conditions in the summer months In parts of the Northeast (Nagaland and Arunachal serve as more immediate predictors of shorter-term Pradesh), satellite fire detections may occur as early as fire danger in the coming days or weeks. Detailed December or January when people set fires to obtain district-level analysis using monthly weather data and certain non-timber forest products (NTFPs) such as satellite fire detections reveals that weather during thatch grass, broomsticks, flamengia, and wild tubers the previous weeks or months in the summer can around that time (to be further discussed in section potentially negate the longer-term effects of above- or 1.2.4 below). For Punjab, the figure shows that fires below-average monsoon rainfall during the previous peak in early November around the time that farmers year.22 One additional wet day during the summer burn rice stubble in their fields before planting (defined as a day with > 0.01 mm precipitation) can wheat; however, in this case it is unclear if the satellite reduce the odds of a fire being detected during the observations of forest fires around this time represent present month by almost 16 percent. A 1°C increase fires on forest lands or on adjacent farm fields. in mean monthly temperature, on the other hand, can raise the odds of fire by 12 percent. The analysis There is a voluminous body of scientific work also shows that a marginal increase in precipitation or examining long-term trends in India’s monsoon, wet days in previous months without higher rainfall yet little research has been done on how shifts in the during the current month can also lead to higher monsoon have affected forest fire seasons. During odds of fire. One possible explanation is that higher the second half of the twentieth century, widespread precipitation in earlier months may stimulate the 20. E. Vikram, FSI, comments to authors, February 2018. 21. See Section 2.1.2 of Annex 1 for details. 22. See Section 2.1.4 of Annex 1 for details. Strengthening Forest Fire Management in India   32 FIGURE 1.4: SEASONALITY OF FOREST FIRES BY STATE AND REGION Western states Gujarat Western Himalayan states Uttarakhand Jammu & Kashmir Himachal Pradesh Southern states Telangana Tamil Nadu Kerala Karnataka Andhra Pradesh Northeastern states Tripura Nagaland Mizoram Meghalaya Manipur Assam Arunachal Pradesh Northern states Uttar Pradesh Punjab Bihar Central states West Bengal Odisha Maharashtra Madhya Pradesh Jharkhand Chhattisgarh 15 15 15 15 5 15 15 15 5 15 15 5 Au 1 r1 1 1 1 1 1 c1 1 c1 1 1 l1 r1 l1 g n ct b ov ar ct ay n n n p ay ov ar b g p Ap Ju Au De De Fe Ju Ap Ju Ja Fe Ju Se Ja O O Se M M N M M N Notes: red line indicates start of year (Jan 1); shaded area represents the period during which 80 percent of forest fires are detected by MODIS in that state; the line within the shaded area is the interquartile range; states are grouped by regions; data are for September 1, 2002 to August 31, 2016; Andaman and Nicobar Islands, Chandigarh, Dadra and Nagar Haveli, Goa, Haryana, Puducherry, Rajasthan, and Sikkim due to insufficient observations. Data source: MODIS monthly data product for active fires (MCD14ML), clipped to forest cover for 2000 (Hansen et al. 2013). growth of grasses and other vegetation and increase Drought is another useful predictor. One way of the availability of fine fuels later (more on this in quantifying the relationship between drought and section 1.2.3). If rainfall continues to be higher than fire potential is the Keetch-Byram Drought Index normal, then the added moisture during the present (KBDI), which FSI is currently considering as an month will decrease the odds of fire. A takeaway from element of its fire danger rating system. The KBDI this finding is that while monsoon rainfall may be a measures the deficit of moisture in the upper soil or useful early indicator of the potential for a severe fire duff layer of a forest. Higher KBDI values indicate season, forecasts and predictions must continually be a lack of available water, leading to the enhanced updated with more near-term indicators of fire danger flammability of fine fuels such as dried-out grasses as the season evolves. and decaying organic material such as buried roots 33 Strengthening Forest Fire Management in India Box 1.1: The El Niño/Southern Oscillation (ENSO) and Forest Fire Season Severity There has been some discussion on ENSO as a good advance predictor of fire season severity at the state or national level (Standing Committee 2016). The logic is that during El Niño years, warmer sea surface temperatures in the Pacific displaces the circulation of upper air flow across the tropics, as dry and stable air descends on the Indian subcontinent and reduces monsoon rainfall. Warmer, drier winters then lead to higher fire danger during the following summer (January to June). FIGURE B1.1: EL NIÑO/LA NIÑA EVENTS AND MONSOON RAINFALL, 1930-2015 2011 20 1988 2013 1975 Monsoon rain % departure from 30 - year average 2010 10 2007 1955 2008 1954 2012 2015 1950 0 1997 2014 -10 2009 1965 1987 -20 1972 -2 -1 0 1 2 Niño 2.4 index, June-September La Niña Neutral EI Niño Notes: Niño3.4 Index compares monthly equatorial sea surface temperatures in the central Pacific (5°N-5°S, 150°W-90°W) against a running 30-year climatological average, for the months of June-September; monsoon rainfall also compared against running 30-year climatological average. Data source: US NOAA, http://origin.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/detrend.nino34.ascii.txt; monsoon precipitation data from University of East Anglia Climate Research Unit, available at World Bank, Climate Change Knowledge Portal, http://sdwebx.worldbank.org/climateportal/ A systematic relationship between sea surface temperatures and wildfires could provide a basis for longer-term forecasting of the severity of the fire season.23 Yet, the link between ENSO, monsoon rains, and forest fires in India is far from straightforward. First, strong El Niño events are not always associated with drought, and the correlation between ENSO and monsoon rains has weakened in recent decades (An et al. 2015). As shown in figure B1.1, one of the strongest El Niño years on record, 1997, was accompanied by above-average monsoon rainfall, while some of the worst drought years in the late twentieth century occurred during relatively mild El Niño events. Research suggests that In the western United States and Canada, for example, researchers have observed a strong correlation between sea surface temperatures 23. and seasonal wildfire activity in the western United States and Canada (Barbero et al. 2015; Hess et al. 2001; Swetnam and Betancourt 1990). Strengthening Forest Fire Management in India   34 the effect of El Niño on the monsoons in India depends more on the location of the warming in the Pacific than on the severity of the El Niño itself. El Niño events marked by warmer seas in the central equatorial Pacific are more likely to produce drought in India than events with warming concentrated in the eastern Pacific (Kumar et al. 1999, 2006). Second, bad fire years do not always occur following drier-than-normal monsoons. Figure B1.2 compares years with above- or below-average precipitation with the number of forest fire detections during the following summer. Three of the six drier-than- average monsoons to occur between 2003 and 2016 were followed by milder than average fire seasons. The two worst fire seasons, in 2009 and 2012, occurred after above-average winter rainfall. Statistical analysis performed at the national, state, and district level does not provide any evidence of a systematic link between ENSO and fire season severity.24 FIGURE B1.2: MONSOON RAINFALL AND NEXT-YEAR FIRE SEASON SEVERITY, 2003-2016 2009 2 Fire detections, standardized anomaly 2012 1 2010 2016 2004 2006 2007 0 2008 2014 2005 2013 2011 2003 -1 2015 -2 -1 0 1 2 Monsoon rainfall, standardized anomaly Drier than normal Wetter than normal Notes: Fire detection average is for the months of January-May from 2003-2016; monsoon rainfall standardized anomaly is calculated against 30-year running climatological average. Data source: MODIS monthly data product for active fires (MCD14ML), provided by FSI; monsoon precipitation data from University of East Anglia Climate Research Unit, available at World Bank, Climate Change Knowledge Portal, http://sdwebx. worldbank.org/climateportal/ or wood (Keetch and Byram 1968). As originally severe drought. Because of the high temperatures formulated, the KBDI is calculated on an 800-point during the dry summer season across much of India, scale, where each point represents 1/100 inch of conditions can easily progress from saturated soils to additional rainfall necessary to restore soils back to a severe drought within a month or two. Inputs to the saturated state. A KBDI value of 0 indicates that soils KBDI include daily rainfall, mean annual rainfall, are saturated, while a KBDI value over 700 indicates and daily maximum temperature. See Section 2.1.3 of Annex 1 for details. 24. 35 Strengthening Forest Fire Management in India Figure 1.5 depicts KBDI values for districts, by regions, likely continue to shift. Combined with habitat on days when fires were detected versus on days when fragmentation and resource extraction, the effects of no fires were detected in the peak months of the fire climate change are likely to put many species under season (February to May). The figure reveals that on greater pressure and weaken their ability to withstand most of the days where forest fires were detected, disturbances such as fire. The impact on forest fire KBDI values were above 650-700. The discrepancy in frequency is yet to be fully understood (Settele et al. KBDI values on days with and without fire was greatest 2014). in the Northeast and Western Himalayan regions. The discrepancy was smallest in the Western states, where The IPCC finds that there is high confidence that the climate is generally much drier, forest is sparser, fires in moist tropical forests are becoming more and fires are less frequent. This suggests that KBDI frequent and severe throughout much of the world may be a useful operational indicator of adverse fire due to interactions between drought and land use, conditions in the Northeast and Western Himalayan that lead to reduced moisture content of fine fuels states, but not the Western states. and lower resistance to fire (Settele et al. 2014). Dry tropical forests are also increasingly under pressure Statistical analysis was then performed to relate the from climate change, deforestation, fragmentation, daily odds of fire occurrence during the peak fire and fire. One study of the effects of climate change on seasons to the drought stage, as indicated by the KBDI tropical dry forests in South Asia cited by the IPCC, scale. The analysis is described in detail in section 2.1.5 for example, finds that by the end of this century of Annex 1. The results are illustrated in figure 1.6, most of India’s dry forests are projected to experience which shows how the predicted probabilities for fire climate conditions beyond the envelope that they can detection on any given day during the peak period from tolerate. Global simulations of fire frequency under February to May increase at the margins as drought the SRES A1B, A2, and B1 emissions scenarios have worsens. The figure reveals how the probability of projected increases in landscape fires across much fire occurrence is much more sensitive to drought of India’s central highlands, the Gangetic plain, and conditions in some regions. In the absence of drought, much of the northern states. However, the IPCC finds when soils are saturated (KBDI 0), there is less than there is generally low agreement about how climate a 1 percent chance of a fire occurring on any given change will affect the frequency or severity of fire in day in any of the regions. As upper soil layers dry out specific locations (Settele et al. 2014: 303-308). Much and the KBDI rises, the differences in the likelihood of the uncertainty in India lies in whether the effects of fire become more pronounced. At KBDI 600, there of higher temperatures will be offset by changes in is about a 21-25 percent chance of fire occurrence in precipitation (Joseph et al. 2009). the districts of the Northeast and Western Himalayas, compared to a 7-10 percent chance in the Central Despite the uncertainties, the interconnected risks and Southern regions, and a 1 percent chance in the posed by longer-term climate change and forest fires districts of the West and North. At the upper end have been widely recognized, as reflected in India’s of the KBDI scale (KBDI > 600), the likelihood of state action plans on climate change. The State fire increases dramatically. In the Northeastern and Strategy and Action Plan on Climate Change for Western Himalayan states, the predicted probability Himachal Pradesh (2012), for instance, highlights that of fire increases to 46-48 percent at KBDI 750. In the the occurrence of forest fires in the state may increase Central and Southern states, the probability doubles to as a result of climate change. The Assam State Action around 17-22 percent. These larger increases suggest Plan on Climate Change (2015) notes that forest fires that reaching drought stages 6 or 7 on the KBDI scale may become the norm with longer dry periods, which may serve as a good operational indicator for high would in turn impact the livelihoods of the people fire potential across a variety of forest environments dependent on timber and non-timber forest produce. in India, outside of the West and North, where the The Kerala State Action Plan on Climate Change incidence of fire is less common in general and less (2014) points to the lengthening of the dry period sensitive to the effects of drought. during the summer, which has resulted in a higher incidence of fires denuding forests and disturbing As India’s climate continues to change with human- the associated watersheds. Jharkhand’s Action Plan driven global warming, scientists expect that the on Climate Change (2014) points to higher expected boundaries and areas of different forest types will forest fire risks, which could pose a threat to mining Strengthening Forest Fire Management in India   36 FIGURE 1.5: DISTRIBUTION OF KBDI VALUES IN DISTRICTS ON DAYS WITH OR WITHOUT FIRES DURING PEAK FOREST FIRE SEASON (FEBRUARY-MAY), 2012-2016 Central North .008 .02 .002 .004 .006 .005 .01 .015 Kernal Density Kernal density 0 0 0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800 Keetch - Byram Drought Index Keetch - Byram Drought Index With Fire Without Fire With Fire Without Fire Northeast South 0 .001 .002 .003 .004 .005 .006 Kernal density Kernal density .002 0 .004 0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800 Keetch - Byram Drought Index Keetch - Byram Drought Index With Fire Without Fire With Fire Without Fire West Western Himalaya .006 .015 Kernal density Kernal density .004 .005 .01 .002 0 0 0 100 200 300 400 500 600 700 800 0 100 200 300 400 500 600 700 800 Keetch - Byram Drought Index Keetch - Byram Drought Index With Fire Without Fire With Fire Without Fire Notes: KBDI = Keetch-Byram Drought Index; kernel density indicates the density or bunching of observations in a sample around different values in the population distribution. Data source: NOAA gridded daily climate datasets; MODIS monthly data product for active fires (MCD14ML). 37 Strengthening Forest Fire Management in India FIGURE 1.6: PREDICTED DAILY PROBABILITY OF FIRE DETECTION AT DIFFERENT LEVELS OF DROUGHT, BY REGION 0.6 0.5 Predicted probability of fire detection 0.4 Central North 0.3 Northeast South 0.2 West W. Himalaya 0.1 0 0 100 200 300 400 500 600 700 800 Keetch - Byram Drought Index (KBDI) Source: Authors, using weather data from Climate Research Unit, University of East Anglia and MODIS active fire detections operations and facilities, since most of the districts at most fires occur in peopled areas is not surprising and risk are also rich in minerals and subject to immense reflects the dominant human influence on the fire mining activities. regime. 1.2.2 Topography Yet, a look at the spatial distribution of fires reveals that there is a long tail of fires that are in more remote Local topography influences the difficulty of fire areas. About 10 percent of all fires are detected 10 km prevention and suppression and can raise the or farther from the nearest roadway. Also, 10 percent potential for out-of-control fires. Moving up steep of all fires are detected 17-18 km to the nearest built- slopes, fires can spread at several times the rate they up area. Response time to fires in these more remote would on level ground. Winds in rugged terrain can areas may be slower, and the potential for the fires to change direction quickly or blow harder, and fuels spread and grow beyond the point at which they can may dry out faster on south-facing slopes. Remoteness be easily contained may be greater. and rugged terrain can also prevent fire crews from reaching fires quickly enough to suppress them before Forest fires in India also tend to occur in flat or gently they become unmanageable (Smith 2017). hilly terrain at lower elevations; however, as in the case with roads and built-up areas, the spatial distribution In general, most forests and fires in India are of fires exhibits a long tail of fires detected in high or distributed close to people and infrastructure. Figures rugged terrain. States in which fires tend to occur in 1.7 and 1.8 show that about half of detected fires were the most rugged terrain include Himachal Pradesh, observed within 3-4 km of the nearest road, and half Jammu and Kashmir, Manipur, Nagaland, Tamil are within 7-8 km of the near built-up settlement. That Nadu, and Uttarakhand, not surprisingly. Figure 1.9 Strengthening Forest Fire Management in India   38 FIGURE 1.7: SPATIAL DISTRIBUTION OF FOREST FIRE DETECTIONS RELATIVE TO NEAREST ROAD (2014-2016) 20% 18% MODIS fire detections in forest 16% 14% Forest with no fires detected Percent observations 12% 10% 8% 6% 4% 2% 0% 0 2 4 6 8 10 12 14 16 18 20 Distance to nearest road (km) Note: Orange line represents distribution of forest areas in which no fires were detected; blue line shows distribution of forest areas in which fires were detected by MODIS. Data sources: MODIS monthly data product for active fires (MCD14ML); Open Street Map data from Geofabrik, http://download.geofabrik.de/ asia/india.html FIGURE 1.8: SPATIAL DISTRIBUTION OF FOREST FIRE DETECTIONS RELATIVE TO NEAREST BUILT-UP AREA (2014-2016) 10% 9% MODIS fire detections in forest 8% 7% Forest with no fires detected Percent observations 6% 5% 4% 3% 2% 1% 0% 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Distance to nearest built-up area (km) Sources: MODIS monthly data product for active fires (MCD14ML); built-up area data from EC JRC (2016) 39 Strengthening Forest Fire Management in India FIGURE 1.9: TERRAIN RUGGEDNESS SCORES FOR FORESTS IN WHICH FIRES WERE DETECTED (2003-2016) 16 flat gentle hills Himachal Pradesh 14 Jammu and Kashmir Manipur 12 Nagaland Percent of fire detections Tamil Nadu 10 Uttarakhand 8 6 highly rugged extremely rugged moderately rugged 4 2 0 0 100 200 300 400 500 600 700 800 900 1,100 1,000 1,200 1,350 Terrain Ruggedness Index (TRI) of 0-100 = level or nearly level ground, TRI of 100-250 = gently hilly terrain; TRI of 250-500 = moderately rugged terrain, TRI of 500 or more = highly rugged mountains (Riley et al. 1999). Data sources: MODIS monthly data product for active fires (MCD14ML); SRTM elevation data illustrates the terrain ruggedness scores for forests vary across forest types, density, composition, and in which fires were detected by MODIS from 2003- structure.25 In contrast with weather and topography, 2016. Among these hill states, Uttarakhand stands fuels are the only corner of the fire triangle (figure 1.3 out as having the most fires in highly or extremely above) that fire managers can control. rugged terrain. Resources and infrastructure needs for fire response are greater in these areas, where fire In tropical broadleaf forests, such as those that response may be impaired. Prevention is even more experience frequent fires in India, fuel load important. accumulation follows an annual cycle, with most (85 percent) of ground litter decomposing each year with 1.2.3 Fuels the monsoon rains (Tuome et al. 2009). This annual cycle limits fuel load accumulation. Fuels determine the potential for fires to ignite, grow, intensify, and spread. Combustible material in forests The build-up of flammable material may also be includes grasses, ground litter, small shrubs, living limited by recurrent fires. Satellite observations for and dead trees, and decomposing humus in soils. Fire 2003-2016 suggest that forest fires revisit the same potential and behavior is affected by the moisture area once every 3 to 6 years, with a nationwide average content, fineness, depth, compactness, and orientation of once every 4 years (figure 1.10). The average fire (vertical or horizontal) of these fuels. Fuel loads recurrence interval varies by region and predominant See Anderson (1982) and Parks and Wildlife Service, Department of biodiversity, Conservation and Attractions, Government of Australia, 25. “Fuel loads and fire intensity,” 19 June 2013, https://www.dpaw.wa.gov.au/management/fire/fire-and-the-environment/51-fuel-loads- and-fire-intensity Strengthening Forest Fire Management in India   40 forest type, with the shortest intervals seen in dry of the country’s total forested area, FSI notes that deciduous and thorn forests (table 1.3). For example, these forests account for a disproportionate number records from 1909-1921 for the present-day area of of fire incidents (FSI 2012). Subtropical pine forests the Mudumalai Wildlife Sanctuary in the Western show the highest portion of moderate or heavy fire Ghats compared with satellite imagery from 1989- disturbance among the country’s forest types (FSI 2002 indicate the average recurrence interval for 2015). Pine needles degrade slowly and have a high forest fires in this area has shortened from 13 years resin content. According to forest department officers to 3-4 years (Kodandapani et al. 2004). Also, in the interviewed in Uttarakhand, the average recurrence Northeast, jhum cycles have shortened from around interval for fire in chir pine (Pinus roxburghii) forests 30 years to around 5 (Pyne 1994). It is unclear whether is about 4-5 years on average. The buildup of highly these limited examples are representative of the rest flammable materials between fires results in more of the nation. intense fires such as the crown fires that were observed in Uttarakhand during 2016. Mondal and Sukumar (2016) have shown how fuel loading and fuel moisture constrain fire potential in 1.2.4 People and the causes of forest fires seasonally dry tropical forests, such as those found across much of Central and Southern India. They Similar to other parts of the world, people are the main find that fire potential is the result of a complex driver of fires in India. Population pressures, current dynamic between previous years’ fires, the monsoon, and historic land management practices, demand for and rainfall during the early part of the dry season. A forest resources, the use of fire as a tool, negligence, severe fire season followed by heavy monsoon rains and anthropogenic climate change all influence the can lead to greater understory plant growth and the other elements in the triangle and shape the forest fire accumulation of biomass that serves as fuel for fires regime today. during the next dry season, raising fire potential. Fire potential may also be high the year after a mild fire At present, nationally representative data on forest season. If there is less post-monsoon rainfall during fires do not exist for India, making a systematic analysis the early months of the current year, fuels will dry of the causes of fires difficult.26 Because of the lack of more quickly and burn more easily, likewise raising field data on the causes of fire, a survey was conducted fire potential. with the forest departments of 11 states, which asked officers about the causes of fires along with other With more limited fuel load accumulation from year information about fire prevention and management to year and more frequent fires overall, forest fires in practices. States in the survey sample were chosen in India’s seasonally dry tropical forests are generally consultation with MoEFCC and intended to reflect characterized by lower-intensity surface fires. Field different forest types, geographies, climates, and visits to forests in Telangana, Madhya Pradesh, Odisha, patterns of forest resource use. The survey is described and Jharkhand revealed no evidence of crown fires. in more detail in Annex 2. In-depth consultations with While surface fires tend not to cause major damage to members of forest-dependent communities were also the forest in the first instance, repeated burning can conducted in three districts in Uttarakhand and five have an impact on forest ecology, the humus content districts in Meghalaya, as described in Annex 3. of soils, species mix and forest growth and quality (discussed in section 1.3.1 below). Forest officers surveyed in the 11 states agreed overwhelmingly that people are the main source of By comparison with dry deciduous forests, there forest fire ignitions. Of the 83 officers who responded is a greater potential for intense fire behavior in to the survey, most (75) said that more than 75 percent India’s subtropical pine forests. Though forests such of forest fires in their area are caused by people; 56 as the chir pine (Pinus roxburghii) forests that are said that more than 90 percent of the fires are caused common at elevations of around 500 m to 2,000 m by people. Officers were also asked to rank the six most above sea level in states such as Himachal Pradesh common causes of forest fires in their area. Responses and Uttarakhand make up a relatively small share were categorized a modified version of the classification See Chapter 3 for further discussion of field data collection on forest fires and their causes. 26. 41 Strengthening Forest Fire Management in India FIGURE 1.10: FOREST-COVERED AREA AFFECTED BY FIRE BY COUNT OF MONTHS THE AREA WAS AFFECTED, 2003-2016 35 30 Percent of total fire-affected area 25 W. Himalaya 20 West South 15 Northeast 10 North Central 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14+ Number of months in which the forest area burned, 2003-2016 Data sources: MODIS monthly data product for burnt area (MCD45A1); forest cover data for 2000 from Hansen et al. (2013) TABLE 1.3: IMPLIED AVERAGE FIRE RECURRENCE INTERVAL BY FOREST TYPE, 2003-2016 Forest type Number of months in which Implied recurrence interval forest burned (2003-2016) (years per fire) Wet Evergreen forest 2.0 7 Semi Evergreen forest 2.3 6 Moist Deciduous forest 3.1 5 Dry Deciduous forest 3.9 4 Dry Evergreen forest 2.3 6 Thorn forest 4.1 3 Subtropical broadleaved forest 1.7 8 Subtropical Pine forest 2.2 6 Subtropical Dry Evergreen forest 2.9 5 Montane Wet Temperate forest 1.7 8 Montane Moist Temperate forest 2.6 5 Montane Dry Temperate forest 2.2 6 Sub Alpine forest 2.3 6 Data sources: MODIS monthly data product for active fires (MCD14ML); forest type data from Reddy et al. (2015), available from National Remote Sensing Centre, Bhuvan data platform, http://bhuvan.nrsc.gov.in/. Strengthening Forest Fire Management in India   42 scheme of the Fire Database of the European Forest (Diospyros melanoxylon) to make wrapping papers for Fire Information System (Camia et al. 2013) that was beedi cigarettes, or to maintain open canopy cover adapted for the Indian context (see Annex 4). Table and space for bodha grass (Cymbopogon), which is 1.4 presents the categorized responses for the most used for roof thatching (Schmerbeck et al. 2015). common causes of fire by state, with importance scores Other NTFPs harvested with the aid of fire include based on the weighted rankings of identified causes.27 fodder, honey, mushrooms, seeds, medicinal plants, charcoal, bamboo shoots, vegetables, fruits, tubers, Although most officers agree that humans are and dammar gum or resin. Tendu leaf collection was responsible for most forest fires, table 1.4 shows the second most important cause of fire named in significant differences between states in the most Chhattisgarh, Madhya Pradesh, and Telangana. It is common reasons why fires are set. Officers cited the also common in Odisha. Setting fire to stimulate the negligent use of fire as the most common cause of growth of new grasses and fodder in the forest for forest fires in three states, Himachal Pradesh, Kerala, livestock was cited as a major cause of fire in Himachal and Uttarakhand. In Himachal Pradesh, officers Pradesh, Telangana, and Uttarakhand. In all, 47 of 76 pointed mainly to agricultural burning on farmlands field-level officers said local people use fire in forests and pasture (ghasnies) adjacent to forest. They also to promote fodder. cited escapes from burning weeds and bushes on privately-owned lands next to reserved forests. In These survey results reinforce findings from previous Kerala, officers pointed to out-of-control fires started studies. As the National Forest Commission had by private citizens doing early-season burning in reported in 2006: “crown fires in coniferous forests forests near settlements and agricultural lands. In and ground fires in the rest…are mostly man caused. Uttarakhand, officers frequently mentioned the use of Fires are purposely set to promote new flush of grass fire to clear paths and other areas in the forest of fallen or tendu leaves, to facilitate collection of honey, sal pine needles, which can be slippery. As in Himachal seeds and mahua and chiraunji and to prepare land Pradesh, escaped fires from burning agricultural for shifting cultivation” (NFC 2006: 94). residues on adjacent farmlands were also a major cause of forest fires. Agricultural burning as a cause of forest The survey results also provide further insight into fires is a widespread problem in other states, too: 56 of the importance of green fodder from forests for rural the 76 field-level officers surveyed agreed that escapes livelihoods, and the use of fire as an input to fodder of agricultural fires on adjoining lands were a cause of production. FSI reports that cattle are grazed in about forest fires in their areas. Officers in all the states also 71 percent of the nation’s forest area (FSI 2015). A blamed the negligent disposal of cigarettes in forests previous study that estimated 30 percent of fodder and nearby areas, especially roadsides, as a common requirement for India’s livestock population comes cause of fire. from forests, including for the 90 million animals that are grazed in the forest (Rai and Saxena 1997). Other than negligence, the collection of non-timber About 10-25 percent of the households in the forest- forest products (NTFPs) was reported as another dependent communities interviewed for the present main cause of fire. Officers in five states identified study said they raise goats or sheep, and community the process of obtaining NTFPs as the most common members said they prefer to graze their animals in cause of forest fires: Chhattisgarh, Jharkhand, the forest. An earlier field survey of eight forest-fringe Madhya Pradesh, Odisha, and Telangana. Officers communities in Assam by the World Bank found in these states and others pointed to a diverse array that green fodder from the forest supplied about 64 of NTFPs obtained with the aid of fire. People may percent of feed requirements for domestic livestock burn to aid in the collection of flowers from the mahua there (World Bank 2005). Across the country, in the tree (Madhuca indica) for food or to brew alcohol (as Western Ghats region, researchers estimated that it is easier to find the fallen flowers when there is cattle consumed an average of 13 kg of green fodder no under growth), to flush the leaves of tendu trees per day while free grazing in forests, translating into The first most common cause was given a weight of 6; the sixth most common cause was given a weight of 1. Responses from each officer 27. in the state were scored with equal importance, regardless of official designation. Scores were then aggregated at the state level and scaled on an index from 0 to 100, where the highest scoring cause is equal to 100. 43 Strengthening Forest Fire Management in India TABLE 1.4: CAUSES OF FOREST FIRES ACCORDING TO SURVEYED FOREST DEPARTMENT OFFICERS, BY STATE (INDEX OF IMPORTANCE, 0-100) Most important cause of fire Second most important cause of fire Category Cause Assam Chhattisgarh Himachal Jharkhand Kerala Madhya Meghalaya Odisha Telangana Tripura Uttarakhand Pradesh Pradesh Unknown 3 38 Natural Natural, not specified 38 5 8 22 2 4 10 6 Lightning 1 2 5 Other natural 1 Accident Accident, not specified 6 36 52 5 2 5 Electric power 7 6 2 5 Works 10 Self-ignition 1 Negligence Negligence, not specified 3 15 10 64 17 7 35 Negligence, use of fire 38 32 100 42 100 69 22 36 100 Negligence, glowing 100 18 51 46 19 8 21 1 19 24 33 objects Deliberate/ Deliberate, not specified 8 21 21 13 incendiary Responsible (arson) 58 12 16 19 92 11 2 12 30 10 24 Not responsible (e.g., fires 6 set by minors) Resource NTFP collection 100 8 100 72 100 100 100 7 3 collection Grazing or fodder 8 73 8 25 22 38 11 64 12 70 Wildlife Burning to deter wildlife 7 10 39 14 3 17 Hunting 12 5 4 19 19 14 10 Other, Traditional practice, not 6 3 1 cultural specified Shifting cultivation 85 5 100 45 9 100 6 No. of responses 6 7 14 10 10 3 11 9 9 9 16 Strengthening Forest Fire Management in India   44 an annual benefit of Rs 3,260 per head assuming the villages has reduced people’s dependence on the average price for paddy, finger millet, and sorghum forests for their livelihoods. Without regular use of the straws that households would otherwise have to buy forests (e.g., for grazing or for the collection of grasses, to feed their animals (Ninan and Kontoleon 2016). animal bedding, and dry wood), and with the reduced In tropical dry forests, households routinely use fire practice of controlled burning to promote fodder or in forests where they graze their cattle to burn away to clear forest litter, fuel loads have accumulated and undergrowth, stem the growth of woody and thorny created the potential for more severe and destructive plants, and promote the growth of grasses under an fires in these areas. Forest department officers also open canopy (Schmerbeck and Fiener 2015). commented on greater difficulty in organizing labor from the local communities to conduct fire prevention After negligence and NTFP collection, the third or awareness-raising activities.28 main cause of forest fires cited by surveyed officers was shifting cultivation (jhum), particularly in the Another cause of forest fires cited by surveyed officers was northeastern states of Assam, Meghalaya and Tripura, burning to deter wildlife. Consultations in Uttarakhand but also in Odisha. Due mainly to jhum cultivation, provided additional insight into the link between forest the three northeastern states also had the highest fires and increased human-wildlife conflict in forest- number of active fire detections per square kilometer fringe areas. Community members said they burn of forested area during the years 2003-2016. Surveyed pine needles, cones, weeds, and so on during the dry officers in these northeastern states noted that local season to keep away wild boars, birds, and leopards. people use fire in other traditional practices, too, for Households grazing their livestock in the forests may example in group hunting. also burn away undergrowth and forest litter to remove cover for wild animals that might threaten their herds. Community consultations in Meghalaya shed light on Yet, the removal of habitat for some unwanted animals how land tenure and management arrangements may by burning grasses, undergrowth, and forest litter has also influence the prevalence of fires in some forests. In brought these animals closer to settlements in forest- Meghalaya, about 88 percent of forests are controlled fringe areas in search of food and shelter, thereby by communities or private individuals, outside the increasing the potential for conflict. jurisdiction of the state forest department. However, with breaking up and weakening of traditional institutions, The link between poverty and forest fires was cited community members said forest management has by only a few of the officers who responded to the become more of a challenge. Individual ownership forest department survey. Additional analysis using without sufficient resources, incentives, or capacity to district-level poverty data and satellite detections of protect forests, combined with a weakening of social forest fires districts helped draw out this connection. norms to manage forests sustainably has made these The analysis, described in more detail in section 3 of forests more vulnerable to fires. Fire incidence is also Annex 1, revealed that districts with a higher share of higher in law raid, forested lands managed jointly by their population living below the national poverty line groups of villages, which tend to be more intensely also tend to experience more forest fires. As table 1.5 exploited and are more likely to fall into neglect. The shows, the average annual number of fires detected creation of Village Fire Control Committees (VFCCs), per unit area of forest cover is more than three times such as the one observed in Jirang, Meghalaya, higher in the poorest districts than in the least-poor has helped strengthen joint management and fire ones. Districts in the table are grouped into quartiles prevention in some of these forests. according to the poverty headcount ratio for the district in 2011, with the least-poor districts (those with the A weakening of traditional land management practices smallest percent of the population below the national was also seen in some of the villages where consultations poverty line) in the first quartile and the poorest were conducted in Uttarakhand, increasing districts in the fourth quartile. About 59 percent of the vulnerability to forest fires in those areas. According poorest districts are in Central India, and 30 percent to community members and forest officers who were are in the Northeast. The rest are spread across the interviewed for this assessment, out-migration in these North, South, West, and Western Himalayas. 28. Community institutions for forest fire management are discussed in greater depth in chapter 3. 45 Strengthening Forest Fire Management in India Although there is clear geographic overlap between 1.3 IMPACTS OF FOREST FIRES districts with high fire density and districts with high poverty rates, poverty by itself cannot explain why Some forest fires are beneficial, but not all. Fire has fires are more concentrated in these areas. District- been a part of India’s landscape since time immemorial level regression analysis reveals that poverty rates are and can play a vital role in healthy forests. Many of not a significant explanatory factor in determining India’s forests have evolved with fire and rely on fire fire density after accounting for population density, to regenerate. Occasional fires can also keep down fuel rainfall, temperature, predominant forest types found loads that feed larger, more destructive conflagrations. in fire-affected districts, and unexplained regional Today, however, large areas of degraded forest are differences.29 In other words, poor rural districts tend subject to burning on an annual or even semi-annual to be in more fire-prone areas, but simple differences basis. in the prevalence of poverty cannot explain why some of these districts in fire-prone areas experience more State forestry policies recognize that fires are taking a fires than other. toll on forests. The Assam Forest Policy (2004) points TABLE 1.5: POVERTY RATES AND FIRE DENSITY IN RURAL FORESTED DISTRICTS, GROUPED BY QUARTILE ACCORDING TO THE POVERTY HEADCOUNT RATIO IN 2011 Quartile 1 2 3 4 (Lowest poverty (Highest poverty rate) rate) Poverty headcount ratio, 4.5 14.1 27.2 48.2 mean, 2011 (% population below national poverty line) Poverty headcount ratio, 0.0-8.8 8.9-20.4 20.5-35.2 35.3-78.7 range, 2011 (% population below national poverty line) Average forest cover, 2000 60.2 55.9 51.1 53.5 (% total district area with ≥ 10% tree canopy cover) Average forest fire density, 1.1 2.0 2.7 3.8 2003-2016 (annual fire detections per 100 km2 treed area) Note: Average forest cover and fire density are weighted by the forested area per district; sample is limited to rural districts (population density < 1,000 per km2) with at least 10-percent forest cover in 2000 (554 of 638 districts with data); forest cover is defined as an area having at least 10-percent tree canopy cover. Sources: Authors, using World Bank subnational poverty data, MODIS fire detections in forested area provided by FSI using the MCD14ML product, forest cover in 2000 from Hansen et al. (2013) 29. Refer to section 3 of Annex 1 for details. Strengthening Forest Fire Management in India   46 47 Strengthening Forest Fire Management in India to forest fires as a cause of considerable damage in Southern India have found that repeated fires over plantation and regeneration areas, and the State short intervals are having a deleterious effect on Afforestation Policy of Tripura also mentions that forest composition, structure, and species diversity. plantations and natural forests are severely damaged In the Nilgiri Biosphere Reserve in the Western by forest fires. The Himachal Pradesh Forest Sector Ghats, Kondandapani et al. (2009) find “drastically Policy (2005) recognizes that forest fires cause altered” species structure and diversity and reduced irreparable damage to forests, biodiversity, wildlife, seedling density in areas of dry deciduous forest with water resources, forest-based livelihoods and well- the shortest fire return intervals compared to forest being. The Andhra Pradesh State Forest Policy (2002) patches with lower fire frequency. Jhariya et al. (2014) also notes the deleterious impact of forest fires, observe a similar pattern in the dry deciduous forests especially on the young plantations. of the Bhoramdeo Wildlife Sanctuary in Chhattisgarh. In Maharashtra, Saha and Howe (2003) find that The current pattern of fire is no longer beneficial to repeated fires favor species in dry deciduous forests forest health, yet the extent to which fires are having that re-sprout clonally from root buds and spread to a longer-term impact on India’s forest ecology and its new ground away from the parent plant by sending wider economy are still poorly understood. out rhizomes or root suckers; fires suppress species that sprout basal shoots from root crowns and spread 1.3.1 Ecological impacts by dispersed seeds. The result is dominance by a few clonal species and lower tree diversity. The ecological impacts of forest fires are specific to the different types of forests, situated in different climates Lower species diversity, reduced biomass, and and geographies, and subject to other disturbances, homogenous structure of dry deciduous forests most particularly from people. Forests that are affected affected by fire may reflect damage from frequent by fire may also be affected by agriculture, grazing, fires to regeneration, as fewer seedlings grow to reach harvesting fuelwood and other NTFPs, encroachment larger size classes. However, the effects of different or fragmentation from road building and construction, intensities and frequencies of fire on regeneration illicit felling, invasive species, and numerous other in dry deciduous forests are still poorly understood pressures. The ability of forests to withstand and (Thekaekara et al. 2017). In another study in the recover from fires will depend largely on how these Nilgiri Biosphere Reserve in the Western Ghats, other pressures are managed. Mondal and Sukumar (2015) compare survival rates for juvenile tree species and find no difference There is limited literature on impacts of forest fire between burnt and unburnt areas, even for plots in India, as assessed through field research. As the affected by multiple fires over a short period. They National Forest Commission noted in 2006: “The also notice that juvenile trees in burnt areas quickly nature and severity of damage depends on the type of bounce back to pre-fire height in 1-2 years. However, forest, availability of fuel and climatic factors. However, after recovering, the trees continue to grow more the damage to forest ecosystem due to fire has not slowly and may require several fire-free years to reach been scientifically studied” (NFC 2006: 94-95). Much a larger, more fire-tolerant height and girth. In their of the existing research has focused on seasonally dry study in Maharashtra, Saha and Hiremath (2003) tropical forests (including dry and moist deciduous also note higher growth rates among young trees in forests) in Central and Southern India and subtropical burnt areas, but they also find that repeatedly burnt pine or mixed-broadleaf forests in the hill states of the forest plots exhibit stunting. Damage to sal seedlings Western Himalayas. from low-intensity surface fires and negative effects on regeneration of trees forming the top canopy layer has 1.3.1.1 Forest composition, structure, and species also been observed in the plains forests of Uttarakhand diversity by Maithani et al. (1986). Tropical dry forests. Fire is found to play a On the other hand, Kondandapani et al. (2009) in the complex role in the tropical dry deciduous forests Nilgiri Biosphere Reserve observe that in dry thorn of India. Although some important tree species in forests, the greatest seedling density, species diversity, dry deciduous forests, such as teak and sal, require and the number of sapling and standing trees was fire to regenerate, several studies in Central and found in areas of moderate or high fire frequency. Strengthening Forest Fire Management in India   48 They hypothesize that patchy, low-intensity fires on their intensity and return interval (Verma and “could actually be recycling nutrients back into the Jayakumar 2012). In a study in South Kashmir, Khaki soil, mitigating invasive species, reducing flammability et al. (2015) find that total soil carbon and nitrogen of vegetation and promoting regeneration of seedlings content were lower in burnt versus unburnt areas, and saplings” (350). while phosphorus and potassium were higher. The findings corroborate those from an earlier study by Tropical moist forests. Moist deciduous forests are Banerjee and Chand (1981) in the Darjeeling hill much more sensitive to fire. In the Nilgiri Biosphere district of West Bengal, which notes fire-affected soils Reserve, Kondandapani et al. (2009) observe that show reduction of organic carbon and nitrogen. Verma areas of moist deciduous forest affected by fire have and Jayakumar (2012) note, however, that while total fewer tree species and a lower density of seedlings, nitrogen tends to decrease in soils after fires, plant- saplings, and standing trees. They are also more available forms (NH4+) increase, spurring a flush of susceptible to invasion and replacement by grasses. A grasses and herbaceous vegetation and promoting study of moist deciduous forests in the Achanakmar- regrowth. In their meta-analysis of previous studies, Amarkanak Biosphere Reserve in Chhattisgarh by Holden and Treseder (2014) find evidence that fires Kittur et al. (2014) which compared plots exposed to reduce the abundance of microbes in soil due to high different frequencies of fire concludes similarly that temperatures and the removal of organic carbon that the regeneration and size structure of economically soil microbes can decompose. Higher-intensity fires important tree species such as sal is harmed by can severely deplete soils and strip them of organic repeated fires. The population structure in the most matter and nutrients (Chandra and Bhardwaj 2015). fire-affected plots consists of seedlings and saplings of a similar age and few older trees. In Thrissur Forest In Northeastern India, the shortening of fire- Division of Kerala, Valappil and Swarupanadan associated jhum cycles has also had a detrimental (1996) also note few trees in upper size classes and low effect on soil fertility. As Ramakrishnan (2007) has survival probabilities for seedlings due to fire, grazing, documented, shorter jhum cycles reduce fallow and browsing in tropical moist forests there. biomass available for burning and gives soil fertility less time to recover, resulting in lower economic yields Subtropical pine forests and mixed broadleaf forests. and efficiency. Economic yields for grains and seeds, Pine forests in the hills of Uttarakhand and Himachal leaf and fruit vegetables, and tubers decline by more Pradesh are highly fire-adapted but may also suffer the than half in moving from a 20-year jhum cycle to a negative effects of overly frequent burning. Parashar 5-year one. and Biswas (2003) report damage to seedlings in pine, oak, and mixed deciduous forests in Uttarakhand Water retention and erosion. Fires increase water due to repeated fire in some areas, though they also repellency of forest soils, reducing infiltration and note that damage to regeneration is due mainly to increasing erosion (Verma and Jayakumar 2012). the combination of fire followed by heavy grazing In their study in West Bengal, Banerjee and Chand and browsing by goats and sheep. Singh et al. (1984) (1981) find moisture retention and available water comment that oak forests in Uttarakhand are gradually in fire-affected soils are “radically” reduced. D. being converted into pine forest because of human Nagbhushanam of the Hyderabad Forest Department pressures such as fire, lopping, grazing, and leaf litter writes, “Because of frequent fire all waste matter such collection; fires promote the expansion of pine forests as leaves, twigs, small branches, grasses etc., are burnt dominated by chir. At the same time, Bhandari et al. and converted into ashes which gets wasted during (1997) notice greater species diversity and richness in rain exposing the top soil as there is no layer of humus. burnt pine stands from greater growth in shrubs and Because of the above phenomenon severe soil erosion ground vegetation promoted by openings in the forest is noticed due to which roots of trees become weak canopy (Bhandari et al. 2011: 171). and natural regeneration is poor” (Hyderabad Forest Department 2013-14 to 22-23 by D. Nagbhushanam). 1.3.1.2 Forest soils 1.3.1.3 Wildlife and other forest animals Soil chemistry and biology. Fires alter the physical, chemical, and biological properties of forest soils. Habitat management. Evaluating the role of fire Fires can be either beneficial or harmful depending in wildlife conservation in India, Rodgers (1986) 49 Strengthening Forest Fire Management in India finds that fires may benefit wild herbivores in some 1.3.1.4 Invasive species areas. Controlled patchwork burning of small areas of moist grassland may enhance habitat for grazing Some invasive species in India’s forests are fire- species such as swamp deer and chital. The benefits assisted. One of the most pernicious of these species of fire diminish, however, as habitats get drier and fire is Lantana camara, a woody plant believed to have frequency increases. Also, though fire may be useful been introduced to India in the 1800s (Bhagwat et in promoting habitat for some wild herbivores to al. 2012). In a 2005 paper, Hiremath and Sundaram some extent, not all species benefit. Even low-intensity hypothesize the existence of a fire-lantana cycle: surface fires may destroy nests, dens, and eggs and kill young animals that cannot escape quickly enough.30 “forest fragmentation, coupled with intensified Burning for habitat management may be appropriate anthropogenic disturbances—especially fires— only under specific, limited, and controlled conditions have resulted in degradation of ecosystems, (Rodgers 1986). making them more vulnerable to invasion by alien species; some invasive species (e.g., Human-animal conflict. According to the Wildlife lantana), in turn, fuel further fires. The Institute of India, intensified human-animal conflict resultant positive feedback has deleterious may result from forest fires. As Maithani et al. (1986) compositional and functional consequences have documented, though fires promote the growth for ecosystems and the goods and services that of some herbaceous species, such as grasses and herbs, society derives from them” (Hiremath and that are palatable to ungulates, fires may eliminate Sundarem 2005: 34). other species eaten by wild animals. As surface fires remove food, water, and shelter, some animals living More recent studies, incorporating indigenous in the forest understorey may be forced to move out knowledge (box 1.2), however, have added nuance of forest to fringe areas. to this hypothesis, revealing that fires can kill lantana seeds in the soil and, under certain conditions, may Livestock grazing. Semwal and Mehta (1996) observe help control one of the ways in which lantana spreads. that low-intensity fires enhance the carrying capacity Early summer burning when the weather is cooler of grazing lands in pine forests, releasing stored and fuels are not as dry would also prevent the nutrients in the biomass pool and promoting the accumulation of fuels in lantana-infested areas that growth of herbaceous vegetation. Bhandari (1995) have caused larger and more intense fires lately. estimates above-ground net primary productivity (NPP) for herbaceous vegetation in pine forests is In Northeastern India, Ramakrishnan and Vitousek highest in areas affected by fire every 2-5 years. (1989) have reported that shorter jhum cycles have Others debate the benefits of frequent fires to grazing. also accompanied the more aggressive spread of Research by Konsam et al. (2017) in chir areas of invasive weeds into forests. One theory for this is the Garhwal Himalaya in Himachal Pradesh finds that nutrient-depleted soils and short fire recurrence no significant difference in regeneration potential of intervals have prevented native plant species from understorey vegetation (including fodder) in burnt producing seeds and growing, creating an absence of versus unburnt sites. Semwal (1990) notes that high- propagules to recolonize burnt patches. The quicker- intensity crown fires followed by heavy rains during growing invasive weeds have filled this gap and the monsoon reduced ground vegetation available to crowded out other species (cited in Hiremath and livestock and other animals because of soil erosion. In Sundaram 2005: 32). the tropical dry forests and savanna that stretch across much of Central India, grasses grown from recently 1.3.1.5 Carbon storage and emissions burned areas may have higher nitrogen and protein content (Lü et al. 2012; Mbatha and Ward 2010) and Forest fires contribute to climate change by releasing provide greater nutritional value to grazing animals. carbon stored in trees, undergrowth, litter, and soils 30. Dr. S.S. Negi, retired Director General of Forests, MoEFCC, written comments to World Bank, February 2018. Strengthening Forest Fire Management in India   50 Box 1.2: Indigenous Knowledge, Early-Season Controlled Burning, and Stemming the Lantana Invasion The forest-dwelling Soliga community in the present-day Bilgiri Rangaswamy Temple Wildlife Sanctuary (BRT) in Karnataka traditionally used controlled burning to manage the area’s forests before the sanctuary was notified in 1974. They call the practice of setting low-intensity fires in the early summer when the weather is cooler, tarugu benki or “litter fires,” which they maintained to promote understorey plant growth and eliminate parasites. Since 1974, the BRT has maintained a policy of total fire exclusion. Partly because of this policy, the weed Lantana camara has invaded much of the area’s forests, as fuel loads have built up and fires have been larger and more destructive, killing native vegetation and creating conditions ripe for lantana to spread. Soligas who were interviewed by researchers note invasion has altered forest structure, causing a decline in understorey plants and natural regeneration of canopy trees, as saplings have struggled under the dense thickets of the weed. They also claim the absence of fire has led to an increase in adult tree mortality due to hemiparasites (Sundaram et al. 2012). Scientific studies in the BRT have backed the observations made by the Soligas. Sundaram et al. (2015) discovered that forest plots that burned most frequently have the lowest density and coverage of lantana. Setty (2004) confirmed that trees in areas that experienced low-intensity surface fires. Still, resistance to the reestablishment of controlled burning in the BRT remains. Controlled burning is not a cure-all and may not be appropriate in all areas. Though tarugu benki may help check the spread of lantana, it would not be feasible where lantana has already established dense thickets and climbed trees, presenting a heavy and vertically-oriented fuel load. Under such conditions, there is a greater risk of crown fires that could result in high tree mortality, creating an opening that would only be filled by more lantana. The case of the BRT illustrates how a policy of total fire exclusion has had a negative ecological effect, replacing large areas of native forest with lantana in the span of a few decades. A pragmatic fire management policy for dry forests should look at the option of learning from indigenous fire management practices and consider when controlled burning may be effective, such as during early-summer when burning may help avoid fires later in the dry season, when temperatures are higher and the potential for more intense fire behavior is greater. Sources: excerpted and adapted from Thekaekara et al. (2017) and Suresh (2017) into the atmosphere. Forest fires also emit heat- al. 2014; Sommers et al. 2014). Frequent fires are trapping gases such as N2O and other aerosols that one such critically important indirect mechanisms. influence the regional and global climate. The net Almost every dry-land vegetation type on the planet effect of a fire on the climate depends on the pre- experiences fire at some point in its life cycle, maybe disturbance characteristics of the forest and the extent on a relatively frequent and regular basis or maybe to which the forest can regenerate. Forest clearing as little as once each several centuries. Following fire, and persistent changes in vegetation composition and the vegetation recovers and situation normal prevails. structure after a fire may result in net emissions. As In the short term there may be a spike in emissions, global warming proceeds, increases in tree mortality but in the long term it doesn’t really matter whether and degradation due to heat stress, drought, pests, and emissions are caused by occasional fire or natural other indirect mechanisms create a positive feedback break down by microorganisms, thereby releasing loop, furthering eroding the forest carbon sink and the constituent elements and compounds. What does further contributing to climate change (Settele et matter is when the fire frequency is altered such that 51 Strengthening Forest Fire Management in India it changes the floristic structure so that the maximum The studies listed in table 1.6 only quantify emissions “carbon storage” capability on a site is reduced. from burning above-ground vegetation; the studies do not consider the effects of fire on the vast pools Scientific research on the contribution of forest fires of carbon stored in below ground biomass and to climate change in India has so far been limited forest soils. According to FSI, below ground biomass to estimates of direct emissions from the burning of and organic soils contained 699 million tons and above-ground biomass and have not considered the 3.979 billion tons of carbon, respectively, in 2017, impact on regeneration. Nation-wide estimates have representing a combined 66 percent of India’s total ranged from 6.34 million tons (Mt) CO2 per year to as forest carbon stocks (FSI 2018). As noted in section much as 123.84 Mt CO2 per year (table 1.6). The wide 1.3.1.2 above, empirical studies in different regions of range of estimates reflects not only the inter-annual India have found that forest fires deplete soil organic variability in fires, but also significant differences in carbon; however, the authors are unaware of any assumed parameters (Badarinath and Vadrevu 2011). existing estimates of nationwide emissions from soils due to fire. Sommers et al. (2014) have noted several major sources of uncertainty involved in quantifying If frequent fires are a “major” cause of degradation, emissions from above-ground biomass burning as MoEFCC has asserted, then the weakening of following the approach taken by each of the studies carbon sinks from limited productivity and damage illustrated in table 1.6. Under this approach, to natural regeneration may prove an even greater emissions are calculated as the product of burnt area contribution to climate change than direct emissions (e.g., hectares), the above-ground fuel load or biomass from the combustion of above-ground biomass. This is per unit area (e.g., tons per hectare), combustion an urgent area for research, as a reduction in carbon completeness (percent of biomass burned), and the uptake and storage in forests could pose a risk to emission factor for vegetative biomass in the forest targets the government has set to create an additional type burned (e.g., grams of carbon released per ton sink of 2.5 billion to 3.0 billion tons worth of carbon of biomass burned). According to Sommers et al., the CO2 stored in its forests in 2030 by expanding forest chief sources of uncertainty are typically burnt area cover and improving forest health. and the heterogeneity of fuels, as the availability of live and dead vegetation, moisture content, and other 1.3.1.6 Summary of ecological impacts characteristics may vary widely from forest to forest. In India’s case, uncertainty is compounded by the The available scientific evidence supports that fires are fact that most fires are low-intensity surface fires, and having a degrading effect on India’s forests. Repeated thus the amount of biomass that is consumed by fire fires in short succession are reducing species richness may be relatively low. Also, as mentioned, much of the and harming natural regeneration, in combination carbon in the ground litter that is burned would be with other pressures such as intense grazing and decomposed and released back into the atmosphere browsing. In some forests, fire may be used in a anyway with the onset of the monsoon, even in the controlled way to manage fuel loads, check invasive absence of fire (Tuome et al. 2009). weeds, and eliminate pathogens. In other forests TABLE 1.6: PREVIOUS NATION-WIDE ASSESSMENTS OF CARBON EMISSIONS FROM FOREST BIOMASS BURNING Study Period Scope Emissions (Tg CO2 year-1) Venkataraman et al. (2006) 2001 National 49-100 Badarinath and Vadrevu (2011) 2000-2007 National 6.34 (mean) Srivastava and Garg (2013) 2003-2010 National 74.95-123.84 (range) Saranya et al. (2016) 2004-2013 Similipal Biosphere 1.26 (mean) Reserve, Odisha Reddy et al. (2017a) 2014 National 98.11 Strengthening Forest Fire Management in India   52 that are less adapted to fire, it should be excluded. same. The Committee also found that the impact of Reductions in biomass, species diversity, and natural forest fire on biodiversity is severely under-estimated regeneration due to fire may pose a risk to policy goals and that the loss of wildlife is not accounted for. for enhancing India’s forest carbon sinks. A fuller accounting of the economic costs and benefits of 1.3.2 Economic impacts forest fires in India would serve several purposes. First, it would provide a clearer picture of the many ways in As with ecological impacts of fires, comprehensive which forest fires affect India’s society, economy, and assessments of the economic losses due to fire in India environment and the dynamics of who gains and who are lacking. One oft-cited figure from India’s former loses from fire. Second, a more inclusive accounting of Deputy Inspector General of Forests, V.K. Bahuguna fire costs would support policymakers in determining puts annual damages from fires at around INR 1,101 the appropriate level of financial resources to devote crore (US$ 164 million, year 2016 prices) (Bahuguna to FFPM. And, third, such an accounting framework 1999).31 Though the details behind this figure are may also help in assessing the results of FFPM policies unclear, Bahuguna notes biodiversity, soil degradation once they are put into action. and erosion, and intangibles are excluded. In official reports and statistics, monetary damages 1.4 SUMMARY due to forest fires are generally assessed only for the loss of standing trees (natural or planted) in terms of Every year, forest fires occur in around half of the their timber value. Table 1.7 below lists monetary losses country’s 647 districts and in nearly all the states. from forest fires as reported by an illustrative sample Though fires are spread throughout the country, they of states. Average damages reported per hectare in occur much more frequently and affect forest more in 2016 ranged from INR 0 in Chhattisgarh (according some districts than in others. Just 20 districts accounted to the forest department, because “only ground fires” for 44 percent of all forest fire detections from 2003 occur in that state, there have been “no losses so far”) to 2016. Similarly, just 20 districts (not necessarily to INR 2,344 in Himachal Pradesh. the same ones) accounted for 48 percent of the total fire-affected area. These districts with the highest fire Reflecting on the limited scope of damages recorded, frequency and largest extent of fire-affected areas an officer surveyed in Kerala offered: should be priorities for intervention, as should areas of significant ecological, cultural, or economic value. “There is no scientific method to assess the loss Data from 2014, for example, showed that about 10 caused by forest fire. In most cases, it is limited percent of forest cover in protected areas was affected to the loss of timber / wood, if any, which is very by fire that year (Reddy et al. 2017b). low compared to the actual loss. Method to assess loss in terms of damage to other vegetation, soil, In India’s seasonally dry forests, most forest fires micro flora / fauna, loss of habitat, impact on are characterized by low-intensity surface fires. The ecosystem services etc. to be developed and then potential for more intense and difficult-to-control put to use. Reporting the actual loss through fires is shaped by a complex dynamic involving the such an assessment will convey the seriousness of monsoon rains, weather during the winter and early the issue to all concerned and will immediately part of the dry season, and fuel accumulation. Also, stir the system to quick response.” although India’s forests are densely populated—and most fires occur within a few kilometers of the nearest Similarly, a Parliamentary Committee report which road or settlement—each year there is a long tail of was presented to the Rajya Sabha in December 2016 fires in more remote and inaccessible areas, where asserted that there is a “gross underestimation of response is slower and the potential for fires to grow losses” due to forest fires and urged the appointment beyond control is greater. of a credible independent agency to estimate the The original estimate from Bahuguna for 1998 was INR 440 crore (US$ 107 million, year 1998 prices). Losses have been adjusted for 31. inflation to year 2016 prices using the GDP deflator. Estimates in US dollars are converted at market exchange rates using the year 2016 period average. 53 Strengthening Forest Fire Management in India TABLE 1.7: REPORTED MONETARY LOSSES DUE TO FOREST FIRES IN SELECT STATES 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Total losses, INR lakh (US$ 1,000) Chhattisgarh 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Himachal Pradesh 255.23 97.69 43.08 276.83 52.31 113.27 134.78 (558.17) (209.33) (80.62) (472.42) (85.71) (176.56) (200.58) Kerala 1.89 1.04 2.46 1.11 0.25 (3.53) (1.77) (4.04) (1.73) (0.38) Uttarakhand 3.84 2.68 4.79 1.90 0.30 42.89 4.39 23.58 7.94 46.50 (9.29) (6.16) (9.90) (4.16) (0.64) (80.27) (7.50) (38.63) (12.38) (69.20) Losses per hectare, INR (US$) Chhattisgarh 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Himachal Pradesh 1,027.10 1,246.47 2,450.24 1,332.56 1,615.75 1,683.74 2,343.97 (22.46) (26.71) (45.85) (22.74) (26.47) (26.25) (34.88) Kerala 33.42 43.91 93.54 65.54 12.87 (0.63) (0.75) (1.53) (1.02) (0.19) Uttarakhand 240.79 113.04 116.42 118.00 128.25 1,518.91 1,144.09 2,534.27 1,132.19 1,048.82 (5.82) (2.60) (2.41) (2.58) (2.75) (28.42) (19.52) (41.53) (17.65) (15.61) Note: In nominal terms, not adjusted for inflation; assuming official exchange rate for INR to US$ by year Sources: data sheets provided by state forest departments; Kerala (2016) Weather, fuels, and topography may influence fire natural regeneration, in combination with other potential and behavior, but virtually all forest fires pressures such as intense grazing and browsing. in India, as in other parts of the world, are caused Reductions in biomass, species diversity, and natural by people. Roughly 150 million people live in or regeneration due to fire may pose a risk to policy nearby forests, and many depend on forests for goals for enhancing India’s forest carbon sinks. Not their livelihoods. Many of the important goods and all fires are bad. The key is to maximize the ecological services that people obtain from forests, such as benefits of fire while minimizing the adverse impacts, fodder for their livestock, are generated or gathered recognizing that the controlled use of fire may play through the aid of fire. Unwanted forest fires may also a positive role in the management of fire-adapted occur due to human negligence, for example, from forests. casually discarded cigarettes or from poor control of burning on adjacent croplands. Shifting societal and Current estimates of the economic costs of forest fires cultural practices also play a role, as with the use of in India are almost certainly underestimates. Monetary fire in traditional shifting cultivation (jhum) in the damages due to forest fires are generally assessed only Northeastern states. In some parts of the country, for the loss of standing trees (natural or planted) in the erosion of traditional community institutions for terms of their timber value, which are usually minimal managing forest lands has also reportedly contributed in the event of low-intensity surface fires such as those to more unwanted forest fires. that commonly occur in India. Estimates could be improved by including the direct and indirect impacts The longer-term impacts of the current pattern of on other sectors including, for example, the effects forest fires on India’s forest ecology and the wider of soil erosion from degraded forest areas on water economy are still poorly understood; however, the supply and the harm from wildfire smoke exposure available scientific evidence supports that fires are on public health. Without credible, empirically based having a degrading effect. Repeated fires in short estimates of the costs of forest fires, it is unlikely that succession are reducing species richness and harming FFPM will be made more of a policy priority. Strengthening Forest Fire Management in India   54 CHAPTER TWO ASSESSMENT OF CURRENT POLICIES, PLANS, AND PRACTICES Forests, being on the concurrent list of subjects under are issued at different levels of government and have the Constitution of India, are the responsibility of distinct functions. National and state forestry policies both the central and state governments, though most provide the overall framework for fire prevention of the forest areas of the country are owned and and management. Standard operating procedures directly managed by the respective state governments. (SOPs) and standing instructions at the state level lay The central government, MoEFCC and agencies out the standard practices and basic requirements. under its purview, are responsible for overall policy Area-specific prescriptions are made as part of guidance, administration of centrally-sponsored forest working plans. Executing these plans requires schemes, coordination of training and research. State sufficient, regular, and predictable funding. governments, on the other hand, being the repository of the manpower of the forest departments carry the 2.1.1 National-level policies and prescriptions primary responsibility of implementing forest fire prevention and management practices. Legislation and policy at the national-level provides the overall framework and direction for FFPM The main implementing agencies to be covered in this in India. Although fire has been used as a land chapter and the next, and their respective roles in management tool by traditional cultures in India for FFPM, are summarized in table 2.1. thousands of years (Pyne 1994), national laws strictly forbid setting fire in forests. Sections 26 and 33 of the Indian Forest Act of 1927 make it a criminal offense to 2.1 POLICIES, PLANS, AND FUNDING burn or to allow a fire to remain burning in reserved FOR FOREST FIRE PREVENTION and protected forests.32 Section 30 of the Wild Life AND MANAGEMENT (Protection) Act of 1972 further prohibits setting fire in wildlife sanctuaries. These laws establish the basis Policies and financing are the foundation for for the strict exclusion of fire from India’s forests—the successful forest fire prevention and management one exception is controlled burning done by the forest (FFPM). In India, policies and prescriptions for FFPM department. Under the India Forest Act of 1927, reserved and protected forests are owned and managed by the government. Hunting, grazing, 32. felling, fuelwood collection, and other extractive activities are prohibited in reserved forests unless specific permission is otherwise granted by the state to rights holders to conduct such activities. Protected forests include any forest lands owned and managed by the government that are not notified as reserved forests, including demarcated and un-demarcated protected forests. In practice, restrictions are generally less strict in these forests areas, and forest-dwelling and forest-fringe communities can hunt, graze their animals, and collect non-timber forest products. 55 Strengthening Forest Fire Management in India TABLE 2.1: AGENCIES INVOLVED IN FFPM IN INDIA Central government entities MoEFCC • Overall policy guidance and standard setting for FFPM • Administers centrally-sponsored schemes and provides funding to states FSI (under MoEFCC) • Issues pre-warning alerts for high fire danger to state forest departments nationwide • Nationwide monitoring and alerts for active fires, provided to state forest departments and the public • Nationwide estimation of burnt forest area ICFRE (under • Apex research organization for forestry in India MoEFCC) • Research institutes under ICFRE include FRI, which has developed training modules for the SFDs and firefighting equipment kits DFE (under MoEFCC) • Coordinates training for frontline staff across the country, including forest rangers and state forest service officers NRSC • Provides near-real time satellite data to FSI for fire monitoring NDMA • Policies and planning for disaster management across the country • So far, has played minor role in FFPM • Deployed NDRF during 2016 forest fires in Uttarakhand • Organized mock drill for forest fire response in April 2017 Military, Paramilitary, • Local units may be called by the SFD to assist in response to large forest fires from time to time and Home Guards State government entities State forest • Primary agency responsible for implementing FFPM department (SFD) • Approves forest working plans for forest divisions within the state, laying out required forest fire prevention activities • Issues state-specific instructions, standard operating procedures, and manuals for field staff • Monitors and collects field-reported data on fire occurrence, burnt area, damages, and forest offences in forest divisions across the state SDMA • Policies and planning for disaster management at the state level • Approves district-level disaster management plans • So far, has played minor role in FFPM District government entities District magistrate • Coordinates among different departments like revenue, health, fire brigade in the event of a large fire • Approves the district-level fire management plan Community/village-level institutions Joint Forest • Primary institution for community-based forest management in India and entry point for SFD Management engagement with communities on FFPM Committee33 • Responsible for developing forest micro-plans for JFMC areas, with technical support from the SFDs • Carries out FFPM activities in coordination with the SFD and may organize labor from the local community for clearing fire lines, conducting controlled area burning, seasonal firewatchers, etc. Other community • A diverse variety of other community-level institutions have evolved in the different states for institutions community-based forest management, such as the Van Panchayats of Uttarakhand and Village Fire Protection Committees in parts of the Northeast Notes: DFE = Directorate of Forest Education; FRI = Forest Research Institute; FSI = Forest Survey of India; ICFRE = Indian Council on Forestry Research and Education; MoEFCC = Ministry of Environment, Forest and Climate Change; NDMA = National Disaster Management Authority; NRSC = National Remote Sensing Centre, ISRO; SDMA = State Disaster Management Authority. Institutional arrangements and rules for JFMCs vary from state to state, and the JFMCs may take various forms, such as the Forest 33. Protection Committees in dense forest areas or Eco-Development Committees in degraded forest areas of Madhya Pradesh. The JFMC is typically under the village-level Gram Sabha, though it may contain members from multiple villages. The Gram Sabha is a village-level body consisting of all voters registered in the electoral rolls for the local Panchayat and is typically responsible for electing the Executive Committee of the JFMC and setting its bylaws. The Gram Panchayat is a village-level administrative body elected by the Gram Sabha and may have an oversight role in the JFMC. See MoEF (undated) and Rose Mary K. Abraham, “Gram Sabha,” Arthapedia, http://www. arthapedia.in/index.php?title=Gram_Sabha. Strengthening Forest Fire Management in India   56 The National Forestry Policy, issued in 1988, makes of human resources and materials in case of only brief mention of forest fires, stating in paragraph an eventuality, (8) active involvement of JFM 4.82: (Joint Forest Management) committees and forest protection committees, including people “The incidents of forest fire in the country is living in and around forest areas and getting high. Standing trees and fodder are destroyed benefits from forests, in prevention and control of on a large scale and natural regeneration forest fires, (9) regular training of communities annihilated by such fires. Special precautions and government staff in prevention and control should be taken during the fire season. Improved of forest fire, (10) emphasis on awareness and modern management practices should be generation programmes including celebration adopted to deal with forest fires.” of a Fire Week to create mass awareness, and (11) enforcement of legal provisions for fire The exigency of special precautions and modern prevention and control” (Saxena 2012: 138). management practices is not elaborated any further.34 Moreover, India does not have a National Forest The 2000 guidelines were not widely known by the Fire Action Plan or Strategy. The need for clearer forest department staff interviewed for this study policy direction on FFPM at the national level has and, according to MoEFCC, are no longer being been recognized by MoEFCC and was echoed by the implemented.35 National Green Tribunal in an August 2017 ruling, which found the ministry “should in consultation with MoEFCC has also continued to provide guidance on the States formulate National policy / Guidelines for specific aspects of FFPM and on formulating working forest fire prevention and control…” (Judgment, M.A. plans in its circulars and letters to the states; however, 397/2017, O.A. 216/2016, sec. 81.i). The NGT has also it has yet to update the 2000 guidelines and various asked the MoEFCC to provide more direction to the other instructions it has issued in the intervening states in preparing and implementing management years and integrate them into a cohesive National plans for fire prevention and control. Action Plan. One of the issues that illustrates the need for such a policy is the uncertainty around the “green MoEFCC had, in fact, issued a set of national guidelines felling ban”, discussed in a subsequent section of this for forest fire prevention and control in 2000. These chapter. guidelines call for: 2.1.2 State-level policies and prescriptions “(1) identification and mapping of all fire prone areas, (2) compilation and analysis Only a few states have issued forest policies and, of of database on forest fire damages, (3) those policies, only a few mention FFPM with varying development and installation of Fire Damage levels of importance. States that have explicitly Rating System and Fire Forecasting system, incorporated aspects of FFPM into their overall (4) making realistic assessment of damage due forestry policies include inter alia Andhra Pradesh, to forest fire, (5) all preventive measures to be Assam, Chhattisgarh, Himachal Pradesh, Madhya taken before the beginning of the fire season, (6) Pradesh, and Telangana. Chhattisgarh’s State Forest deputizing a Nodal Officer in each state to be Policy, for instance, specifically recommends the liaison during fire season with various agencies use of GIS and remote sensing for fire control. In including Government of India on issues addition to preventive measures such as control pertaining to forest fire, (7) constitution of a burning and clearing fire lines, the Himachal Pradesh ‘Crisis Management Group’ in each State/UT Forest Sector Policy (2005) identifies strategies such at State/UT headquarters, Circle and District as engaging fire watchers during the fire season, level during the fire season to closely monitor adopting efficient communication systems and quickly the situation, coordinate various preventive mobilizing adequate human resources with modern measures and arrange adequate enforcement firefighting equipment and tools, particularly in fire A similar call was repeated in the National Forestry Action Programme of 1999 (MoEF 1999). 34. Mr. A.K. Mohanty, Deputy Inspector General of Forests, MoEFCC, telephonic conversation with authors, 8 March 2018. 35. 57 Strengthening Forest Fire Management in India prone forest divisions and ranges. Other strategies fire prevention at the division, range, section, and beat include developing incentives for Panchayats and level, including actions required before and during the communities to involve them in forest fire prevention fire season. The SOP also sets out the requirements and control, and providing educational, extension for a “Model Fire Prevention and Reclamation Plan” and training programmes to create awareness and provides instructions regarding actions to be regarding the causes and ill effects of forest fires. In taken when a fire occurs, as well as post-fire reporting Madhya Pradesh, strategies for fire management that (Odisha 2016). have been identified in the State Forest Policy (2005) include developing a “new fire protection system” 2.1.3 Local-level policies and prescriptions after a detailed study of the effects of fire on forests in the state (including beneficial and harmful effects), as Working plans are required for all state-managed well as using “modern techniques and equipment” to forest areas. Plans are prepared by the officers of the control forest fires. state forest departments and approved by MoEFCC. Guidelines for the preparation of working plans Standing instructions on FFPM have also been are contained in the Working Plan Code. The Code issued by states from time to time. Again, the level of 2014 requires, “Details of all fire cases (range of detail varies across states. In Chhattisgarh, these wise) should be given, for at least past three years to include clearing fire lines, carrying out controlled identity fire prone areas along with specific remarks burning, ensuring NTFP collection without using with regard to severity and burnt area.” It is also fire, engaging fire watchers or JFMC members to suggested that other information, such as on fire monitor forest areas adjacent to agricultural land lines, be provided, and, “Details of the locations along and habitation, using fire watch towers, providing with area affected by fire incidents and appropriate adequate firefighting equipment to field staff, and measures taken may be analysed from the records of keeping water tankers ready. Standing instructions for the fire register and appropriate prescription given”. FFPM in Himachal Pradesh are provided in the state’s Forest fire management is also on a “suggestive list” recently updated Forest Manual. The Manual notes, of exclusive or overlapping mandatory working circles “Detailed information about the causes and thorough (MoEFCC 2014). understanding of the motives behind the forest fire… provide the back ground (sic) fire prevention work” A review of division-level working plans in 11 states (HP 2015: 72). The Manual then lays out a two- suggests that the amount of detail contained in the pronged strategy of public “fire prevention education” plans for fire prevention and management varies and “compulsions” (mandatory prevention measures greatly from area to area. In some cases, the working by the department such as the clearance of fire lines). plans contain exhaustive instructions and guidance Special measures are provided for fire prevention in for field officers; in other cases, fire prevention and chir pine forests and controlled burning in forests not management are given only passing reference. In under regeneration. Specifications for fire detection some detailed working plans, officers have gone to through observatory towers and instructions on the extent of creating overlapping working circles suppressing fire are also provided (HP 2015: 72-76). on grazing, fire protection, and other management concerns—each with specific prescriptions. In 2016, Odisha issued Standard Operating Procedures (SOP) for fire prevention and management, which Good examples of detailed working plans for FFPM stands out as a best-practice example. Issuing a SOP include those for Nainital Division, Uttarakhand and is a standard management practice and an effective Rampur Division, Himachal Pradesh. The Nainital method of communicating the objectives, principles, Division Working Plan discusses aspects of FFPM and actions for FFPM to field staff in the state forest including fire danger rating (with information on departments. The SOPs also provide a medium apparatus to be used for calculating a fire danger for the states to consolidate the various orders, rating index), controlled burning, the maintenance instructions, and letters they have issued from time of fire lines, information on firefighting equipment to time on different aspects of FFPM. It is important (including blowers, water pumps, crew carriers and that the SOPs be updated regularly (i.e., biennially or hand tools such as Macleod, Pulaski and brush hooks), quinquennially). The Odisha SOP sets out a coherent as well as precautions to be taken during firefighting. strategy and clarifies the responsibilities of officers for The Rampur Division Working Plan identifies the Strengthening Forest Fire Management in India   58 causes of fire and provides instructions for firefighting, ensure sustainable non-destructive harvesting of guidelines for controlled burning as well as suggestions NTFPs…and for this, the Committees should be given for firefighting equipment (including brooms, shovels authority to act, monetary and other incentives as and axes), among other requirements for FFPM. genuine stakeholders” (MOEF No. 22-8/2000-JFM [FPD], 24 Dec. 2002). As of 2011, there were 118,213 Not all divisions in all states may require the level of JFM committees managing 22.9 million hectares of detail found in the Nainital and Rampur working forest nationwide,37 equal to about 30 percent of total plans; however, there should be a clear and empirically- forest cover (figure 2.1).38 based method for states to determine which fire-prone divisions warrant special attention through fire risk The JFM program yielded some positive early gains. zoning. Odisha, for example, has created a fire atlas For example, a field study by the International Centre using historic data on the locations of forest fires and for Community Forestry in 2006 found that forest information on fire-sensitive forest types to map zones fires decreased 40 percent at JFM sites in Madhya of high risk. Field officers in high-risk zones have Pradesh, Chhattisgarh, and Jharkhand. A study been deputed to provide an assessment of the causes funded by FAO found a reduction of unwanted fire of fire in their zones, allowing the state to identify and improvements in natural regeneration and areas of shifting cultivation, burning for tendu leaf biodiversity in forests under JFM in Andhra Pradesh. collection, frequent accidental fires, and so on. This Yet, since the mid-2000s, the program’s momentum analysis provides a solid basis for identifying targeted has slowed, and there is less evidence for longer-term measures for fire prevention in the working plans of results in improving forest cover and reducing forest those areas and a way for the state and MoEFCC to degradation. Many of the states with the highest levels determine whether the measures contained in the of JFM participation continue to experience the most plan will address the main causes of unwanted fire. widespread and frequent burning. The way in which From the results of the analysis, the working plans of JFM has been implemented in many areas has also fire-prone divisions should create a Fire (Overlapping) drawn criticism. In practice, the program has often Working Circle, clearly delineating the responsibilities been top-down, with decision-making powers and of officers at different levels for fire prevention and management authority concentrated in the forest control. department, and increasingly low levels of investment and participation on the part of communities. With 2.1.4 Policies and institutions for community forest limited resources and support for management at management the local level, micro-plans and working schemes (providing for forest fire prevention and management) The role of local communities in fire prevention have been implemented in fewer than half of the JFM and management has been institutionalized through committees (Bhattacharya et al. 2010). Joint Forest Management (JFM).36 Under national guidelines, all JFM sites must be covered by working Coexisting with formal state policies and institutions schemes, prepared in consultation with the community for forest fire prevention are indigenous and and approved by the state forest department, which traditional institutions. The uneven uptake of should include prescriptions for fire protection. The JFM seen in figure 2.1 reflects in part the diverse guidelines also require the forest departments and landscape of land tenure and forest rights in states the JFM committees to enter into MOUs. As part of where institutions for forest management by local the MOUs, “All JFM committees should be assigned communities have historically evolved from the specific roles for…fire prevention and control of bottom up. This is especially true in the Northeast, grazing, encroachments and illicit felling as well as the most fire-prone region in the country, where the 36. JFM refers to a cooperative arrangement between a forest-dwelling or forest-fringe community and the forest department, whereby the people of that community organize into a committee to protect and manage local state-held forests in exchange for accessing benefits from the forest, for example, by collecting NTFPs or receiving a share of timber revenues. National guidelines for JFM were introduced in 1990, with refinements in 2000, 2002, and 2009. 37. ENVIS Centre on Forestry and Forest Related Livelihoods, “JFM Committees and Forest Area,” http://frienvis.nic.in/Database/JFM- Committees-and-Forest-Area-Under-JFM_1994.aspx. 38. Total forest cover in 2015 is as per FSI (2015). 59 Strengthening Forest Fire Management in India state forest departments have direct control over only Consultations conducted by the World Bank with about one-third of forested areas (Poffenberger et al. forest communities in Meghalaya reinforced how 2006). Most of the forest lands in the region are held strong community institutions for forest management by communities or are privately-owned. Community can reduce vulnerability to forest fires. According to forests are managed for a variety of purposes, under community members and field-level forest officers a diverse array of traditional institutions. Tiwari et interviewed for this study, fire is most common— al (2013) have documented 11 different categories and the negative effects of frequent fire are most of community forests in the states of Meghalaya, apparent—in forested areas where these institutions Mizoram, and Nagaland, each with varying degrees are the weakest, including individually-owned forests of access and protection against fire (table 2.2). The and law raid (forests managed by groups of villages community forests are managed under an overlapping with few restrictions). set of rules and regulations involving the village (or group of villages), autonomous district councils The ADCs have struggled to fulfill their role as the (ADCs), and the state forest department (Tiwari et legal overseer of a vast estate of community-held al. 2010). Generally, the private and community-held forest in the Northeast. As Poffenberger et al. (2006) forests are under the legal authority of the ADCs, have documented, the ADCs often do not have the which have administrative responsibility for fire administrative capacity, expertise, or resources to craft prevention and management on all non-state forest and implement policies to support the indigenous lands and are required to prepare working schemes and traditional management of forests and fires. for the forests under their jurisdiction. The state Instead, they have leaned heavily on the state forest forest departments assist the ADCs in preparing the departments and the approaches to fire prevention working schemes and have sign-off authority. Direct and management that the state departments have management powers remain with the private owners practiced. The restrictions on the green felling of and villages that control the non-state forest lands. timber and requirements for working schemes in all FIGURE 2.1: PERCENTAGE OF FOREST AREA UNDER JOINT FOREST MANAGEMENT (JFM) 73 71 71 58 56 54 39 38 33 30 26 26 24 22 21 20 18 16 15 11 10 8 2 2 2 2 6 5 0 3 D AR SH AB H AL RA RA U A N H AT S A A A A K SS L ND IM SH UR O ESH ND AM SAM ESH MIR AY BAR DI A A AN H E J AR G T U AD YAN HA DE AR ATA RI ERA HA IKK DE NIP G D A R S D L O H BI AD UN SG EN SH IP N R T A J L O A A S H A I N K P TI B RA TR IL HA JAS PR U O K AK S RA A A A Z IC AR PR G RN P M PR AG MI PR KA EGH N JH YA T AT ES HA A M R A A R K A TAR A R A L N A L & M & H W T H A H T TT H H U AN AD CH M D U AC AC MM M AN U M N A AM I U J D H AR AN Source: Nair (2017) Strengthening Forest Fire Management in India   60 TABLE 2.2: TYPES OF COMMUNITY-HELD FORESTS IN MEGHALAYA, MIZORAM, AND NAGALAND Cluster Forest type Local Size Management Degree of Access Shifting Collection Collection Collection Hunting Grazing name (ha) institution protection to forest cultivation of timbers of NTFPs of resources fuelwood 1 Raid forest A1 35-50 Group of Low All Allowed Allowed Allowed Allowed Prohibited Allowed Villages council 2 Village A2 20-27 Village Low All Allowed Allowed Allowed Allowed Prohibited Allowed forest council 3 Restricted A3 4-10 Village High Prior Prohibited Prohibited Allowed Prohibited Prohibited Prohibited forest council permission 4 Sacred A4 1-100 Village Very High None Prohibited Prohibited Prohibited Prohibited Prohibited Prohibited forest council 5 Clan forest A5 5-20 Clan council Very Low Clan Allowed Allowed Allowed Allowed Allowed Allowed members 6 Cemetery A6 1-30 Church High All Prohibited Prohibited Allowed Prohibited Allowed Prohibited forest 7 Regenera- A7 3-5 Village Very High None Prohibited Prohibited Prohibited Prohibited Prohibited Prohibited tion forest council 8 Bamboo A8 10-15 Village Low All Prohibited Allowed Allowed Allowed Prohibited Allowed Reserve council 9 Recreation A9 10 Village High None Prohibited Prohibited Prohibited Prohibited Prohibited Prohibited forest council/YMA 10 Reserve A10 5-10 Village High Prior Prohibited Allowed Allowed Allowed Prohibited Prohibited forest council/YMA permission 11 Medicinal A11 50 YMA/YLA Very High Prior Prohibited Prohibited Prohibited Prohibited Prohibited Prohibited Plantation permission Notes: YMA = Young Mizo Association; YLA = Young Lai Association; YMA and YLA are indigenous civil society organizations. Source: Tiwari et al. (2013) forested areas ordered by India’s Supreme Court 2.1.5 Funding for Forest Fire Prevention and (discussed in a subsequent section) have added to Management the pressure on ADCs to adopt traditional, state-led management practices. The difficulties faced by the Financial resources for FFPM are provided at both the ADCs are evidenced in Meghalaya, where, of the central and state level. 837,100 hectares of forest under the councils, working schemes have been implemented for only 8,553 The states depend on the central government for hectares. nearly half of their revenue (Busch and Mukherjee 2017), making financial support from the central There are signs that the traditional community government a crucial element for many public policy practices for fire prevention and management in programs at the state and local level. FFPM is no the Northeast are also under strain (Darlong 2002; exception. The federal government provides financing Poffenberger et al. 2006). Communities have started to the states and local entities for forest management growing cash crops like cashew nut, betel nut, through three major Centrally Sponsored Schemes coffee, etc., which require permanent area, but are (CSS) administered by MoEFCC. Funding released also continuing the traditional practice of jhum for from CSSs to the states in recent years is summarized agriculture produce for personal consumption. As a in table 2.3: result, burning cycles have shortened from 30 years to 4-5 years, and regeneration of forests is not taking 1. The National Afforestation Programme (NAP): place at the desired rate. Also, fire prevention practices The NAP delivers central financing to village-level are mostly non-existent. With traditional practices Joint Forest Management Committees (JFMCs) eroding, villages have become increasingly reliant on for rehabilitating and afforesting degraded lands. the state forest department, the ADCs, and hired fire Funds are distributed to JFMCs through local-level watchers to monitor and respond to fires in private Forest Development Agencies, comprised of forest and community-held forests. department staff and village-level representatives. 61 Strengthening Forest Fire Management in India From its inception in 2000 to 2016, the NAP has 2017: 47). Funds are allocated under the FPM distributed INR 3,640 crore (US$ 541.7 million)39 according to a center-state cost-sharing formula, to fund afforestation projects on over 2.1 million with a 90:10 ratio of central to state funding in the hectares.40 The NAP is a 100-percent centrally Northeast and Western Himalayan regions and funded scheme. a 60:40 ratio for all other states (MoEFCC 2017). Funds are released to the states as per approved 2. The Mission for Green India (GIM): Approved Annual Plans of Operation. MoEFCC approved the in 2014, the GIM was established with the goal immediate release of 60-percent of budgeted FPM of realizing the government’s target to increase funds (INR 26.7 crore) in December 2017 to assist forest cover by 50,000 km2 and to improve the with preparation activities in advance of the peak quality of forest on another 50,000 km2. The GIM forest fire season. provides grants to the states for landscape-level forestry projects, with a focus on areas vulnerable Although the FPM is the only dedicated CSS for fire to climate change and with significant biodiversity prevention and management, states may have the and ecological value. As of mid-2017, the GIM was flexibility to direct a portion of NAP and GIM funding being implemented in 13 states and had released a toward forest fire work. Per instructions issued by total of INR 113 crore (US$ 16.9 million) for tree NITI Aayog in August 2016, up to 25 percent of planting activities on 30,000 hectares.41 As with the financing to the states under the CSS may be applied NAP, the GIM is a 100-percent centrally funded as “flexi-funds.” These flexi-funds may be used to scheme. “meet local needs and requirements within the overall objective of any given Scheme,” to “pilot innovation 3. The Forest Fire Prevention and Management to improve efficiency within the overall objective of Scheme (FPM): The FPM is the only centrally- any given Scheme,” or to “undertake mitigation/ funded program specifically dedicated to assist the restoration activities in case of natural calamities” (F. states in dealing with forest fires. The FPM replaced No. 55(5)/PF-II/2011, 6 September 2016). In areas the Intensification of Forest Management Scheme where frequent fires are a cause of forest degradation, (IFMS) in December 2017 (MoEFCC Doc. F. No. it could be argued that FFPM supports the overall 3-1/2017-FPD, 6 December 2017). Up until then, objectives of the NAP and GIM and thus would be an the IFMS had provided financing to the states appropriate item for flexi-funding. for various aspects of forest protection, including fire prevention and management, surveying and Additional funds are available to the states through the demarcation of forested areas, eradication of Compensatory Afforestation Fund Management and invasive species, conservation and restoration, the Planning Authority (CAMPA).42 An ad-hoc CAMPA preparation of forest working plans, and so on. was created following a 2006 ruling by Supreme About INR 52 crore (US$ 7.7 million) in IFMS Court which ordered that compensation paid to the funds were released to the states in the 2015-16 fiscal state governments by users diverting forest lands year, a third of which went to fire prevention and (e.g., for mining, infrastructure building, and other management (MoEFCC 2017). By revamping the projects) be transferred to the fund. The payments IFMS, the FPM has increased the amount dedicated included funding for afforestation on non-forest land for forest fire work. For the 2017-18 fiscal year, INR or degraded forest areas, compensation for the value 49.4 crore (US$ 7.4 million) has been allocated of forgone ecological services, money for watershed under the scheme, with the maintenance of 70,000 protection in catchments where dams are built (Kohli km of fire lines and construction of 60 field crew et al. 2011). For years, payments accumulated unused stations expected as outputs of this support (DEA in the ad-hoc CAMPA, with the states only able to 39. Except where otherwise noted, the average official exchange rate for 2016 is used (INR 67.20 = USD 1.00). 40. Anil Madhav Dave, Minister of State (Independent Charge) for Environment, Forest and Climate Change, response to Rajya Sabha Unstarred Question No. 45, 18 July 2016. 41. Anil Madhav Dave, responses to Rajya Sabha Unstarred Question No. 2771, 27 March 2017, and Rajya Sabha Unstarred Question No. 2922, 12 December 2016. 42. This discussion of CAMPA draws primarily from Kohli et al. (2011). Strengthening Forest Fire Management in India   62 draw a small portion.43 Legislation passed in 2016 was In parallel with the CSSs, states may also have intended to unlock more of these funds for the states;44 dedicated state-sponsored budgeting schemes for however, the permanent institutional mechanism forest management. Uttarakhand, for example, has a created by the CAMPA bill has yet to become State-Sector Forest Fire Protection Scheme. operational (MoEFCC 2017). By the end of 2017, the ad-hoc CAMPA had paid out INR 12,241.5 crore An assessment of individual state budgets and (US$ 1.82 billion) to the states to fund tree planting on spending on FFPM is beyond the scope of this study; 8,130 km2 (including on 3,270 km2 of non-forest land however, audits by the Comptroller and Auditor and 4,860 km2 of degraded forest land).45 Yet, as of General (CAG) indicate that funding shortages September 2017, about INR 50,500 crore (US$ 7.51 do exist, as reflected by insufficient staffing and billion) sat unused in the ad-hoc fund.46 equipment for executing FFPM. In one recent audit, CAG examined budgeting for FFPM in for four of The stated goals of CAMPA to “accelerate activities 35 forest divisions of Uttarakhand from 2013-14 to for preservation of natural forests, management of 2015-16. In these three years, the divisions requested wildlife, infrastructure development in the (forestry) a total of INR 775 lakh (US$ 1.15 million) for FFPM sector and other allied works” make the fund an under the centrally sponsored Intensification of amenable—though underutilized—source of financing Forest Management Scheme (IFMS) and the state- for FFPM.47 Data on the use of compensatory funds level Forest Fire Protection Scheme (FFPS). Each year, by the various states for FFPM purposes are lacking. they were allotted between 40 percent and 65 percent Anecdotally, it appears that at least some of the of their total requested amount (CAG 2017: 41). As a CAMPA financing has gone toward FFPM. State forest result, CAG found: department officials interviewed by the World Bank in Odisha, for example, said they used CAMPA funds to “The (Forest) Department lacked insufficient purchase of hand tools, leaf blowers, and vehicles for funds for preventing and controlling forest forest firefighters. fires which translated into shortages of essential-firefighting equipment, vehicles, Reforms to the formula for distributing tax revenue communications as well as manpower. Shortages from the central government among the states may of equipment, accessories and vehicles required provide additional incentives for states to improve how for fire-fighting in the fire season ranged from they manage their forests. The reforms, introduced in 31 to 100 percent while shortage of manpower February 2014 by India’s 14th Finance Commission, ranged from 16 to 55 percent in cadres of provide that 7.5 percent of the tax revenue transferred foresters and forest guards” (CAG 2017: 40). from the central government among the states will be determined according to the area of each state’s forest Responding to CAG’s audit, Rajender Mahajan, cover. The government estimates that this translates to Principal Chief Conservator of Forest and Head INR 10,956 (US$ 174) per hectare of forest per year of Forest Force for the Uttarakhand state forest in fiscal transfers (GoI 2015). However, these funds department, explained: are transferred to the states’ general accounts with no requirements for spending on forest management or FFPM. 43. Amitabh Sinha, “CAMPA: The Manager of Afforestation Funds,” The Indian Express 25 May 2016, http://indianexpress.com/article/ explained/campa-afforestation-bill-rajya-sabha-green-india-mission-narendra-modi-2817475/. 44. MoEFCC, “Compensatory Afforestation Fund Bill, 2016 Passed by Rajya Sabha,” Press Information Bureau, Government of India, 28 July 2016, http://pib.nic.in/newsite/mbErel.aspx?relid=147937 45. Mahesh Sharma, Ministry of State in MoEFCC, Rajya Sabha Unstarred Question No. 203, 18 December 2017. 46. Dhananjay Mahapatra, “Don’t Allow Govts Access to Forest Funds: SC Amicus,” Times of India, 12 December 2017, https://timesofindia. indiatimes.com/india/dont-allow-govts-access-to-forest-funds-sc-amicus/articleshow/62030590.cms. 47. The overarching objectives and core principles of the state-level CAMPA are quoted from MoEF (2009 b: 3). 63 Strengthening Forest Fire Management in India “There was meagre budgetary support to tackle these staffing shortages are not due entirely to the forest fire. Just five crore in year 2015 and, lack of funds, funding does play a part. In the case during massive forest fire of 2016, it was Rs 22 of Karnataka, CAG attributed the shortages in part crore, we place demand of Rs 446 crore before to the underutilization and suboptimal allocation of the state government but were sanctioned merely available funds. Rs 22 crore. It is next to impossible to depute manpower, purchase equipment, maintain fire lines or hire people for controlled burning.”48 2.2 FOREST FIRE PREVENTION AND MANAGEMENT (FFPM) Shortfalls in resources at the field level have been PRACTICES documented in other states, too. In Madhya Pradesh in 2013, CAG found 3,870 posts for forest officers at the rank of range officer or below had gone unfilled.49 Effective FFPM entails a continual management In Andhra Pradesh, about 60 percent of the sanctioned process, as illustrated in figure 2.2 below. The positions in the forest department were empty, with stages of this process include prevention, detection, only 915 forest beat officers on the ground, about one suppression, and post-fire management. Prevention is per 40 km2.50 In Karnataka, about one-quarter (2,929) the beginning and most critical stage of the process. of all posts in the forest department were unfilled, At the end of the process, after a fire is extinguished, with CAG noting a “large number of vacancies… post-fire management should aim to inform and amongst the frontline staff ” (CAG 2014: 34). Although improve future prevention activities, hence the cycle. TABLE 2.3: FUNDING FOR FOREST MANAGEMENT UNDER CENTRALLY SPONSORED SCHEMES AND OTHER PROGRAMS, 2011-2016 2011-12 2012-13 2013-14 2014-15 2015-16 Funding/expenditure INR crore USD INR crore USD million INR crore USD INR crore USD INR crore USD category million million million million NAP funds released 303 57 193 33 258 42 244 38 94 14 GIM funds released --- --- --- --- 13 2 0* 0* 70 10 IFMS funds released 63 12 41 7 59 10 56 9 52 8 Ad-hoc CAMPA funds 942 176 1,029 176 1,085 178 1,980 309 1,402 209 released Total MoEFCC budget/ 1,982 371 1,753 299 1,890 310 1,514 236 1,521 226 expenditures Total central government 1,286,997 240,843 1,393,577 237,821 1,541,466 252,577 1,670,220 260,354 1,761,812 262,193 expenditures Notes: * GIM funds before 2015-16 were allocated for preparatory activities prior to formal approval of the mission; NAP = National Afforestation Programme, GIM = National Mission for Green India; IFMS = Intensification of Forest Management Scheme; CAMPA = Compensatory MoEFCC = Ministry of Environment, Forest and Climate Change; INR converted to US$ at official exchange rate for period average; INR and US$ are in nominal amounts not adjusted for inflation. Sources: NAP data from Indiastat.com and Rajya Sabha Unstarred Question No. 2922 (12 December 2016); GIM data from Lok Sabha Unstarred Question No. 361 (19 July 2016) and Rajya Sabha Unstarred Question No. 2922 (12 December 2016); IFMS data from Indiastat.com; MoEFCC budget/expenditure data from Union Budgets for 2013-14 to 2017-18; central government expenditure data from Department of Economic Affairs, Ministry of Finance, India Public Finance Statistics 2015-16 (August 2016). 48. Seema Sharma, “CAG Reports Serious Irregularities in Forest Fire Fighting Measures,” The Times of India, 11 May 2017, https:// timesofindia.indiatimes.com/city/dehradun/cag-reports-serious-irregularities-in-forest-fire-fighting-measures/articleshow/58631115. cms. 49. Shashikant Trivedi, “MP Govt Lacks Action Plan to Curb Forest Crimes: CAG,” Business Standard, 24 July 2014, http://www.business- standard.com/article/economy-policy/mp-govt-lacks-action-plan-to-curb-forest-crimes-cag-114072401225_1.html. 50. Sumit Kumar Onka, “Acute Shortage of Beat Officers in Andhra Pradesh,” Deccan Chronicle, 25 July 2016, https://www.deccanchronicle. com/nation/current-affairs/250616/acute-shortage-of-beat-officers-in-andhra-pradesh-forests.html Strengthening Forest Fire Management in India   64 FIGURE 2.2: THE FOREST FIRE warning and fire danger rating systems are also part PREVENTION AND of the prevention process and allow fire managers to MANAGEMENT (FFPM) put in place an appropriate state of readiness when CYCLE hazardous conditions develop that could lead to more severe fire behavior. Forest-using communities play a pivotal role in fire prevention in India. The need to improve the effectiveness of community engagement PREVENTION on forest fires is elucidated further in Chapter 3. Challenges to effective prevention of forest fires in India identified as part of the survey of forest officers DETECTION in 11 states included: lack of public awareness and engagement; difficulties in changing traditional community practices with the use of fire; the inaccessibility and ruggedness of fire-affected forests; limitations in the forest department’s equipment, SUPPRESSION technology, and infrastructure; shortages of labor; and insufficient financial resources (figure 2.3). While the issues identified by officers varied across states, the first or second most-mentioned challenges in all states (except Kerala) were difficulties with public POST-FIRE MANAGEMENT engagement and the lack of department resources (table 2.4). Source: Authors 2.2.1.1 Fire lines and controlled burning The challenges identified by officers have led to gaps in the implementation of measures for fire This section evaluates each of the stages of the FFPM prevention, including in the maintenance of fire process, with a focus on implementation and on- lines and controlled burning to remove built-up the-ground practices. The analysis aims to identify fuel loads. Only half of the officers responding to constraints or shortfalls in implementation, with an the survey said that the fire lines required in their eye toward identifying opportunities for improvement area were all clear (47 of 94). Figures 2.4 provides that are further elaborated in Chapter 4. Institutional state-wise information, which indicates that not all coordination issues and engaging with communities fire lines are cleared in any of the states surveyed. Of throughout the FFPM process is dealt with separately those officers who noted gaps in the maintenance of in Chapter 3. fire lines, most cited a lack of resources as the main reason for the lines in their area not being cleared. 2.2.1 Prevention The one exception was in Uttarakhand, where most respondents said a ban on green felling in areas The aim of effective prevention is not to entirely exclude above 1,000 m was the main reason behind fire lines fires from forests, but rather to avoid damaging and not being maintained. According to FSI (2015), more unwanted fires, thus maximizing the environmental than 70 percent of the state’s forest cover can be found benefits of fire while minimizing its adverse impacts. above an altitude of 1,000 meters, underscoring the The most common methods of prevention employed importance of scientific fire prevention measures in by forest departments in India include the clearance such areas. of fire lines and conducting controlled burning to limit fuel loads. Other methods may include silvicultural Varying degrees of information on fire lines were practices such as selective thinning and planting available from the state forest departments. Table fire-adapted tree species in fire- prone areas. Early 2.5 shows the total length of fire lines in the states 65 Strengthening Forest Fire Management in India FIGURE 2.3: BIGGEST CHALLENGES TO EFFECTIVE FOREST FIRE PREVENTION IDENTIFIED BY RESPONDING OFFICERS Lack of awareness/education 34 Traditional practices (non-specific) 24 Public engagement (non-specific) 15 Conflict or animosity toward forest department 7 Public engagement Missing or perverse incentives 3 Negligence 1 Inaccessibility and difficult terrain 21 Forest structure or species composition 9 Water availability 2 Weather or climate conditions 2 Environmental factors Environmental factors (non-specific) 2 Invasive species 1 Lacking equipment, tech, infrastructure 22 Labor shortage 18 Insufficient financial resources 14 Department resources Knowledge gaps (expertise, training, etc.) 5 Poor coordination between agencies 8 Problems with policies, laws, regulations 6 Institutional factors Weak implementation of existing plans/policies 4 Low priority given to fire prevention 2 Lack of management plan 1 Land encroachment 2 Illegal/criminal activity Illicit activities 2 Illegal felling 1 Lack of alternative livelihoods 1 Lack of development Lack of agricultural land/capital 1 Poverty 1 0 10 20 30 40 Frequency (times mentioned by surveyed officers) Note: responding officers = 96; each responding officer may mention more than one challenge Source: World Bank survey of state forest department officers, April-August 2017 that responded to requests for data.51 Chhattisgarh officers did indicate that this was not the case in both and Telangana indicated that all fire lines stipulated states. The length of fire lines maintained annually in in Working Plans for SFD lands are maintained Himachal Pradesh and Uttarakhand is less than the annually, although in the forest officials’ survey, some length stipulated in the Working Plans. In Tripura, 51. Data request sheets were sent by MoEFCC to nodal officers in the state forest departments in March 2017 to collect basic information about forest area, fire lines, controlled burning, causes of fire, reporting of fire incidents, and burnt area in each state. As of the time of writing, data sheets had been received by 7 states (Chhattisgarh, Himachal Pradesh, Kerala, Meghalaya, Telangana, Tripura, and Uttarakhand). Strengthening Forest Fire Management in India   66 TABLE 2.4: RANKING OF CHALLENGES TO EFFECTIVE FOREST FIRE PREVENTION IDENTIFIED BY 67 RESPONDENTS (1ST = MOST-MENTIONED) Most mentioned challenge Second most mentioned challenge Category of challenge Uttarakhand Tripura Telangana Odisha Meghalaya Madhya Kerala Jharkhand Himachal Chhattisgarh Assam Pradesh Pradesh Public engagement 1st 1st 1st 1st 1st 1st 2nd 2nd 1st 1st 1st Environmental factors 3rd 3rd 2nd 3rd 3rd 3rd 1st 3rd 3rd 4th 4th nd nd nd nd st st rd st nd nd Lack of department 2 2 2 2 1 1 3 1 2 2 2nd resources Institutional issues 4th 4th 2nd 3rd 4th 4th 4th 2nd Illegal activity 4th 2nd 2nd th rd Economic development 4 3 Note: responding officers = 96; Rankings are tied if categories are mentioned the same number of times. Source: World Bank survey of state forest department officers TABLE 2.5: INFORMATION ON FIRE LINES PROVIDED BY STATE FOREST DEPARTMENTS Length of fire lines (km) Chhattisgarh Himachal Kerala Telangana Tripura Uttarakhand Meghalaya Pradesh Fire lines on SFD lands (per 91,001 2,750 No info 3,866 No info 16,443 Not working plans) prescribed Fire lines mapped and digitized 0 0 0 3,866 0 0 0 on a GIS layer Fire lines maintained annually 91,001 1,000 15,000 3,866 No info No info 274 Fire lines on non-SFD lands No info No info 0 No info No info No info 130 Note: SFD = State Forest Department. Source: State forest department data sheets provided to World Bank Strengthening Forest Fire Management in India FIGURE 2.4: SHARE OF FOREST DEPARTMENT SURVEY RESPONDENTS WHO SAID FIRE LINES IN THEIR AREA WERE ALL CLEARED PER THE WORKING PLAN Uttarakhand 56% 44% Yes Tripura 56% 44% No Telangana 50% 50% Odisha 29% 71% Meghalaya 64% 36% Kerala 56% 44% Jharkhand 10% 90% Himachal Pradesh 50% 50% Chhattisgarh 57% 43% Assam 33% 67% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percent of responses Note: responding officers = 93 Source: World Bank survey of state forest department officers, April-August 2017 it is unclear if fire lines exist or are required, while 96); however, of those officers, nearly two-thirds said in Meghalaya fire lines are reportedly maintained in that controlled burning was not regularly performed both SFD and non-SFD lands even though they are (43 of 70). Reasons for failing to perform controlled not required as per the Working Plan. Telangana is burning on all areas as required mirrored those given the only state to have digitized locations of fire lines. for gaps in fire line maintenance, with a lack of resources being cited as the main constraint in all states surveyed. Similarly, data on the performance of controlled burning is not readily available. Based on information 2.2.1.2 Early Warning and Fire Danger Rating provided by state forest departments, controlled Systems burning is most commonly practiced in Uttarakhand and Himachal Pradesh, which have large areas of chir Early warning is a process that provides warning weeks pine forests. In some states, the area of controlled or months in advance of deteriorating conditions that burning required is not specified, or controlled could easily translate into severe fire behavior, for burning is not required at all. example following a below-average monsoon season or with severe drought. Early warning systems can Three-quarters of forest officials responding to the help fire agencies ensure that an appropriate level survey said that controlled burning was required in of readiness is in place and that preparatory work their area per forest department working plans (72 of for a potentially severe fire season is completed. Fire Strengthening Forest Fire Management in India   68 Strengthening 69 Forest Fire Management in India   69 Strengthening Forest Fire Management in India Strengthening Forest Fire Management in India   70 Box 2.1: Court-Ordered Green Felling Bans and Fire Prevention Despite clear statements from MOEFCC and the Parliamentary Committee that there are no legal obstacles to the clearing of fire lines and fuel removal, there appears to be some confusion on this issue. The ban on felling green timber arose from a series of court decisions and actions taken by the central and state governments over many years, beginning in 1980 (Uttarakhand Biodiversity Board 2017). Responding to the Chipko andolan, a popular protest movement led by rural women against logging in the Himalayas, the then Prime Minister Indira Gandhi called for a ban on felling green timber in the hill forests of Uttar Pradesh (present-day Uttarakhand). A year later, Uttar Pradesh instituted a ban on felling in areas above 1,000 m (G.O. No. 1913/1-81, 18 Mar 1981). In the face of recommendations of an expert committee in 1982 that tree removal should be allowed in these areas per scientific prescriptions in the forest working plans, in 1986 the state government issued a follow-on order that upheld the ban without exception (G.O. No. 6241/14-2-124/82, 21 Aug 1986). The ban was extended by the state in 1993 and again for another 10 years in 1996 (G.O. No. 6373/14-3-700(385)/93, 15 Sept. 1993, and G.O. No. 9371/14-2-96-124/1982, 27 Sept. 1996). Following the lead of Uttar Pradesh, green felling bans were put in place at the state level in Himachal Pradesh in 1986, and other states outside the Himalayan region, including Gujarat in 1986, Karnataka in 1990, and Odisha in 1992 (Springate-Baginski and Blaikie 2007). India’s Supreme Court intervened in 1996, issuing a pivotal order in the case of T.N. Godavarman Thirimulkpad v. Union of India that bolstered the movement in the states to halt felling (Order W.P. 202/1995 and W.P. 171/1996, 12 Dec. 1996). In its order, the Court laid out a set of general instructions for the country, directing, “The felling of trees in all forests is to remain suspended except in accordance with the Working Plans of the State Governments, as approved by the Central Government.” The Court further specified that for Himachal Pradesh, Jammu and Kashmir, and the hill regions of Uttar Pradesh and West Bengal, “There will be no felling of trees permitted in any forest, public or private.” In Tamil Nadu, it said, “There will be a complete ban on felling of trees in all ‘forest areas.’” In Arunachal Pradesh, “there would be a complete ban on felling of any kind of trees therein because of their particular significance to maintain ecological balance needed to preserve biodiversity.” After the 1996 ruling was issued, Uttar Pradesh and Himachal Pradesh filed affidavits in the Supreme Court stipulating that no felling should be carried out above 1,000 m. The ambiguity created by the general provision that cutting down green trees is suspended except where prescribed by working plans, and the more restrictive language for the handful of states has been a source of great confusion. Forest managers in the hill regions of present-day Uttarakhand have generally interpreted the Court’s order to mean that fuel removal, fire line clearance, and other fire prevention activities involving the cutting of green trees in areas above 1,000 m continue to be prohibited. A similar situation has been observed in other Himalayan states. Since 1996, the Supreme Court has continued to issue a string of orders in the ongoing T.N. Godavarman case expounding on the green felling ban. A January 1998 order reinforced the Court’s position that working plans should be prepared and implemented for all forest divisions and stipulated that “future felling will remain suspended” in areas that fail to prepare plans within the prescribed timeframe of two years (15 Jan. 1998). Exceptions have been made for extraction and use of forest resources by local communities. A February 2000 order prohibited “the removal of dead, diseased, dying or wind-fallen trees, drift wood and grasses, etc. from any National Park or Game Sanctuary or forest” (I.A. 548, 14 Feb. 2000). A May 2001 order clarified that working schemes should also be required for cutting green trees in forest areas outside of the lands managed by the forest department (I.A. 295, 12 May 2001). A February 2002 order further clarified that the green felling ban does not apply to bamboo or cane 71 Strengthening Forest Fire Management in India (I.A. 707, 18 Feb. 2002). The order would in principle allow for the clearing of stands of dry flowered bamboo that pose a fire hazard. Further interpretation of the Court’s ban on the cutting or removal of trees from protected areas has held that such activities may be allowed if they support biodiversity and wildlife conservation (which is not necessarily the same as fire protection). As noted in a 2002 order issued by Karnataka state, amending the state government’s total ban on green felling, “Many of the activities of salvaging of dead and fallen timber and flowered bamboo, gradual reduction in number of Teak trees in plantation to encourage other indigenous species, removal of exotics like Eucalyptus, etc. indeed contribute to the improvement of the habitat for wildlife” (Karnataka G.O. FEE 101 FAP 2001, 23 Oct. 2002). However, these actions can only be carried out only upon obtaining special permission from the Court or a Court-appointed committee. The Standing Committee of the National Board for Wildlife has been designated by the Court to handle permissions in protected areas.52 The Court has instructed states to appoint suitable committees for handling of permissions to remove dead, dying, and diseased trees in the hill regions (Uttarakhand Biodiversity Board 2017). The National Green Tribunal (NGT) has added to the mix of court decisions and state orders banning the felling of trees. For example, in May 2016, the NGT issued a total and complete ban on the felling of trees in Punjab after a complaint of trees being removed for infrastructure projects: “we hereby restrain the State of Punjab…and Departments of State of Punjab from felling and cutting of any tree in the entire State of Punjab without specific permission of the Tribunal” (Order, Items 16-17, O.A. 161/2016 and O.A. 162/2016, 8 Jul. 2016). This followed an order by the NGT in November 2014 to “restrain any person, company, authority from carrying out cutting of trees from forests anywhere in the country without obtaining environmental clearance from MoEF [the Ministry of Environment and Forests]/SEIAA [State Level Environment Impact Assessment Authority] and license from the competent authorities.”53 Thus, the issue of the green felling ban remains unresolved. The courts have generally permitted the felling of trees per forest working plans approved by MoEFCC, but it is still unclear whether trees may be removed to mitigate fire hazards in those specific states and areas where the Supreme Court has categorically banned the felling of trees in all forest areas, public or private. These areas include fire- prone forests in Himachal Pradesh, Jammu and Kashmir, and the hill regions of Uttarakhand. The Parliamentary Committee has disagreed with the position that there are legal barriers standing in the way of fire prevention in these areas, asserting that “the ball is in the court of the Central Ministry of Environment, Forest and Climate Change to plan to remove dead and fallen tree (sic) even in the protected forest areas” (Parliamentary Committee 2016: 110). In meetings and interviews with the World Bank team, forest officers in Uttarakhand said they would seek a waiver to the ban and have sent a proposal to MoEFCC to resume clearing fire lines and conducting selective thinning in reserve forests in areas above 1,000 m. Studies done by the Forest Research Institute, Dehradun and the research wing of the Uttarakhand forest department on the impact of the green felling ban in the hill areas of Uttarakhand support such a move, finding that forest cover and treed species diversity had not improved in unfelled plots versus plots where felling had been done as part of silvicultural interventions, while fine fuel loads in the unfelled plots were greater, presenting more of a fire hazard.54 52. See Supreme Court Record of Proceedings, Items 301-308, 311, 314-319, Sections PIL, XIA, IX, X, XVIA, IVA, 5 Oct 2015. 53. The Hindu, “NGT Bans Cutting of Trees Without Clearance,” 26 Nov. 2014, http://www.thehindu.com/todays-paper/tp-national/ tp-newdelhi/ngt-bans-cutting-of-trees-without-clearance/article6634799.ece. 54. The study by Manoj Chandran (2012) for the forest department recommended the “scientific green felling of Chir pine in fully regenerated sun facing slopes” be allowed, per working plans and with the approval of MoEFCC and the Court. The study for FRI by the Mishra Committee (2011) on the effects of the ban on deodar, kail, spruce, and fir forests found the composition of tree species was the same in felled and unfelled plots and that the ban has not improved forest cover or regeneration by those species. At the same time, shrub level vegetation had increased, and there was more soil carbon and organic matter in the unfelled plots (cited in Uttarakhand Biodiversity Board 2017: 35-37), thus the fine fuel load was greater. Strengthening Forest Fire Management in India   72 danger rating systems (FDRS) warn of short-term fire Developing the FDRS in these countries took many potential and allow fire agencies to quantify different years, and refinements continue to be made. Reflecting aspects of fire behavior, for example, how fast fires on the experience of Canada, Taylor and Alexander are likely to spread, how intensely they may burn (2006) have identified key elements of an effective under current conditions, and how difficult they national FDRS: may be to control.55,56 FDRS are intended to inform fire managers and other responsible agencies about • Indicators backed by empirical scientific research hazardous fire weather conditions so that they can and tailored to the fire environments in which they ensure an appropriate state of readiness, alert the will be deployed; public of the danger, and take actions to prevent or mitigate damaging fires (e.g., by putting in place • A reliable infrastructure to gather, analyze, restrictions on the use of fire). As a decision-support disseminate, and archive data for the FDRS. This tool, FDRS may enable fire managers to allocate their would include physical infrastructure that need to resources for FFPM in a more efficient and cost- be built and maintained, such as weather stations, effective way (Taylor and Alexander 2006). as well as a supporting institutional infrastructure, such as standards and policies to clearly define the Early warning systems, such as seasonal fire weather FDRS and assign roles and responsibilities for its forecasting, do not currently exist in India. However, production; several states do conduct fire risk zonation as part of their prevention and preparedness planning • Guidelines, decision aids, and training for the to identify areas that are vulnerable or frequently application of the FDRS. An operational FDRS affected by fire. Most typically, states have done this by should trigger actions for various levels of fire mapping historic patterns in satellite fire detections. danger by forest managers, local communities, fire Telangana provides a good example of this practice responders, and others. Ongoing support should (box 2.2). Tamil Nadu has also analyzed historic be provided to make sure these users to make sure patterns of forest fires to identify high-risk areas (box they understand the FDRS and what these actions 2.3). Other states include inter alia Madhya Pradesh, are; Chhattisgarh, and Odisha. FSI conducted a nationwide assessment of forest fire vulnerability (FSI 2012). Such • Cooperation between fire management agencies exercises may be completed every few years as part of and fire scientists for ongoing development of the developing forest working plans or on an annual basis FDRS and to ensure research meets the practical before the start of each fire season. Because people needs of those responsible for applying the FDRS. are the dominant influence on the forest fire regime, and fires are often set in the same areas year after year Underlying the system should be a clear mandate for promoting fodder growth, clearing forest litter, for the creation of the FDRS and a statement of or obtaining certain NTFPs, by showing where fires objectives defining what the system should do. Once have occurred in the past, fire risk zonation maps can this is done, practical decisions about what to measure, help inform fire managers where they are likely to when to measure, where to measure, how to measure, occur in upcoming seasons. Data about historic fire how to integrate measurements, and how to apply occurrence may be also overlaid with information measurements can be made (Alexander 2008). about ecologically sensitive areas, protected habitat, plantations, regeneration zones, etc. to further identify The recent experiences of South Africa and Indonesia priorities for fire protection. are also instructive of some of the key considerations in developing a national FDRS (boxes 2.3 and 2.4). FDRS have been used by fire managers for the better Perhaps even more important than the outcomes part of a century in North America and Australia. of the FDRS development process was the process 55. Ross Smith, “Land and Forest Fire – Adoption vs Adaption of Fire Danger Rating Systems,” presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017. 56. Fire danger is defined as a “general term used to express an assessment of both fixed and variable factors of the fire environment which determine the ease of ignition, rate of spread, difficulty of control and fire impact” (Merrill and Alexander 1987). Factors influencing fire danger may include weather, fuel, and topography. 73 Strengthening Forest Fire Management in India Box 2.2: Fire Risk Zonation in Telangana, India One of Telangana’s successes in FFPM has been the creation of fire risk maps, working with field staff in the most fire-prone areas to assess why those areas experience more fires than others and to identify appropriate solutions for the management of those areas, from providing extension services to fire- reliant communities to increasing enforcement. The IT wing of the state forest department (SFD) has carried out a forest fire risk assessment by integrating various parameters governing forest fires based on their degrees of influence on fire, using modern IT and Geomatics tools. Forest fire risk zonation mapping was done for the entire state in 2003, earning the department the “Silver Icon” award from the Government of India in 2004. Factors influencing fire occurrence and behavior considered include vegetation (canopy density and vegetation type), topography (slope and aspect) and proximity (roads and villages). Based on field observations, past fire data and vegetation characteristics, variables were weighed in order of influence (vegetation, aspect, slope and road etc.) and the modeling was carried out. The maps are being used for forest fire management. The fire risk zonation maps have been prepared and communicated for the use of field officers so that they can take preventive measures before the commencement of the fire season and avoid or minimize fires. An example of Bhadrachalam South Division is provided below. FIGURE B2.1: FIRE RISK ZONATION MAP, BHADRACHALAM SOUTH DIVISION, TELANGANA Source: Telangana (2015); P. Raghuveer, Forest Department, Government of Telangana, India, “Forest Fire Prevention and Management – Experiences from Telangana”, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017 Strengthening Forest Fire Management in India   74 Box 2.3: Assessing Forest Fire Hazards and Risk in Tamil Nadu Highlighting a lack of studies carried out on forest fire vulnerability, the Tamil Nadu Forest Department has analyzed information covering 2006-2015 on forest fires within the state. The sensitivity to forest fires within different forest density classes57, forest types and beats was studied, and it emerged that Moderately Dense Forests were the most prone to forest fires between 2006 and 2015, followed by Open Forests and Scrub Forests. Moreover, Tropical Dry Deciduous Forests were found to be more prone to fire (accounting for 43 percent of the 3272 fire incidents that occurred over 2006-2015), followed by Tropical Moist Deciduous Forests (in which 14 percent of fire incidents were recorded over these years). Furthermore, fire-prone beats were categorized into 5 sensitivity classes based on the number of fire detections, and 41 beats were classified as being either highly or very highly prone to forest fire. It was also highlighted that the road network provides a gateway for human activities within forests and plays an important role in the origin of forest fires (with about 73 percent of hotspots sensed within a 1.5 km buffer of the roads over this decade) and it was noted that fire initiation was the highest during the early hours of the day. Based on the annual pattern of fires over this decade, the number of fires was found to have been above average in 2007, 2009, 2012 and 2014, with more fire incidents in alternate years. Around 72 percent of fire incidents over this decade was detected during the months of February and March. Source: Tamil Nadu Forest Department (2017) itself, clearly defined, and the ongoing trials and The requirement to create a national FDRS was set improvements that continued after the initial review forth by the Government of India in 2001 (Govt. was done. of India vide No.9-6/99-FFD, 22 June 2001). The Department of Science and Technology was instructed The experiences of South Africa and Indonesia to take the lead on the FDRS, though this never illustrate how the chances of a FDRS being accepted happened. Instead, the task of early warning has been and utilized are inherently much more probable if taken up by FSI, which began issuing “pre-warning a FDRS is developed or adapted locally than if an alerts” for dangerous fire conditions nationwide in existing FDRS is imported directly. Especially in the 2016 after piloting its system in Uttarakhand in 2015. case of South Africa, the country ultimately decided FSI’s stated goal in producing pre-warnings “is not to to abandon the American system it had begun predict forest fire locations but to identify areas which implementing in the mid-2000s and to replace it are more vulnerable to severe forest fires” (FSI 2017a). with the old Lowveld system that fire managers Though the pre-warning alerts do not yet constitute a had been using for decades. The need for a locally full-fledged FDRS with each of the elements outlined appropriate system is also evident for India, where the above by Taylor and Alexander (2006), FSI is moving difficulties and pitfalls in calibrating a FDRS for a wide in that direction. range of forest types in different latitudes are plain. The prospect of developing a local system, albeit Figure 2.5 depicts the current system for FSI’s pre- with components from existing systems, is likely to be warning alerts, as modified in 2017. FSI determines more successful. areas of high fire danger using data on forest types, The Forest Survey of India (FSI) classifies forests as follows: (i) tree canopy density of 70 percent and above: very dense forests (VDF); 57. between 40 and 70 percent: moderately dense forests (MDF); between 10 and 40 percent: open forests (OF); and less than 10 percent: scrub areas. 75 Strengthening Forest Fire Management in India locations of recent burning, and weather conditions. then overlays vulnerable forests with areas currently FSI maps forest cover and forest types has identified experiencing drought. Areas with rainfall forecasted which forest types are most susceptible to fire (see by IITM Pune for the next 24, 48, 72, or 96 hours FSI 2012).58 Using hotspot data from MODIS, FSI are then removed. The result is a map of 5 km x 5 screens areas with vulnerable forest types to identify km grid cells containing locations determined to locations where there has not been recent burning have high fire danger. FSI distributes weekly alerts or only low-intensity fires have occurred. Fire with these locations via emails to nodal officers and intensity is measured by fire radiative power. The field staff in the forest departments. Alerts contain logic is that if a high-intensity has already burned an KML files to view areas of high fire danger on area, fuels will have been depleted and the chance Google Earth. Though FSI does not send pre-alert of repeated burning in the same season will be low. warnings to public users, the forest departments may Using interpolated rainfall and relative humidity data disseminate information on fire danger to the public from weather stations in the CRIS-IMD network, FSI through their own channels. Box 2.4: South Africa’s Lowveld Fire Danger Index South Africa’s national fire danger rating system (NFDRS) exemplifies how a system tailored to local conditions may receive greater acceptance and ultimately prove more effective than imported systems. The development of a national FDRS was mandated by the National Veld and Forest Fire Act of 1998. The Act set in motion an extensive review of rating systems and indices that had been employed in South Africa and other countries (Willis et al. 2001). Eight systems were initially considered, including the Swedish Angstrom Index, the Russian Ignition Index, the Canadian Fire Weather Index (FWI), the French NFDRS, the American NFDRS, the Australian (McArthur) NFDRS for forests, the Australian NFDRS for grasslands, and, lastly, the Lowveld Fire Danger Index from South Africa. The review identified the American, Australian, and Canadian systems as the best candidate models but concluded that further work was needed to investigate how the models performed in the different regions of the country before any specific model could be formally adopted. At the end of this process, South Africa’s Department of Water Affairs and Forestry (DWAF) issued a notice in 2005 that it would use two of the fuel models from the American NFDRS, with additional fuel models to be customized for specific regions after more research (Isaacs 2004; DWAF 2005). Ultimately, however, DWAF (now the Department of Agriculture, Forestry and Fisheries) decided to revert to using the old Lowveld Fire Danger Index (DAFF 2013). The Lowveld index, adapted from a system originally developed in 1968 in Zimbabwe (Rhodesia), had been used for decades in the vast savanna and open woodlands that stretch across the province of Mpumalanga and the surrounding region. Though the system has its weaknesses and was never intended for use in other parts of South Africa such as the Cape region, subsequent reviews using historical weather data and fire maps have found that it works almost as well as the Canadian FWI and the American NFDRS in predicting fire potential. Perhaps more importantly, it is easy to understand and calculate, and fire managers already had many years of using the model in practice (Oxford 2017). Forest fire experts in South Africa lobbied successfully for the switch to the old Lowveld-based model, and the model was adopted by DAFF in 2013 (DAFF 2013). Forest cover is mapped on a biannual basis. Forest types were mapped in 2011. The survey of forest types is currently being updated 58. and is likely to be completed by 2019. E. Vikram, Forest Survey of India, correspondence with authors, February 2018. Strengthening Forest Fire Management in India   76 Box 2.5: Fire Danger Rating Systems in Canada and Indonesia Building on years of research starting from the 1920s, the Canadian Forest Fire Danger Rating System (CFFDRS) formally came into being in 1968, and its development continues even today. The CFFDRS helps with prevention by allowing fire managers to know where the risk of fires is higher. It helps with detection by giving fire managers a place and time to look for new fires. It also helps with suppression by providing some guidance about how the fire will behave. Beyond fire prevention, detection and suppression, it helps with planning, response, risk assessment, smoke modelling, and even assessing carbon emissions from these fires. The provinces and territories in Canada have been involved at each stage of development of the national FDRS, as well as other agencies such as the Department of Meteorology. The FDRS is modular, with different pieces for fire weather, behavior, prediction, and possible impacts. The major components of the CFFDRS are the Fire Weather Index (FWI), the Fire Behaviour Prediction system (FBP), and the Fire Occurrence Prediction system (FOP). The FWI is an accounting system for moisture that uses temperature, relative humidity, wind speed, and precipitation, each taken once a day at noon. For most jurisdictions around the world, the FWI is the most important component and is analogous for fire danger rating. The system is designed to derive the maximum amount of information from the least amount of data and is therefore easily adapted to regions outside Canada. However, before it can be used, it needs to be calibrated, which means analyzing FWI output and comparing it to actual fires and fire behavior to gain a proper appreciation of what the numbers really mean. If fuels data is added to the system, it is possible to predict fire behavior as well. Fire behavior prediction (FBP) is how the relative indices of the FWI are converted into real units such as rate of spread (how fast the fire can grow, in meters per minute), head fire intensity (how big the flames are, in kW/m), and fuel consumption (how much biomass is consumed by the fire, in kg/m2). In Canada, the CFFDRS is used to deploy firefighting resources in advance. It has helped provide for the safety and security of people, reducing deaths and injuries from wildfire. A high correspondence between areas of high fire danger and actual fire activity has been observed. Numerous countries have calibrated Canada’s FDRS for their own use, since it can be adapted to a variety of environments and is relatively simple to use. Indeed, after a season of especially devastating forest fires in 1997-1998, Indonesia embarked on the development of a national fire danger rating system. The system, based on the Canadian Fire Weather Index (FWI), was developed through a collaboration with the Canadian Forest Service. The FWI was adapted to local vegetation, climate, and fire conditions to identify periods of high ignition potential, dangerous fire behavior, and serious haze. Three components of the FWI were tailored for use in Indonesia, including the Fine Fuel Moisture Code (FFMC), the Initial Spread Index (ISI), and the Drought Code (DC). The FFMC serves as an indicator of ignition potential and was calibrated to local conditions through a historical analysis of satellite-based fire detections and weather conditions in Indonesia and field studies testing the moisture content and flammability of dead grasses. 77 Strengthening Forest Fire Management in India The ISI is used as an indicator for dangerous fire behavior. Fires in Indonesia’s tropical forests are typically of low intensity and easier to extinguish, while grassland fires can spread quickly and burn at too high an intensity to control. The ISI measures the potential rates of spread for grassland fires. The DC is a measure of the moisture content of deep soils and is used as a proxy measure for potential for serious haze events. Researchers discovered that when the DC crossed a certain threshold, there was a high probability of poor visibility at airports in the region because of the drying of deep peat layers. Peat fires are responsible for much of the smoke associated with regional air quality problems. Since the Indonesia FDRS became operational in the mid-2000s, it has been implemented by several agencies from the national to local scale. The Indonesian meteorological agency produces fire danger ratings nationally and for provinces using weather station data as well as 3- and 7-day forecasts of fire danger. To supplement the FDRS in areas where there are few ground stations, the Indonesian space agency produces danger ratings for the country using satellite data. The FDRS is also calculated locally at the district or sub-district level in some areas using weather data gathered from instruments at crew stations. Tailoring the FDRS for Indonesia involved a considerable investment of time and resources, but stakeholders in the Indonesian government have credited the FDRS with improving conditions on the ground, for example, increasing public awareness of fire danger and providing support for more informed decision-making by local governments, industries, and private individuals in using and responding to fires. Sources: Simpson (2017); Brian Simpson, Canadian Forest Service, “The Canadian Forest Fire Danger Rating System”, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017; de Groot et al. (2007); Guswanto et al. (2008) FIGURE 2.5: GRID-BASED PRE-WARNING ALERT SYSTEM IMPLEMENTED BY FSI STARTING IN 2017 Grid 5 km X 5 km Forest Cover Temperature Data Relative Humidity Forest Grid Forest Type Map –Vulnerable Forest types Drought Intersect Mask out 3 Rainfall Area 3 Fire Point Data to identify grids that are already depleted of fuels – High FRP Value – Knowledge base Decision on the basis of Forest cover density & Forest/ Admin Boundary – Selection of Pre-Warning Alert Grids Pre-warning email Alerts (KML format) Sources: FSI, “Forest Fire Pre-Warning by FSI,” presentation to World Bank, May 2017 Strengthening Forest Fire Management in India   78 To further improve its pre-warning alerts, FSI is are also provided on Bhuvan, an online platform exploring how to integrate social elements of fire for geospatial data and applications provided by the risk (e.g., population characteristics or areas of National Remote Sensing Centre (NSRC). regular burning by forest users). As Lin (2000) notes, FDRS have typically not considered human Up until 2017, systems for the detection of active factors influencing fire danger. This may represent a forest fires relied primarily on the satellite-based significant shortcoming of existing systems if they are observation of hotspots by the Moderate Resolution to be imported and used in India, where fire regimes Imaging Spectroradiometer (MODIS) instrument. are overwhelming dominated by human-caused MODIS is flown aboard the Terra and Aqua satellites. ignitions. FSI’s attempt to incorporate human factors Each of the two satellites passes over India about into its pre-warning alerts will require additional twice a day. MODIS can detect open fires burning research. A solid empirical basis for the FDRS can on the ground by sensing temperature anomalies at enhance its credibility with stakeholders and users. different bandwidths in the infrared range of the light Further research is also needed to determine how spectrum. Though MODIS has a spatial resolution of closely the current pre-warning alerts are correlated 1 km at nadir, the imager can detect flaming fires as with actual fire potential or behavior. small as 100 m2 depending on the intensity of the fire and the absence of any clouds, smoke, or tree cover 2.2.2 Detection that might obscure the view of the fire from space.64 Beginning in 2016, FSI and the states have also begun 2.2.2.1 Satellite-based detection systems to use observations of hotspots by the Visible Infrared Imaging Radiometer Suite (VIIRS), which is flown Over the past 10-15 years, remote sensing has become aboard the Suomi National Polar-Orbiting Partnership an indispensable part of forest fire detection in India. (Suomi-NPP) spacecraft. VIIRS offers a higher spatial FSI implemented its first nationwide system for resolution (375 m at nadir), making it better able to monitoring active forest fires using remote sensing sense lower intensity ground fires burning under and providing alerts to local forest departments in canopy cover. The downside is that VIIRS only makes 2004. In parallel with the FSI system, Madhya Pradesh about two passes over India every day. has established its own full-fledged alert system. Other states have created online platforms to re-distribute The ways in which satellite-based alerts are the FSI fire alert data and/or gather and report disseminated to field staff and the public have field verification data, such as Andhra Pradesh,59 progressed greatly since FSI first began sending active Chhattisgarh,60 Telangana,61 and Uttarakhand.62 fire alerts to the states via fax in 2004. In 2007-2008, Supplementing the FSI active fire alerts with additional Madhya Pradesh pioneered a new system to distribute data on fire locations such as slope, aspect, nearest alerts to field staff via SMS text alerts—its Fire Alert road, nearest village, and land cover type, the North Messaging System (FAMS) (box 2.6). Following Eastern Space Applications Centre (NESAC) provides Madhya Pradesh’s lead, FSI began sending text alerts twice-daily fire alerts to the forest departments of to registered users nationwide in 2010. FSI’s new Arunachal Pradesh, Assam, Manipur, Meghalaya, Forest Fire Alert System 2.0,65 launched in January Mizoram, Nagaland, and Tripura.63 Fire alerts for 2017, now provides SMS and email alerts for user- Himachal Pradesh, Karnataka, and Uttarakhand specified areas down to the beat level in 17 states, to 59. Andhra Pradesh Forest Department Geomatics Information System, “Forest Fire Status,” http://www.fgis.ap.gov.in/AP/FIRE/FIRE17/ Fire17.htm (accessed 1 October 2017). 60. Chhattisgarh Forest Department, http://www.fmisonline.org/fire.aspx (accessed 1 October 2017). 61. Telangana Forest Department Geomatics Information System, http://202.53.71.73/tgfdgis/TG/FIRE/FIRE17/Fire17.htm (accessed 1 October 2017). 62. Uttarakhand Forest Department, http://forest.uk.gov.in/contents/view/6/44/75-forest-fire-info 63. North Eastern Space Applications Centre (NESAC), “Current Forest Fire Alert,” North Eastern Regional Node for Disaster Risk Reductions (NER-DRR), http://www.nerdrr.gov.in/fire_alerts.html (accessed 1 October 2017). 64. In describing the development of the detection algorithm for active fires using MODIS imagery, Giglio et al. (2003) report, “Over all biomes considered, the size of the smallest flaming fire having at least a 50% chance of being detected under both daytime and nighttime conditions was ~100 m2” (279). 65. See FSI, “Forest Fire Alerts System 2.0,” http://117.239.115.44:81/smsalerts/index.php. 79 Strengthening Forest Fire Management in India Box 2.6: Satellite Fire Detection in Madhya Pradesh Since the mid-2000s, Madhya Pradesh has emerged as a leader in the use of satellite remote sensing and information technologies for forest fire detection in India. In 2007, it launched a Fire Alert Messaging System (FAMS) to begin sending text messages to field staff alerting them of active fires burning in their area, as detected by the satellite. Satellite-based fire detections were derived from observations by MODIS and provided initially via the University of Maryland and NASA’s Fire Information Resource Management System (FIRMS). The state forest department leveraged this freely available online data by building an automatic SMS-based system to disseminate information on fires burning in forested areas to field staff in near-real time. To achieve this, the state provided low-cost mobile phones to field-level officers and negotiated with the telecommunications carrier BSNL to provide texting services at a discounted rate. The department also negotiated for BSNL to erect new cell towers, expanding network coverage to more than 80 percent of forested areas. The FAMS provided immediate results. A year after the system was rolled out, the average time to extinguish fires fell from 11-12 hours to 2-4 hours, and the average area burnt per fire dropped from to 12.9 hectares to 7.1 hectares. Average burnt area has continued to drop to a low of 3.0 hectares.in 2016. The forest department has continued to make improvements to FAMS, developing an online platform and database for field staff to report back to headquarters on fire alerts in their area. All the alerts and field reports received by the department are provided to the public on the FAMS website, further enhancing the transparency and accountability introduced by the system. Sources: OneWorld (2010); Agnihotri (2017); and Madhya Pradesh, “Forest Fire Alert Messaging System,” http://www.mpforest.org/ intranet/fire2014/DashBoard.aspx. the range level in two states, and to the block level understand that “someone is watching”. A similar in one state.66 In other states, alerts are provided at sentiment was expressed by forest officers in other the district level. As of October 2017, FSI had 11,639 states with monitoring systems in place. Whether registered users, including forest department officers these systems have demonstrably reduced the number as well as public users. The number of registered users of ignitions and area of forest that is burnt each year by state is illustrated in figure 2.6. deserves further research. The benefits of satellite-based detection systems While the use of remote sensing for monitoring forest for forest fires in India have yet to be systematically fires has expanded quickly over the past 10-15 years, evaluated and quantified. Still, there is substantial its continued development faces several constraints: anecdotal evidence that the systems have helped, as evidenced from the Madhya Pradesh example in box • Incomplete digitization of forest boundaries: The 2.6, cutting down response times and the average area boundaries of lands under the management of the burned per fire. As one high-ranking forest officer state forest departments have been mapped and in Madhya Pradesh put it, FAMS has been effective digitized down to the lowest administrative level because fire users—and field-level personnel—now in only 10 states. In most states, the boundaries of These states receiving alerts down to the beat level as of February 2018 are: Andhra Pradesh, Bihar, Goa, Gujarat, Haryana, Himachal 66. Pradesh, Jharkhand, Karnataka, Maharashtra, Manipur, Mizoram, Punjab, Tamil Nadu, Telangana, Tripura, and Uttarakhand. States receiving alerts at the range level are: Chhattisgarh and Kerala. Alerts are provided at the block level in Meghalaya. Strengthening Forest Fire Management in India   80 forest department lands have yet to be completely nature, improving uptake may be more difficult digitized. FSI and the states use these digitized and may require, for example, working with mobile boundaries to screen satellite detections of service providers to install new cell towers or other landscape fires in issuing alerts to local field staff. infrastructure to expand network coverage in more In the absence of clearly defined boundaries, FSI remote forested areas. screens fire detections according to areas with forest cover in the most recent year it has surveyed. Some • Lack of ground verification, feedback, and lands under forest department jurisdiction are not evaluation: Satellite-based systems for fire technically forested (with at least 10 percent tree monitoring can be highly cost-effective. The data canopy cover by area) and thus are excluded from are free, computing costs are minimal, and the fire alerts. Forest departments are often expected alerts can be generated automatically. With many to respond to forest fires on non-department lands, states reporting shortages in resources, and field and fires started on these non-department lands personnel having to cover large territories, the can spread to reserve forests. Thus, having clearly satellite detection systems can help fill gaps in defined boundaries to notify the appropriate monitoring by crews on the ground. But there has field staff and providing alerts for forested areas been little investment in evaluating the accuracy of surrounding these boundaries can enhance the the satellite detection systems and how they can be utility of the alerts. improved. • Uneven adoption: In some states there more than A handful of states have built online platforms for 2,000 users registered with FSI’s Forest Fire Alert field-level personnel to submit ground verification System, while in others there are fewer than a dozen. reports for fire alerts in their areas. These states Some of these other states have put in place their include Andhra Pradesh, Chhattisgarh, Kerala, own robust systems for monitoring active fires and Madhya Pradesh, Meghalaya, and Telangana alerting field staff. Yet, in many other states without among others. In the vast majority of states, no such systems, use of the satellite-based alerts has such platform exists. Though FSI has established been minimal. The reasons for this limited uptake a way for users registered with its Forest Fire Alert are not completely understood, although there are System 2.0 to provide feedback, it has hardly some possible explanations. Lack of understanding received any so far. Even in those states where among forest guards as to how to use the system ground verification reports are collected, little has also contributes to its ineffective use. In poorer been done to systematically evaluate these data to or more remote forest areas, uptake by field see how the accuracy or utility of fire alerts may be personnel and local communities may be limited improved. by a lack of internet connectivity and mobile phone reception. Frequent turnover or rotation in field- As indicated in table 2.6, field verification rates level personnel (including seasonal fire watchers) have varied widely among these states, ranging also contributes to difficulties in providing alerts from 3 percent in Chhattisgarh to a self-reported to the right people at the right time, as databases 100 percent in Himachal Pradesh. The reported of users may quickly become out of date. Where accuracy of satellite-based alerts in these states has lack of understanding is the limiting factor, also varied widely, from more than 99 percent in encouraging greater adoption of satellite-based Andhra Pradesh to less than 25 percent in Kerala. monitoring will require continued outreach by FSI The lack of standard protocols for ground truthing and the state forest department, with trainings and by field staff in different states, however, makes it sensitization provided at the beginning of each fire difficult to interpret these numbers. In some cases, season. Especially in regions where large areas of field officers may report that an alert is false if they forest are contained within non-forest department observe that the fire is outside the department- lands, outreach will need to be extended beyond managed forest area and is on an adjacent land department staff to reach other stakeholders in managed by another entity. A quick glance at the JFMCs, VSSs, van panchayats, and other table 2.6 suggests that accuracy has been highest community institutions responsible for managing in the Southern and Central states with gentler forests and responding to fires. Where the barriers terrain and larger areas of open dry deciduous to improving adoption are more structural in forest, though this cannot be confirmed without 81 Strengthening Forest Fire Management in India TABLE 2.6: FIELD REPORTING ON SATELLITE-BASED FIRE ALERTS IN EIGHT STATES Andhra Chhattisgarh Himachal Kerala Madhya Meghalaya Telangana Uttarakhand Pradesh Pradesh Pradesh Field reporting rate for 89.4 3 100 70 60 90 satellite-based fire alerts (%) Rate of false alerts for 0.5 1 70 75.3 10 17 13 satellite-based detections (%) Fires reported by field staff 0 30 1.6 0 20 but not detected by satellite (%) Source [1] [2], [3] [4] [5] [6] [7] [8] [9] Notes and sources: [1] Data are for 2017, Andhra Pradesh Forest Department Geomatics Information System, http://www.fgis.ap.gov.in/AP/ FIRE/FIRE17/Fire17.htm (accessed 30 Sept 2017); [2] Feedback reported for 979 of 33,179 fire locations between February and June 2017, Chhattisgarh Forest Department, http://www.fmisonline.org/fire.aspx (accessed 30 September 2017); [3] Rate of false alerts as per Chhattisgarh Forest Department data sheet, sent to World Bank study team August 2017; [4] Himachal Pradesh Forest Department data sheet, sent to World Bank study team August 2017; [5] Kerala Forest Department data sheet, sent to World Bank study team August 2017; [6] Interviews with Dr. Atul Kumar Srivastava, Add’l PCCF, and Mr. Anurag Srivastava, Add’l PCCF (IT), Bhopal, Madhya Pradesh, 28 January 2017; Meghalaya Forest Department data sheet, sent to World Bank study team September 2017; [8] Average of reporting rate and false alert rate for 2016 and 2017, data from Telangana Forest Department Geomatics Information System, http://www.tgfgis.com/ (accessed 30 September 2017); [9] Uttarakhand Forest Department data sheet, sent to World Bank study team August 2017. more complete and reliable data from other states. ensuring greater accountability in responding to Without sufficient and representative ground fires. For example, in Madhya Pradesh, the forest verification data, it will be extremely difficult to department’s IT cell closely monitors which fire refine the out-of-the-box algorithms and methods locations are verified by field reports. If field staff that are currently used to generate fire alerts and to do not submit feedback for a location, then the IT make any modifications to improve their accuracy. cell will check if an active fire is observed in that Field-reported data should also include records of location upon the subsequent satellite overpass. fire incidents observed by ground crews that were Thus, satellite-based fire detection systems can not picked up by the satellite-based systems. be an effective administrative tool for the state forest departments. • Limited integration: Parallel alert management systems have been developed by FSI and the Integrating the national and state/regional states (using the FSI data). Yet, there is very little detection systems will require a clearer definition of integration or interfacing between these systems. the roles and responsibilities of the various agencies This lack of integration represents a missed and departments involved in fire monitoring. opportunity. Establishing a means for the regular FSI can play a crucial central role in providing exchange of ground verification data with FSI will technical support and advice to the states. Together be vital to refining the methods for generating alerts. with ISRO, NRSC, the Indian Institute of Remote Sensing (IIRS), and the regional space applications Improving the integration of alert systems does centers, FSI can also assist with the improvement of not necessarily mean replacing one with the other. existing methods and technologies for detection. There are important reasons for why state forest departments would choose to maintain their own There is also a need for greater integration fire alert management systems apart from the between on-the-ground fire monitoring and the nationwide FSI system. The state systems provide a satellite detection systems.67 Madhya Pradesh is direct line of communication between headquarters already moving in this direction by experimenting and field-level personnel and a mechanism for with a new mobile app that would allow field 67. This need to integrate “information generated from both field-level fire monitoring and reporting, and satellite-based fire monitoring” was identified in 2007 in the recommendations of a national workshop on “Rethinking forest Fire” organized by the Government of India and the Ashoka Trust for Research in Ecology and the Environment (Recommendations, Session II, 1). Strengthening Forest Fire Management in India   82 FIGURE 2.6: STATE-WISE NUMBER OF USERS REGISTERED WITH FSI FOREST FIRE ALERT SYSTEM 2.0 Maharashtra 4,437 Himachal Pradesh 3,444 Telangana 2,657 Chhattisgarh 2,302 Uttarakhand 1,942 Karnataka 1,772 Odisha 1,648 Andhra Pradesh 973 Kerala 872 Punjab 801 Tamil Nadu 650 Uttar Pradesh 209 Goa 165 Rajasthan 147 Madhya Pradesh 133 Mizoram 100 Haryana 80 Jharkhand 59 Andaman and Nicobar 56 Bihar 40 Gujarat 29 Manipur 29 West Bengal 15 Jammu and Kashmir 15 Delhi 12 Arunachal Pradesh 10 Chandigarh 9 Meghalaya 9 Assam 8 Sikkim 5 Tripura 5 Nagaland 5 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 5,000 Number of registered users Note: As of 8 April 2018; Source: FSI, http://webline.co.in/smsalerts/index.php. staff to send validation reports for fire alerts and including the inability to detect fires under to submit reports for fires that were not detected very cloudy or smoky conditions. In addition to by the satellite instruments. Such an app would this general limitation, the VIIRS and MODIS allow the forest department to track whether fires instruments have their own specific constraints are observed first by field staff or by satellite, the which must be weighted. VIIRS has a higher spatial location and time of ignitions or detections, the resolution than MODIS and is better able to detect time required for field crews to arrive on-site to lower-intensity surface fires. Still, it is difficult for verify alerts, and other valuable information that VIIRS to distinguish forest fires from crop burning can assist with fire management. The app may and other open fires on lands adjacent to forests, also improve estimates of burnt area, which is especially in fringe areas where forests are more currently estimated by ocular inspection only, by fragmented, and villages are scattered in and allowing field staff to take perimeter measurements around areas of forest. Also, because they only using GPS. pass over India a few times per day, MODIS and VIIRS may not be able to detect fires until many • Technological limitations: Satellite-based remote hours after they are ignited and grow large enough sensing of forest fires has progressed quickly but to be spotted during the next satellite overpass. faces some fundamental technological limitations, Geostationary satellites such as India’s INSAT-3D 83 Strengthening Forest Fire Management in India and INSAT-3DR imagers have a much shorter surveyed. All the surveyed officers in those areas said revisit time, but much coarser spatial resolution.68 fire watchers are provided wages in exchange for their services; however, about half noted delays or shortages How the MODIS and VIIRS data are obtained can in payments (30 of 66). Most respondents also said influence the lag time for alerts and thus the amount that fire watchers are not provided any training of time required before ground crews can respond to or equipment (38 of 65); all agreed that additional an alert and extinguish a figure. Reducing lag time training and equipment is needed. As one officer in can potentially reduce response time and prevent Kerala commented, “The fire watchers use crudest fires from growing to an uncontrollable size. Before form of firefighting. A simple training from the local FSI can generate alerts, the raw satellite data from firefighting office can improve their efficiency to a MODIS and VIIRS must be acquired by a site on the good extent.” ground, processed, and then distributed to users. FSI receives data directly from the National Remote Funding is variable but seems to be a constant challenge Sensing Centre (NRSC) in Shadnagar, Telangana. in most states, often leading to shortage of frontline staff The NRSC processes the MODIS and VIIRS satellite and hence a greater reliance on community members data to obtain the locations of fires using the standard for fire detection and response. As an example, one algorithms from NASA’s Direct Readout Portal. FSI officer in Jharkhand reported that only 6 of 117 forest then retrieves the data from NRSC via an ftp server. guard positions had been filled, and the number of Ideally, the NRSC can relay data and FSI can send firewatchers had also been reduced to 6, resulting in out alerts within 45 minutes of satellite overpass. In watchtowers remaining unstaffed and thus useless for practice, FSI has reported that delays of more than 8 fire detection. An unstaffed watchtower is pictured in hours between the time of satellite overpass and receipt figure 2.7. of data from NRSC have occurred in “many instances” (FSI 2017b). FSI has experienced frequent service As India continues to develop the use of remote disruptions due to server outages and other technical sensing for forest fire detection, MoEFCC and the difficulties on the NRSC side, mainly with the new state forest departments may consider testing new VIIRS data, that have prevented it from providing technologies to supplement existing systems. Examples timely alerts to the state forest departments.69 of technologies that have been deployed in other countries for fire monitoring include optical sensor 2.2.2.2 Ground-based detection of fires by field staff systems and wireless sensor networks (WSNs).70 Tower- mounted optical sensors may include video cameras On-the-ground monitoring of forest fires will capable of recognizing smoke from long distances, continue to be essential, even with the advances in thermal imagers that can sense heat rising from remote sensing technologies and alert systems. And flames, infrared spectroradiometers that can detect yet this function remains under-resourced. Of the 74 particulates in the air from biomass burning, and light field-level officers who were surveyed in the forest detection and ranging systems (LIDAR). Such systems departments of 11 states, 40 said that watchtowers tend to be more expensive and require significant or crew stations were maintained in their area. Of supporting infrastructure but may be appropriate for those 40, 31 said the watchtowers and crew stations in monitoring large tracts of forest in flat or gently hilly their area were all functional. Surveyed officers were terrain. Optical sensor systems have been used for fire unanimous in citing a need for additional watchtowers monitoring in many countries, including Belarus (box and crew stations in all states. 2.7), Canada, Chile, Kazakhstan, Mexico, Portugal, South Africa, Swaziland, the United States, and others. To assist with fire detection and response, the forest By comparison, WSNs have emerged more recently department hires seasonal fire watchers from the and are still in an earlier phase of development but local community in most areas where officers were have some potential advantages that are relevant for 68. The INSAT-3D and INSAT-3DR imagers produce an active fire data product for all of India every half hour. Active fires are derived using the MIR (T3, 3.8-4.0 µm) and TIR-1 (T5, 10.3-11.3 µm) channels. Spatial resolution is 4 km x 4 km at nadir. 69. Teleconference with FSI, World Bank, and NASA, 3 March 2017. 70. The discussion of optical sensor systems and WSNs draws heavily from Alkhatib (2014). Strengthening Forest Fire Management in India   84 FIGURE 2.7: UNSTAFFED WATCHTOWER Box 2.7: Detecting Forest Fires in Belarus The detection of forest fires in Belarus involves the use of fire observation towers, ground based detection through state forest guards, satellite- based monitoring as well as video surveillance. Aerial patrolling is also carried out by the State Aviation Emergency Rescue Institution, which is part of the Ministry of Emergency Situations of the Republic of Belarus. As in the case of India, remote sensing technologies have played a pivotal role in fire detection and prevention in Belarus. In 2015, Belarus implemented an innovative system for automated tracking and early forest fire detection using remote monitoring methods with video surveillance to detect smoke. While the existence of such a system is certainly of interest, this approach may not be appropriate in the context of a country like India, owing to the high risk of false alarms that may be generated (as a result of activities such as garbage burning). Therefore, the need for improving ground-based detection, with more crew stations and people on the ground, remains important. Sources: Aleksander A. Kulik and Dmitry Krasovsky, Ministry of Forestry, Republic of Belarus, “Forest Fire Protection in Source: Ross Smith, World Bank the Republic of Belarus”, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, India, including lower costs and reduced infrastructure November 2017 needs. WSNs consist of a distributed network of small sensors, which may detect changes in temperature, humidity, pressure, or other environmental conditions a WSN in Algeria for monitoring fire weather and such as the presence of smoke and communicate these calculating an index fire danger. Other examples of changes to other nodes in the network. Lutakamale WSNs for fire monitoring are provided by Alkhatib and Kaijage (2017), for example, have proposed a (2014). WSN for wildfire monitoring in Tanzania which can operate off the electrical grid, without the installation The use of geostationary satellite data can also help of watchtowers or other built infrastructure and which reduce the latency of fire detections. In other countries, can communicate with each other via SMS over the the GEOS and Himwari satellites have been used to cellular network, mitigating the need for internet provide near-real-time observations of active fires. In connectivity which has hitherto constrained the India, FSI has already experimented with using data application of WSNs in many developing countries. from the INSAT satellites; however, the results were Bouabdellah et al. (2013) have tested the design of “not encouraging.”71 Active fires are derived using 71. E. Vikram, FSI, review draft comments provided to World Bank team, February 2018. 85 Strengthening Forest Fire Management in India the MIR (T3, 3.8-4.0 µm) and TIR-1 (T5, 10.3-11.3 When fire behavior is modest, it is feasible for people µm) channels of the INSAT imagers. Because the to work right on the edge of the fire to create such a instruments have a coarse spatial resolution of around break and use hand tools to push or rake burning, 4 km at nadir, they are not able to detect small fires and some unburnt, material back into the fire, thereby that have just ignited or low-intensity surface fires. creating a fuel free break. When fire behaviour is of higher intensity and it is no longer feasible for people 2.2.3 Forest Fire Suppression72 to work at the edge of the fire, it is often possible to stand back a few metres from the active fire edge In general, forest fire suppression relies very heavily and create a mineral earth break about a metre or so on “dry” firefighting techniques because it is usually wide by raking or pushing the fuel off the proposed not possible to directly and accurately attack the fire line. When this tactic is employed, sometimes fire edge with water along the entire fire line73. The the active fire can be allowed to burn up to the fire location and nature of the terrain where the fire is line but at other times it is safer to burn back from burning may preclude the use of wheeled vehicles the newly created fire line and allow the fire to burn and ultimately if the fire cannot be surrounded by a back against the wind towards the main fire. At more trafficable track or road, the use of wheeled equipment extreme levels of fire behaviour, it may be necessary to is not a practicable option. retreat to a much greater distance, either to existing fire barriers such as roads or fallow fields, or to create Dry techniques include directly beating out the fire a substantial break using heavy machinery fitted with with hand tools to smother the flames (for very low blades (e.g. grader, crawler tractor). In such instances, intensity fires) or by separating the fuel in advance of it is almost never feasible to allow the wildfire to burn the active fire, either by natural breaks in the fuel or up to the break, and active lighting along the edge of by deliberately creating mineral earth breaks devoid these breaks (backburning or backfiring) is essential, of fuel. In many such instances, hand tools can play a else the main fire is likely to overrun the break as the significant role, as can heavy machines. fire approaches (see box 2.8). Box 2.8: Forest Fire Suppression Techniques A. Direct attack This is usually implemented against small fires burning at low intensity where it is feasible for firefighters to work right at the edge of the fire, pushing burning material back into the fire or smothering flames with suitable beaters. Generally, fuel on the edge of the fire is pushed back into to the fire, to minimize the risk of dragging lighted material onto unburnt fuel, while creating a narrow break, bared of vegetative matter. Because it is arduous work, firefighters need some protection from radiant heat such as long- sleeved and long-legged clothing and realistically can only operate under these conditions when flames are not much more than a metre or so tall. Because firefighters are working very close to the edge of the fire, it is a relatively safe operation. In the event of a sudden change in conditions, (e.g. increase in wind velocity) that may elevate fire intensity, firefighters can rapidly move into already burnt areas. Direct attack with a variety of hand tools is feasible up to fire line intensities of about 800 kW/m, so its use is generally limited to cooler and milder conditions. The two major determinants as to whether direct attack can be utilized are the impact of heat and smoke on the firefighters. Heat uptake can be regulated to some extent by appropriate clothing but smoke exposure is a different issue and the solution is to keep firefighters out of heavy concentrations of smoke. 72. This section draws primarily from a background note prepared by Ross Smith, World Bank consultant, June 2017. 73. The extensive use of water tankers and aircraft to make direct attack on the fire front constitutes “wet firefighting”. It is rarely possible to accurately position sufficient water on land and forest fires to effectively control them - this is one of the reasons why aerial firefighting is inordinately expensive. Strengthening Forest Fire Management in India   86 Strengthening 87 Forest Fire Management in India   87 Strengthening Forest Fire Management in India Strengthening 88 Forest Fire Management in India   88 Strengthening Forest Fire Management in India Advantages of this method are that use can be made of the existing sections of the fire boundary where the fire has self-extinguished (e.g lack of fuel, presence of large areas of exposed rock, green moist gullies and creek side areas). Because there is very little “burning out” necessary, total area burnt is minimized and very little additional fire is added to the environment. A concurrent disadvantage is that the control line follows the fire edge and it may be tortuous and longer, increasing the difficulty of patrol. Large items of heavy fuel (e.g downed logs and dead trees) right on the fire edge need to be accounted for, perhaps by extending the fire line to safely include them within the fire boundary. Alternatively, larger pieces of woody fuel may need to be cut into smaller pieces and moved well inside the fire boundary or safely outside of it. B. Parallel attack When conditions are too intense for firefighters to work right at the fire edge – if it is too hot, the flames are too high, and/or it is too smoky – firefighters can withdraw a short distance from the active fire edge and create a “fire-line” by baring the forest floor down to mineral earth and backburning from that line. Burning out follows closely behind fire-line preparation and there is always a need to watch out for spot overs as the back fire and the main fire come together, creating a junction zone with temporarily increased intensity. Advantages of this method are that the fire line can be much more uniform in direction, there is the opportunity to bypass heavy fuels such as downed logs and trees (or to clean around them and remove the fuels in close proximity to the intended fire edge to reduce the chances of their ignition). Advantage can also be taken of any significant lengths of fire line where the fire is extinguished and other natural barriers. Operating conditions for firefighters are less arduous by way of reduced heat radiation and smoke. Disadvantages are that more fire is applied – there will be junction zone effects as the backfire and wildfire meet so there is a need for increased patrol and vigilance when this occurs. C. Indirect attack It is used against fires that are too intense for close in suppression action, against fires that may be causing downwind spot fires and fires that are spreading too quickly to allow closer suppression. This method usually involves withdrawing to previously prepared lines such as roads or major mineral earth fire breaks. Fresh breaks can be constructed using major plant items to ensure there is a trafficable road/trail as a boundary line from which backburning operations can proceed and from which patrol activities can subsequently be undertaken. Often, these control lines may be some kilometers downwind of the current fire location. The rationale for such a large distance is: • It is futile to attempt close in suppression against very intense fires – it will certainly fail. • It is very dangerous to position firefighters downwind in reasonably close proximity to high intensity fires. • Some amount of time is often necessary to clean existing fire breaks or construct new breaks and prepare for burning out between those lines and the wildfire. 89 Strengthening Forest Fire Management in India FIGURE 2.8: PRINCIPAL TECHNIQUES USED TO SUPPRESS FOREST FIRES (FREQUENCY COUNT OF RESPONSES) Manual beating or smothering 73 Use of fire in control 35 Creation of fire lines 32 Surface clearing 17 Dousing 13 0 10 20 30 40 50 60 70 80 Frequency (number of times mentioned by respondents) Note: responding officers = 85; officers may indicate more than one suppression technique. Source: World Bank survey of state forest department officers, April-August 2017 The main techniques used for suppressing forest cannot be met, then it is irrelevant how much or fires, as reported by forest officers, are illustrated what capabilities exist in the way of hi-tech specialist in figure 2.8. Ground crews manually beating or equipment and the ability to collect data about fire smothering fires was the most common method cited occurrence and behavior either on site or by remote by officers. Most said that beating is done by hand means. There is little point in possessing those with bushes, tree branches, or self-fashioned brooms capabilities unless there is a capability at the very local (jhapas); only a minority said that ground crews level to make a practical and useful response and used manufactured tools. According to respondents, implement effective actions to restrict the fire/fires. beating is typically done by department staff and locally-hired fire watchers. Other members of the Areas that are suitable for use of hand tools include local community, including Joint Forest Management cleaning along the proposed fire line itself, clearing Committees (JFMCs), may also be involved in fire fuel from around large flammable trees (dead snags response and suppression. The use of fire in control that may ignite and cause spotfires across the control (e.g., setting back fires or counterfires) is done by line) or large downed fuels such as old logs or piles department staff. of woody debris from road construction. The normal modus operandi is to carefully assess the fire line Irrespective of the technique used, “people on the and identify any such areas within 20-30 metres of ground” are key to effective fire suppression. In spite the fire line and either remove surface fuel from of the availability of hi-tech equipment globally, the the close proximity of these targets, to minimize the principal need is always to have a competent, trained risk of igniting them during subsequent burning and equipped workforce on the ground ready to out operations or adjust the fire line, if possible, to respond and take immediate action. If that target exclude them from the burn area. Strengthening Forest Fire Management in India   90 A second critical area requiring hand tools is the and litter must be dragged off the earthen break. In backfiring or burning out operation itself. The tried recent years, small motorized equipment has been and tested method is to light along the fire side edge added to the armory, including leaf blowers, such as of the control line and allow the backfire to progress those used in Odisha (box 2.9). back toward the wildfire. Depending on burning conditions, the timing of this operation is critical. There Hand tools need to be robust but also need to be must be sufficient time for the back fire to penetrate of modest weight as their continued use can be the unburnt area between the wildfire and the control very tiring for firefighters. Typically, the handled line to mitigate the junction zone74 effects that will implements such as rakes and hoes should have undoubtedly occur. In large burnout areas where there wooden (hardwood) or synthetic handles for weight can be as much as several hundred meters to several and balance purposes as well as for heat transfer. kilometers between the proposed control line and the Metal handles render tools too heavy and may fire, burning out can be speeded up by use of aerial interfere with tool balance so they are uncomfortable lighting techniques but the ignition of fuels along the to use and can become hot. Some manufacturers fire line remains very much a manual task. promote multi section synthetic handles that enable quick conversion between short and long handled At the same time, the need for equipment to manage tools and also allow the use of interchangeable forest fires was emphasized by forest officials heads - but they usually come at a price which can be responding to the survey (see figure 2.9). Only very high. a handful (5) of officers agreed that firefighting equipment is adequate and sufficiently available in Many tools serve a dual purpose with different their area. When asked what additional fire-fighting configurations on a single head, effectively being equipment is needed, about half (48) pointed to two tools in one. Often, one side serves a cutting the need for basic safety equipment and clothing, or digging purpose and the opposite side offers a including mainly fire-resistant uniforms, boots, raking function. The important features are robust helmets, and gloves. Respondents said that basic construction whilst maintaining a reasonable weight, safety equipment and clothing was also needed for fire comfortable handles and good balance. Joining of watchers. Other common equipment needs identified metal head pieces must be achieved by full length fillet by officers included more manual hand tools such welds to achieve maximum strength of metal joins – as beaters and metal rakes with wooden handles, simple spot tack welds will quickly break apart due to mechanical tools (especially leaf blowers for clearing the heavy nature of tool use. Additional hand tools are ground litter), and transport vehicles for field staff. discussed in Annex 5. Respondents also voiced the need for greater support for field operations. As one officer in Himachal Of the handtools discussed in Annex 5, little evidence Pradesh remarked, “Besides the actual fire operations of any use of them has been observed during field needs are much more difficult, the personnel need a visits apart from the use of leaf blowers and rakes. back-up continuous support of logistics and food and The team was only able to view specific equipment water supply for which there is no set up.” for fire suppression in Odisha and Uttarakhand. In Odisha, extensive use is made of leaf blowers but that There is a range of hand tools that are useful for usage is relatively unknown in other states. There is a the above activities. Typically, hand tools are cutting, good opportunity to acquire a range of handtools and hoeing or raking tools. To clear fuel from fire lines, supply them to areas where community members take small vegetation needs to be chopped off or dug out an active role in response to unwanted fires and/or Junction zone - occurs when two fires approach each other. A point will be reached when each fire begins to influence the other. When 74. that occurs, fire behavior changes rapidly, flames become taller and rate of spread increases. Fire intensity can dramatically increase and potential for spot fires to be caused from the junction zone can escalate significantly. 91 Strengthening Forest Fire Management in India FIGURE 2.9: ADDITIONAL EQUIPMENT NEEDS MENTIONED BY OFFICERS Rake 6 Beaters 6 Other manual hand tools 2 Cutting tools 2 Axe 1 Spade 1 Leaf blower 21 Lightweight fire extinguishers 15 Water tanker (truck or tractor) 4 Portable sprayers 4 Other dousing equipment 3 Bucket 1 Fire-resistant uniform 30 Boots 21 Helmet 19 Other equipment and clothing 15 Gloves 6 Protective eyewear 4 Torch/flashlight 4 First aid kit 2 Manual hand tools Dust mask 1 Transport vehicles 18 Mechanized hand tools Fire engine 2 Dousing equipment Helicopter or other aircraft 2 Tractor 1 Safety equipment and clothing Communications devices or GPS 5 Vehicles Food and water 2 Communications / GPS Nothing 6 Additional manpower more useful than equipment 1 Food and water Light equipment that is easy to carry 2 Other Simple firefighting tools in sufficient quantity 1 Drones 1 0 5 10 15 20 25 30 35 Frequency (number of times mentioned by responding offices) Note: responding officers = 74; officers may indicate more than one type of equipment needed. Source: World Bank survey of state forest department officers, April-August 2017 participate in early burning activities to restrict such 2.2.4.1 Post-fire data collection and the assessment fires to their intended area. of fire impacts 2.2.4 Post-fire management Post-fire data collection is an essential part of the fire management process and crucial to producing The FFPM process continues after fires are informed FFPM plans and policies. However, this extinguished with two main activities: (1) post-fire data part of the management process is given little priority collection and the assessment of forest fire impacts; and is often performed solely for the sake of fulfilling and (2) restoration and rehabilitation. administrative requirements. There is a need to reorient post-fire data collection and analysis toward Strengthening Forest Fire Management in India   92 Box 2.9: The Use of Leaf Blowers in Odisha, India A range of equipment for fighting forest fires is currently available in Odisha, including fire swatters, bill-hooks/axes, torch-light, water bottles, fire suits, masks, boots, helmets and gloves. Odisha has also pioneered the use of leaf blowers, of a two-stroke engine backpack blower with an overall weight of 9.8 kilograms, for combatting forest fires. FIGURE B2.2: LEAF BLOWERS Backpack model leaf blower Fire line cleared of flammable litter This unit was designed as a blower for heavy duty landscaping purposes and is capable of blowing leaves and other fine fuel from intended fire lines. It admirably fills the role of removing dry fuel, comprising leaves and other small pieces of litter from proposed fire lines. The powerful and sustained blast of air can shift all detached litter from treated areas. Although it was originally acquired for clearing mineral earth lines to set counter fires against unwanted fires, this type of unit has proven to be useful in direct attack on fires by blowing litter into the fire, simultaneously creating a bare mineral earth break. It is likely that one or perhaps two units operated in tandem would be effective in clearing fuel from designated fire lines, eliminating the need for heaping and burning. The modus operandi is to walk through the forest and blow material away from proposed fire control lines. Any plants attached to the soil will not be disturbed and it may be necessary for a follow up operator, equipped with a rake or slasher to walk behind the blower operator to remove any vegetation that could provide a conduit for fire to cross the fire line. It is important to note that while this equipment is very useful in lowland broadleaf forests on easy terrain, its utility under Chir Pine is not known. It may be that near surface grasses, herbs and small shrubs will sufficiently bind the Chir Pine needles to render them too difficult to move enough to make a clear fire line. Likewise, the steeper topography may preclude the use of blowers. Sources: World Bank Field Visit to Odisha, May 2017; T. A. K. Sinha, Forest Department, Government of Odisha, India, “Forest Fire Protection in Odisha”, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017 93 Strengthening Forest Fire Management in India the goal of strengthening prevention. The need for with the police department, and the incident will be such a reorienting was recognized as long ago as 1976, investigated as a criminal matter. when the National Commission on Agriculture (NCA), commenting on the lack of complete and accurate forest Post-fire field reporting is hindered by insufficient field fire statistics, “It is important to ascertain and maintain staff, difficult terrain, and a lack of communications an authentic record, as far as possible, of the causes infrastructure in more remote areas. Underreporting of forest fires, with a view to planning fire prevention of forest fires by field staff may also occur because of measures” (NCA 1976: 45.2.2, emphasis added). More institutional disincentives. As M.K. Sharma, Additional than 40 years later, the importance of data collection Inspector of Forests, wrote in 2001, “It is generally and analysis for informing prevention still holds true. observed that field staff do not report the actual fire damage due to fear of action and this practice needs Post-fire data collection includes the gathering of to be curbed” (GoI No. 9-6/99-FFD, 22 June 2001). information on fire incidents via field reporting as well Losses above a certain amount must be reported to as the use of remote sensing. According to information the Accountant General’s Office, which may affect gathered through field surveys, field reporting is career prospects and result in monetary loss to typically done at the lowest level, by forest guards. The officials. Thus, the fear of punitive action may lead to requirements for fire reports by field staff are mostly fewer reports being filed and underestimation of the consistent across the states and include: actual area affected by fires. Even with the advent of satellite monitoring of fires, because of how incentives • Location of fire (including administrative unit, are aligned, field-level officers may be more inclined nearest village, and GPS coordinates, if available) to report back that alerts in their area are false. • Time of fire occurrence • Name of reporting officer Reflecting on the dilemma of underreporting, the • Cause of fire National Forest Commission recommended, “Since • Person(s) responsible for igniting the fire, if fire cases are underreported, in terms of number identified of occurrences, the qualitative damage and the area • Witnesses, if any affected, by the field functionaries, a mechanism should • Extent of area affected be developed for higher authorities to crosscheck these • Type of forest affected (natural forest or plantation, reports” (NFC 2006: para 53). To some extent, such tree species affected) an institutional mechanism does exist. Range officers • Damages to forest caused by fire and DFOs are required to submit inspection reports • Damages to property, injury, or loss of life for fires after initial incident reports are filed by forest • Actions taken to extinguish the fire guards, and fire reports are sent up to the CF and PCCF for review. Yet, these authorities all face similar Once this information is collected, field reports are then disincentives to the full reporting of fires, and merely sent up to the range officer (RO), who compiles a daily strengthening oversight within the department may summary of fire incidents to send to the divisional forest not be effective unless the fundamental problem of officer (DFO). From the DFO, reporting continues up incentives is addressed. the chain of command to the conservator of forests (CF) and eventually the principal chief conservator of Possible remedies to the problem of poorly aligned forests (PCCF) in charge of forest protection or fire. incentives for accurate and full reporting include: Respondents in Uttarakhand noted that reports are also sent to the district magistrate and that the forest • Delinking department financing and career department headquarters provides daily and weekly prospects from fire damages. Exempting forest updates to the state government during peak fire fires from the reporting requirements to the season. If a fire reaches a large enough size (e.g., more Accountant Generals Office was one of the key than 2 hectares), causes damage to property, or results recommendations to be issued by MoEFCC after in injury or loss of life, a first information report (FIR) the National Workshop on Forest Fires in 2007, or preliminary offense report (POR) may also be filed though has yet to be implemented; Strengthening Forest Fire Management in India   94 • Holding field officers accountable for the fulfillment fire were unknown in more than 50 percent of cases of required prevention and control activities but not they encountered in their respective areas. punishing them for the occurrence and reporting of fire. Unless damaging or unwanted fires are More useful information on the causes of fire could caused by negligence or poor management on the be gathered for planning and management purposes part of field officers, the reported incidence of fire if field officers could report the probable or suspected within an officer’s jurisdiction should not be tied cause of fire using a general classification scheme. The to the determination of job performance, monetary need for “a uniform classification of forest fires by compensation, or career advancement. Fires are a types and causes…evolved and adopted by the States” semi-natural occurrence and not completely within was also recognized long ago by the NCA (NCA 1976: the control of field personnel. Also, the complete 45.2.3). To this end, in 2007, the MoEF (now MoEFCC) exclusion of fires from forests is not the aim of issued guidelines proposing a basic categorization the department. Field-level personnel should be of fire causes as part of a revised proforma for field rewarded for providing accurate and thorough reporting. The guidelines list four broad categories of data on fires, not punished. Reporting should be causes: (1) graziers; (2) escape from agricultural fields; reframed as an important management activity, not (3) accidental; or (4) other. Since then, some states have simply an administrative requirement. created more detailed classification schemes. Kerala, for example, has maintained statistics on the number The underreporting of fires is one issue; a separate of forest fire incidents by ignition source listed in table issue is the limited investigation of the causes of fire. 2.7 since 2011. The share of unknown or unascertained Limited resources are a major constraint, especially causes is relatively low and has improved greatly since during the peak fire season. As one officer surveyed 2011, aided by the new classification scheme. in Uttarakhand explained, “Causes of forest fires are not investigated in detail because during the Building on the experience of Kerala and other fire season the forest staff is occupied completely states, the creation of a common classification scheme in fire suppression and does not have time to spare for reporting the causes of fire across states could for investigation work”. Only in a few cases are such facilitate the aggregation of forest fire statistics at investigations “fruitful”. Consequently, about one- the national level. Many countries and regions have quarter of survey respondents (23 of 88) said that developed such schemes, including Australia, Canada, the causes of forest fires are usually investigated only the EU, New Zealand, Russia, and the United States. partly or not at all. In some cases, investigations may A common classification scheme for India would need be delayed until after the fire season is over and more to recognize the variety of circumstances and uses resources are available. In cases where an FIR or POR of fire in the different regions of the country, which is filed, field staff may defer to the police to investigate. may be much different than in countries with already- Little guidance or training is provided to field staff on established schemes. Importing a classification scheme methods for investigation. directly from these other countries and regions will not be workable. As a starting point for discussion, a Limited collection of data on the causes of fire may possible categorization is presented in Annex 4. also owe to the legal nature of such investigations. Because all human-ignited fires in state-managed Apart from increased reporting on the causes of fire, forests are treated as an offence under the Indian post-fire data collection could also be improved by Forest Act of 1927, the causes of forest fires are a more complete and accurate reporting of the area typically investigated by field-level personnel with affected by fire. Forest department officers who were the aim of determining responsibility for the offence. surveyed said that field reporting of area affected is The legal determination of responsibility requires usually done solely by visual inspection and making a a high threshold of certainty that is not easily met. rough estimate, not by using GPS or remote sensing. Also, as noted, in cases of large or damaging fires, Also, field reports of fire-affected area usually exclude responsibility for fully investigating the cause of fire any areas burnt as part of fire prevention or control is often handed over to the police. Consequently, half operations. Thus, there is a large possibility for error. of the responding officers surveyed said the causes of Significant discrepancies exist between self-reported 95 Strengthening Forest Fire Management in India TABLE 2.7: FOREST FIRES REPORTED BY IGNITION SOURCE, KERALA (NUMBER OF INCIDENTS) Cause 2011 2012 2013 2014 2015 2016 Accidental 37 391 112 103 83 96 Incendiary 1 2 1 4 2 4 Deliberate 0 2 3 0 1 1 Lightning 0 2 16 0 3 2 MFP collection 1 14 3 2 5 1 Natural 9 117 71 30 53 121 Not ascertained 49 320 91 79 30 67 Power line 0 2 1 6 4 2 Settlements 1 2 3 3 1 1 Travelers/truckers 8 66 42 11 14 27 Fringe dwellers 1 4 0 0 2 1 Forest offenders 13 95 92 98 74 162 Graziers 0 0 0 0 2 1 Unknown 340 0 69 189 65 0 Sources: data sheet provided by Kerala Forest Department to the World Bank team estimates of burnt area and remote sensing-based imagery of burn scars are provided by Reddy et al. estimates, as in table 2.8 for Chhattisgarh, Kerala, and (2017b). FSI has also done a nationwide assessment Uttarakhand, for example. of burnt area in forests using AWiFS, for 2015 and 2016. FSI plans on continuing to perform nationwide TABLE 2.8: REMOTE SENSING-BASED assessments each year toward the end of the fire VERSUS FIELD-REPORTED season, before the arrival of the monsoon rains. For ESTIMATES OF BURNT the validation purposes, FSI has conducted its own FOREST AREA IN 2014 (KM2) extensive ground-truthing in Uttarakhand, visiting 339 recorded burnt areas in May 2016, and has State Reddy et al. Self-reported done additional comparisons against Landsat-8 OLI (2017b) and MODIS imagery. From these comparisons, FSI Chhattisgarh 4,924 15 estimates overall user accuracy of the AWiFS-derived burn scars at around 85 percent, above the threshold Kerala 113 17 for what is typically considered good.75 Uttarakhand 57 7 An important part of assessing the impacts of fire is a Note: Reddy et al. (2017b) estimates are for February-May using satellite imagery from AWiFS valuation of economic losses. As noted, field staff are required to provide information on monetary losses as Reliable information on burnt area is invaluable to the part of the reporting requirements for fire incidents in assessment of forest fire impacts. Due to incomplete most states. Methods for valuation are stipulated by the field reporting, at the national level, remote sensing state forest departments and described in the working is currently the best option for the assessment of fire- plans. Though not all states have issued a standard affected area, including areas subject to controlled methodology for the accounting of losses,76 in the burning as well as unwanted fire. Nationwide estimates states that have, the most common approach is to value of burnt area for 2014 using higher-resolution AWiFS damages in terms of timber losses or replanting costs 75. FSI, “Forest Fire Burnt Area Assessment,” presentation to World Bank, 14 December 2016. 76. Eleven states were asked about their prescribed methodologies for assessing the monetary value of damages caused by forest fires. Among these states, the team received responses from 6 states: Chhattisgarh, Himachal Pradesh, Kerala, Meghalaya, Telangana, and Uttarakhand. Among the state forest department surveyed, the Kerala department noted that it has not prescribed a standard methodology. Yet, the state does publish annual statistics on losses due to forest fires (see Kerala 2016). In Meghalaya, where most forested lands are outside the management of the forest department, no assessment is made of damages to non-department forests from fires. Strengthening Forest Fire Management in India   96 TABLE 2.9: SCHEDULE OF RATES FOR THE CALCULATION OF DAMAGES FROM FOREST FIRE IN UTTARAKHAND Item Rate before 2016 (INR) Revised rate starting 2016 (INR) A. Plantation 1st year 10.00 per plant 15.00 per plant 2 year nd 11.20 per plant 16.80 per plant 3 year rd 12.48 per plant 18.72 per plant 4 year th 14.00 per plant 21.00 per plant 5 year th 16.00 per plant 24.00 per plant B. Natural forest (Surface fire) Chir forest 1,500.00 per ha 2,250.00 per ha Sal forest 1,000.00 per ha 1,500.00 per ha Mixed forest 500.00 per ha 750.00 per ha C. Natural forest (Crown fire) Chir forest 600.00 per ha 900.00 per ha Sal forest 332.00 per ha 498.00 per ha Mixed forest 168.00 per ha 252.00 per ha Source: CCF Working Plan, Uttarakhand, Letter No. 598/27-2, 16 May 2016 for plantation areas and natural forest. An example of The first kind of assessment is what is currently being this approach is provided in table 2.9, which details the done at the field level in booking forest offences. schedule of rates used in Uttarakhand to calculate losses. The second and third kind of assessment are new but would provide a more complete picture of the Himachal Pradesh has instituted processes for the most economic impacts of forest fires and would assist in inclusive accounting of losses among the states that were the FFPM process. assessed. As an illustrative example, guidance from the working plan of Rampur Division on valuing losses to Disaster impact assessment would be performed only conifer forests due to fire is given in table 2.10 below.77 for particularly large or destructive fires. The need to conduct such an assessment would be triggered The assessment of the economic costs due to forest by a formal disaster declaration. The National Forest fires should be tailored to the appropriate objective Commission has suggested, for example, that all fires and scale. Economic valuation may serve different larger than 20 km2 in size in forests and grasslands purposes, and each of these purposes has distinct should be declared a state disaster (NFC 2006, sec features in terms of the data and analysis involved. 8.10, para 52). Disaster impact assessments may also Relevant purposes for India include: be appropriate if a fire affects an area of particular value or significance for biodiversity conservation or • The determination of compensation amounts for cultural heritage or if there is a significant loss of life legal or administrative purposes; or property. Depending on the nature of the disaster • Disaster impact assessment for particularly large or declaration (national or state), such an assessment destructive fires; would likely not be conducted by the local forest • Strategic-level evaluation of FFPM programs by department but rather by MoEFCC or an independent the states or MoEFCC. entity appointed by the state, such as a university, consulting firm, or panel of experts. 77. The methodology used by Rampur Division for the valuation of losses was developed by the CCF, Forest Protection, Bilaspur. 97 Strengthening Forest Fire Management in India TABLE 2.10: GUIDELINES FOR ASSESSING THE COSTS OF FOREST FIRES IN RAMPUR DIVISION, HIMACHAL PRADESH Type of loss Loss calculation (INR) Remarks Plantation areas Area burnt in ha x plants/ha x% survival (1st year) rate x cost/plant Plantation areas Total expenditure on plantation up to (2nd – 10th year) date, including maintenance x plants burnt x (1 + .10 x years beyond the 1st year after planting) Area under natural For fully stocked areas, assessed as for Loss of natural regenerated seedlings regeneration (seedling plantations due to fire loss) Area consisting of pole- 50% value of dried or salvage trees Loss is assessed after the rainy season; stage, middle-age, or for salvage timber, royalty is taken as mature crop 50% total market value of the tree Resin blazes As per royalty rate fixed by the state government Fodder and grasses None Fodder is a renewable resource, so loss is not calculated; grasses are usually cut by right holders before the fire season, so losses are usually not sustained; little fodder/grass production in chir pine forests Beneficial herbs, shrubs, Loss of harvest equal to area burnt in wild fruits ha x average annual yield per ha x unit price (according to permit issued in the past or local market prices) Domestic losses Not considered Domestic losses include physical property (structures, livestock, etc.) as well as injuries to people or loss of life; these losses are calculated separately by the revenue department Fire prevention costs Expenditures for controlled burning, Cost is based on average cost in clearance of fire lines, etc. previous years; focus of prevention activities is typically pine forest Other environmental and Not considered Difficult to assess due to lack of wildlife losses requisite methodology Sources: adapted from Rampur Division Working Plan for 2013-14 to 2028-29, Himachal Pradesh, by Mr. B.L. Negi Strengthening Forest Fire Management in India   98 Box 2.10: The Economic Costs of Forest Fires in Indonesia Between June and October 2015, fires across Indonesia burned an estimated 2.6 million hectares across the country, an area equal in size to the entire state of Tamil Nadu. Fires have long been used as a cheap tool to prepare fields and gain access to lands for palm oil cultivation and other activities, but in the absence of controlled burning measures or sufficient law enforcement, the fires can easily grow out of control. An analysis by the World Bank (2016) estimated the fires of 2015 cost the Indonesia economy US$ 16.1 billion, equivalent to about 1.9 percent of GDP. Losses were more than twice the reconstruction cost following the Aceh tsunami, and more than 1.5 times the value added from the country’s entire palm oil production in 2014. The World Bank team assessed the costs of the Indonesian fires by applying the methodology for disaster assessment developed by the UN Economic Commission for Latin America and the Caribbean (ECLAC). Under this methodology, damages are estimated as the amount of financing needed for reconstruction and rehabilitation, while losses represent the reduction in economic activities and income resulting from the disaster. A wide range of damages and losses were considered, including those categories shown in the table below. The analysis also drew on previous research by the Center for International Forestry Research (CIFOR) to show that 85 percent of the cashflow generated by using fire to convert forests and peatlands to oil palm went to local elites and plantation developers. A boon to some, the fires were a disaster for many more. In response to the fires and their devastating cost to the economy, on October 23, 2015, the Indonesian president called for a moratorium on new peatland concessions and a cancellation of existing concessions that have not been developed, thereby halting the legal conversion of peatland and peat swamp forests into agricultural land. TABLE B2.1: DAMAGES FROM 2015 FOREST FIRES IN INDONESIA Category of damage/loss US$ millions Agriculture 4,839 Estate crops 3,112 Food crops 1,727 Environment 4,253 Biodiversity loss 287 Carbon emissions 3,966 Forestry 3,931 Manufacturing and mining 610 Trade 1,333 Transportation 372 Tourism 399 Health 151 Education 39 Firefighting costs 197 Total 16,124 Sources: excerpted and adapted from World Bank (2016) 99 Strengthening Forest Fire Management in India Much of the existing literature on the economic costs different framework for valuation is needed for a of forest fires is focused on assessing the costs of large strategic-level assessment than for a disaster impact forest fires. A recent example of economic valuation assessment. Such a framework is provided by the by the World Bank of the costs of fires in Indonesia cost-plus-net-value-change (C+NVC) model, which is provided in box 2.10. A useful methodological has emerged as a standard for determining the “most framework for conducting such an analysis has efficient level of fire management” by weighing the been outlined by the UN Economic Commission for costs and benefits of fires and FFPM programs (see Latin American and the Caribbean (UN ECLAC). Rideout and Omi 1990, Donovan and Rideout 2003). The ECLAC framework involves a damage-plus-loss Under the C+NVC model, costs, C, include spending approach. Damages represent the destruction of on fire prevention, suppression, and protection. The physical assets and are measured as changes in the net value change, NVC, is equal to the sum of damages stocks of assets or other goods relative to a baseline or and losses minus any potential benefits of fire. The pre-disaster situation. Losses are a flow-based concept aim of the analysis is to determine the optimal level of and represent changes in economic activity following prevention, suppression, and protection to minimize a fire disturbance. Indirect losses may include flow- the sum of C+NVC. The C+NVC model first emerged on effects to other sectors, for example, due to delays in the United States after Congress began asking the in the supply of raw materials from fire-affected areas Forest Service to justify its ever-increasing budget to manufacturers. Losses may also include additional requests for fire control in the 1970s and 1980s (see outlays for goods and services such as evacuation Lundgren 1999). costs, overtime pay for health workers, treating higher caseloads of patients with respiratory illnesses The calculation of NVC would consider a range of during large wildfires, and the re-composition of environmental services provided by forests, including public spending as resources are diverted from carbon storage (box 2.11). Forest fires contribute to one sector to another through the provision of climate change by releasing carbon stored in trees, emergency assistance funds to fire-affected areas. The undergrowth, litter, and soils into the atmosphere. quantification of damages and losses should be limited Forest fires also emit heat-trapping gases such as N2O to a well-defined area or region (UN ECLAC 2014). and other aerosols that influence the regional and Distributional impacts within the affected area should global climate. The net effect of a fire on the climate also be considered. Some groups may be less resilient depends on the pre-disturbance characteristics of the than others in coping with the economic shocks from forest and the extent to which the forest is able to fires, particularly those in poor rural areas (Abt et al. regenerate. Forest clearing and persistent changes in 2008). vegetation composition and structure after a fire may result in net emissions (IPCC 2014; Sommers et al. At the strategic level, economic valuation should be 2014). performed to evaluate FFPM programs by the states and MoEFCC. This is the third kind of assessment. The Once estimated, GHG emissions may be valued using objective of such an evaluation might be to determine an indicative carbon price for payments to the forestry the efficient level of budgetary allocation for FFPM, sector. The Comptroller and Auditor General of India to support a request for additional financial resources has previously suggested an illustrative value of US$ if a shortfall is found, and to weigh the costs and 5 per ton of CO2 equivalent (CO2e) for forest carbon benefits of investments in FFPM. Evaluation should be stocks (CAG undated). The assessment by the World performed at regular intervals (e.g., every five years), Bank of the 2015 forest fires in Indonesia also used to gauge progress, establish priorities, and clarify a price of US$ 5 per ton CO2e (World Bank 2016). budgetary requirements over the next planning cycle. By comparison, voluntary offsets have historically averaged US$ 4.6 per ton CO2e (Ecosystem Because the scope of the valuation would be broader Marketplace 2016).78 An assessment by MoEFCC and than tallying the costs of a single event, a slightly GIZ India of the value of carbon regulatory services As of the end of 2015, includes prices for all years since the establishment of voluntary carbon markets, in current US dollars, as 78. estimated by Ecosystem Marketplace (2016). Strengthening Forest Fire Management in India   100 Box 2.11: Estimating Carbon Emissions from Forest Fires Two implementable approaches for estimating emissions of CO2 and other trace gases: (1) the bookkeeping approach, and (2) the fire radiative power (FRP) approach. Under the bookkeeping approach, above-ground emissions are calculated as: Ei,g=BAg∙ FLg∙ CCg∙ EFi,g where BA is the burned area in forest g (hectares); FL is the above-ground fuel load or biomass per unit area (tons per hectare); CC is combustion completeness (percent of biomass burned); and EF is the emission factor for trace gas or aerosol i for vegetative biomass in forest type g (grams of i released per ton of biomass burned). In a review of the state of the science, Sommers et al. (2014) have noted several major sources of uncertainty involved in quantifying emissions under the bookkeeping approach. Chief among these is burned area. Because ground measurements are resource-prohibitive for large areas, satellite-based measurements of burned area are typically used. The heterogeneity of fuels is the second major source of uncertainty in emissions estimates, as the availability of life and dead vegetation, moisture condition, and other characteristics may vary widely from forest to forest.79 An alternative to the bookkeeping approach is to estimate emissions through the measurement of fire radiative power (FRP). FRP is the amount of energy released by a fire, measured in watts per pixel or unit area. As first proposed by Ichoku and Kaufman (2005) and further refined by Ichoku and Ellison (2014), this method is based on the intuition that emissions are directly proportional to the amount of energy released by a fire, FRP, measured in megawatts or mega-joules per second. The rate of PM emissions is determined by multiplying FRP by a spatially-explicit emissions coefficient, Ce, which is calculated on a per-pixel basis. Total PM emissions may then be converted into different species of trace gases and aerosols using emissions factors such as those developed by Andreae and Merlet (2001). As noted by van Leeuwen et al. (2014), the FRP method has advantages in that it avoids the uncertainties involved with the FL and CC parameters in the bookkeeping approach and may be used to detect smaller fires; however, it suffers from the same limitations of the bookkeeping approach in relying on satellite detections of burned area, which may be obscured by heavy clouds and smoke. from forests in India applied a social value of carbon reducing the infiltration of water and increasing surface of around US$ 73 per ton CO2e, derived from a review runoff and sediment yields from post-fire rainstorms by Atkinson and Gundimeda (2006) (MoEFCC-GIZ (Shakesby and Doerr 2006). In India, heavy rains 2014). with the onset of monsoon season may lead to muddy flood events in fire-affected catchments, particularly Other environmental services that may be degraded in steeply-sloped areas (Schmerbeck and Fiener by severe or repeated fire include the stabilization of 2015). The economic costs associated with increased soils and the regulation of water supply. Fires remove soil erosion and reduced water regulation following vegetative cover and leaf litter and can cause changes severe fires may include increased water treatment to the chemical and physical properties of soils, often costs for downstream cities as well as sediment removal 79. A recent review of FL and CC parameters for different forests around the world is provided by van Leeuwen et al. (2014). 101 Strengthening Forest Fire Management in India from reservoirs and irrigation systems. For example, 2.2.4.2 Post-fire restoration and rehabilitation the Indonesian Ministry of National Development Planning (BAPPENAS) and Asian Development Bank Forest officers surveyed were asked about post-fire (ADB) estimated the costs of erosion and siltation recovery assistance to communities affected by fires. following the forest fires of 1997/1998 in Indonesia at Officers responded that financial assistance may be US$ 1.35 billion (BAPPENAS-ADB 1999). Soil loss may provided by other government departments, including also be valued in terms of nutrient loss, particularly the revenue, welfare, and agricultural departments. if the forest is part of an agro-pastoral production Communities may also receive assistance from non- system. In such cases, damages to soils may be valued governmental organizations such as the Red Cross in according to the costs of artificial fertilizers or manure the event of a disaster. However, many officers said needed to restore soil nutrients (UN ECLAC 2014). that such cases of financial loss to communities had never occurred in their area. The strategic-level assessment of the economic impacts of forest fires should also recognize that some of the According to surveyed officers, ecological restoration environmental services provided by forests to rural and rehabilitation activities in fire-affected forests households are secured by using fire. Recognition of are also limited. An exception is in Uttarakhand, the benefits of fire has largely been absent in the past. where the forest department has engaged in works Even though setting fire in state-managed forests is to rehabilitate water retention and erosion control illegal, fire continues to be an important tool, including services in frequently burned chir pine forests (see as an input to livestock production. In states where figure 2.10). The check dams of chir pine needles fire is an important input to livestock production serves a dual purpose, preventing erosion in gullies and other economic activities by rural households, while also removing flammable material. The an assessment of the economic impacts of fire should effectiveness of these measures has not been assessed. demonstrate how damages from fires can be reduced without adversely affecting rural livelihoods. 2.3 SUMMARY Acknowledging the vital role of local communities in preventing and responding to forest fires, the This chapter assessed policies and prescriptions for assessment of FFPM programs at the strategic level FFPM in India and how those policies and prescriptions should also include the creation of a mechanism are being implemented on the ground at each stage of for monitoring and evaluating the effectiveness of the FFPM process, including prevention, detection, JFM committees and other community participation suppression, and post-fire management. schemes (Saxena 2012). Many case studies have been published of the results of community participation A cohesive policy framework with a clear strategic efforts in specific areas and for specific projects, but direction provides the foundation for successful FFPM. there has yet to be a systematic evaluation of these A vacuum currently exists at the policy level, which programs to judge their effectiveness in controlling the National Green Tribunal has ordered MoEFCC unwanted fire and to identify gaps or constraints to fill by developing a national policy for FFPM and how the programs may be improved. Because in consultation with the states. Though MoEFCC of the diversity of institutional arrangements for had released FFPM guidelines in 2000, they are no forest management by communities across different longer being implemented. The guidelines should be states and regions, such an evaluation would be updated to incorporate the various other guidance most appropriate at the level of the states. The state and instructions that MoEFCC has issued since 2000. assessments should be based on a periodic monitoring Without stronger guidance and standard setting from of activities and outcomes at the village and district above, there will continue to be significant variations level. from state to state and district to district in terms of the detail and substance on FFPM found in local policies and working plans. Strengthening Forest Fire Management in India   102 FIGURE 2.10: CHECK DAMS AND WATER RETENTION WORKS IN FIRE-AFFECTED CHIR PINE FORESTS OF UTTARAKHAND Source: photos from Dr. Rajendra Singh Bisht, “Forest Fires in Uttarakhand: Status, Management issues and challenges,” Presentation to World Bank, Haldwani, Uttarakhand, 30 January 2017 Policies and prescriptions for FFPM should be Detection has been aided tremendously by satellite supported by adequate and predictable financing. technologies, as India has emerged as a leading user A shortage of dedicated funding for FFPM at the of these technologies as part of forest fire monitoring central and state level has been a perennial issue, and response. Using satellite data, Madhya Pradesh which has been documented by the Comptroller and was the first state to develop an SMS-based system to Auditor General in various states. Along with a lack alert field staff to active fires burning in their area. of public engagement, forest officers surveyed for the Since then, the Forest Survey of India (FSI) has rolled assessment cited insufficient equipment, labor, and out a nationwide system. Satellite-based detection financial resources as some of the main challenges has helped fill the gap left by under-resourced for effective FFPM. Revamping the Intensification of ground detection. As these satellite systems continue Forest Management Scheme to focus exclusively on to be upgraded, they would benefit from greater FFPM represents a positive development. Directing integration, including the increased collection of more resources specifically for FFPM will need to field-based reporting for verifying satellite-derived happen at the state level too. fire alerts, as well as improved data sharing between the states and FSI. Only through systematic ground Prevention is the most crucial link in the FFPM chain verification and evaluation can the existing techniques and should receive the greatest support. Prevention for satellite detection be improved. activities have included primarily the creation and maintenance of fire lines and controlled area burning. Forest fire suppression in India mainly involves Only half of the forest officers surveyed in 11 states dryland firefighting. Although the tools used in India said that all the fire lines in their area were being may differ from those used in other countries, the cleared as required per the forest working plans; two- principle of effective suppression remains the same: thirds said controlled burning was not being regularly having a competent, well-trained, and adequately- performed. Other than fire lines and controlled equipped workforce on the ground, ready to respond burning, less emphasis has been given to silvicultural and take immediate action. This workforce includes practices, such as selective thinning and planting field staff from the forest department as well as fire-adapted species. Officers commonly cited a need seasonally-employed fire watchers and volunteers for greater participation by local forest-dependent from the local community. Only a handful of forest communities in fire prevention. department officers surveyed and interviewed agreed 103 Strengthening Forest Fire Management in India that the equipment currently used in their area is Fires larger than a few hectares trigger extra work for adequate. Most cited a lack of basic safety gear and field staff to report and investigate offenses, and the clothing, and many agreed there is a need for more department and its officers may be held responsible training on fire safety and response, especially for for reported monetary damages due to fires. seasonal firewatchers and community volunteers. A more complete assessment of the economic damages Post-fire management is not being treated as part of from forest fires will help make FFPM more of a policy the FFPM process and is probably the weakest link. priority. Current estimates of the economic costs of Post-fire data collection is an essential part of the forest fires in India, at around INR 1,101 crore (US$ fire management process and crucial to producing 164 million, 2016 prices) per year, are almost certainly informed FFPM plans and policies. However, this underestimates. Damages due to forest fires are part of the management process is given little priority generally assessed only for the loss of standing trees and is often performed solely for the sake of fulfilling (natural or planted) in terms of their timber value, administrative requirements. Field reporting and which are usually minimal with low-intensity surface the investigation of fire causes may be hindered by fires such as those that commonly occur in India. insufficient field staff, difficult terrain, and a lack Estimates could be improved by including the loss of of communications infrastructure in more remote environmental services and direct and indirect impacts areas. A lack of standard protocols for collecting and on other sectors, though the states will need help from reporting information on fires, including their causes, MoEFCC and the research community in developing has made it impossible to aggregate data across standard methods and protocols for conducting such states. The greater issue, though, are the institutional assessments. disincentives for accurate and complete reporting. Strengthening Forest Fire Management in India   104 CHAPTER THREE INSTITUTIONAL COORDINATION AND COMMUNITY ENGAGEMENT Preventing and managing forest fires has been and will Among the states providing data on forest area by remain the responsibility of the state forest departments jurisdiction, Uttarakhand and Meghalaya indicated (SFDs), but there are many other stakeholders that also the largest share of forest land not under SFD play a part in FFPM. Chief among these stakeholders management (table 3.1). Uttarakhand is the only state are members of local forest-using communities. to have maintained or reported data on the incidence Others include disaster risk management agencies, of forest fires on non-SFD lands. From figure 3.1, non-SFD public land managers, and the scientific one can see that the risk of fire on non-SFD lands in research community. This chapter draws on a review Uttarakhand is not trivial: fires on non-SFD lands of the academic literature, observations from field accounted for about 35 percent of state-wide burnt visits, and information provided by forest officials, forest area in 2016. disaster management officials and local communities to evaluate how the SFDs can engage more effectively According to the officers surveyed, in most cases, non- with these stakeholders on FFPM. SFD forests are not covered by working plans or similar planning documents. Although the SFD does not have formal jurisdiction, in practice the department is often 3.1 COORDINATING ACROSS held responsible for FFPM on these lands. Explained AGENCIES one officer in Chhattisgarh: 3.1.1 Working with other public and private entities “The same prevention measures are required in managing forested lands the other forested areas as it is done in the forest under the control of forest department. Only Although the SFDs manage the largest part of India’s those other forested areas which are adjoining forest estate, a diverse variety of private, government, to forest areas, controlled burning and fire and community entities also have responsibility for line cutting is done by the forest department forested lands. Nationwide data on the extent of forest to prevent fire from spreading to forest areas cover held by these other entities is lacking. Yet, in all under the control of the forest department.” the states surveyed, forest officers noted that there are forested lands in their area managed by non-SFD The department faces a similar situation in Uttarakhand entities.80 where several officers noted they are “expected to For details on the forest department survey, refer to Annex 2. 80. 105 Strengthening Forest Fire Management in India TABLE 3.1: FOREST AREAS MANAGED BY DIFFERENT ENTITIES Area in square Chhattis- Himachal Kerala Meghalaya Telangana Tripura Uttara- kilometres garh Pradesh khand Forest lands managed by 49,818 32,374 16,071 1,121 32,760 6,294 25,863 SFD Non-forest lands managed No info 21,580 No info No info 72 No info No info by SFD Community-held forest No info*** 725 2 8,371** 6,756 No info 7,350 lands Forest lands under other No info*** 24 No info No info 0 No info 4,926* government entity Forested lands not under No info No info No info No info 2,325 No info No info management Notes: SFD: State Forest Department; * For Uttarakhand, this includes 4,769 km2 under the Revenue Department and 158 km2 of private forest lands and lands held by other government entities (e.g., the military); ** For Meghalaya, this includes the area of private forests under Autonomous District Councils; the breakdown in area is not provided; *** Although no data are available, a nodal officer in the Chhattisgarh forest department indicated the area of community forest land and forest managed by other government entities is “negligible.” Source: SFD data sheets provided to World Bank FIGURE 3.1: BURNT FOREST Thus, it is important for them to prevent fires in those AREA REPORTED IN non-department areas. In Meghalaya, officers said UTTARAKHAND ON STATE the department provides technical advisory services FOREST DEPARTMENT to land managers and financial assistance to village AND NON-STATE FOREST committees charged with fire prevention. However, DEPARTMENT LAND as noted in chapter 2, of the 837,100 hectares of (SQUARE KILOMETERS) private and community-held forest in Meghalaya, working schemes have been implemented for only 30 8,553 hectares. Officers cited funding constraints as State forest the main reason for this shortfall. Non-state forest Burnt forest area (sq km) 25 In addition to forest fire prevention, the SFDs are also 20 often responsible for suppression on non-SFD lands. 15 In at least 5 of the 11 states surveyed, officers said that the forest department was the sole agency responsible 10 for suppressing forest fires. Uttarakhand stands out as an exception. Officers in Uttarakhand suggested 5 that the forest fires of 2016 marked a turning point, as 0 numerous other agencies and departments, including the National Disaster Response Force (NDRF), State 09 15 16 07 13 14 08 10 12 11 Disaster Response Force (SDRF), army, paramilitary, 20 20 20 20 20 20 20 20 20 20 police, revenue department, and health department, Source: SFD data sheets provided to World Bank were involved in a coordinated response under the supervision of the state government. Officers in other prevent and manage the fires in such areas.” As the states said other departments and agencies generally officers explained, the boundaries of non-SFD forest only become involved if a forest fire occurs on the edge areas have yet to be clearly demarcated, and in many of an urban area, town, or village and threatens people’s cases, fires may originate in non-SFD forests and then lives or property. An example of inter-agency forest fire spread to reserved forests under their management. response from Rajasthan is provided in box 3.1. Strengthening Forest Fire Management in India   106 Box 3.1: Coordinating to Control Forest Fires in Rajasthan A major wildfire broke out in Mount Abu, Rajasthan in April 2017, prompting the district administration to request assistance from the Indian Air Force (IAF). Multiple agencies were reportedly involved in responding to the fire. Along with the IAF, army troops, police and forest department personnel worked together to control the fire, and this has been cited as an instance in which there was very good coordination in responding to a forest fire. IAF helicopters airlifted water with Bambi buckets and dropped lakhs of liters of water at various locations in order to douse the fire. IAF crash fire tenders, army personnel and the Central Reserve Police Force (CRPF) also worked on the ground to control the fire. The operations were coordinated by an Air Force station commander, a lieutenant colonel, a sub-divisional magistrate, and an assistant conservator of forests. Earlier, in March 2017, there was another coordinated effort to control a fire in Udaipur, Rajasthan near the Eklinggarh army cantonment. A troop of soldiers reported the fire to the authorities and a joint firefighting operation was launched by the army, forest department and district administration. The troops created fire lanes to check the spread of the flames and to rule out the possibility of the fire spreading into the cantonment or adjoining residential areas. An IAF helicopter was also called in to carry out firefighting operations. In addition, fire brigades from the Udaipur Municipal Corporation, Hindustan Zinc and Eklinggarh cantonment made rounds and NDRF teams from Gandhinagar and Ajmer were also called in for support. The massive operation reportedly involved more than 100 forest department employees and 300 soldiers from the army. Sources: (Press Trust of India, 2017); (Sharma, 2017); (TNN, 2017); Discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017 3.1.2 Working with disaster management authorities and SDMAs to respond to specific events, whether it be an earthquake, flood, fire, or other disaster.82 Disaster management authorities have so far played a relatively minor role in forest fire preparedness Members of the SDMAs in five states were surveyed and response in India. These authorities include to explore the arrangements currently in place for the National Disaster Management Authority incorporating forest fires into disaster management (NDMA), state and district-level disaster management planning and for inter-agency coordination in authorities (SDMAs and DDMAs), National Disaster response to fires.83 Overall, the survey revealed that Response Force (NDRF), state disaster response forces the state disaster authorities have little experience in (SDRFs), army, police, fire departments, and other responding to forest fires and are better equipped and supporting agencies. Under the NDMA, the NDMA, trained for fires in urban or built-up areas. SDMAs, and DDMAs are principally responsible for overseeing the coordination, planning, preparedness, The survey revealed wide variation in the level of and response for all kinds of disasters in their integration of forest fires in the state and district-level respective jurisdictions.81 The NDRF and SDRFs are disaster management plans. In Uttarakhand, forest specialized forces that may be deployed by the NDMA fires are not presently included in the state’s plan. In 81. The NDMA, SDMAs, and DDMAs were established under the Disaster Management Act, 2005. The NDMA has “responsibility for laying down the policies, plans and guidelines for disaster management for ensuring timely and effective response to disaster[s].” The SDMAs are charged with developing “policies and plans for disaster management in the State” and approving district-level disaster management plans. The DDMAs are tasked with “planning, coordinating and implementing body for disaster management and take all measures for the purposes of disaster management in the district in accordance with the guidelines laid down by the National Authority and the State Authority.” Disaster Management Act, 2005, secs. 6, 18, and 30. 82. The NDRF and SDRFs were also created under the Disaster Management Act, 2005. 83. The SDMAs surveyed included those in Kerala, Madhya Pradesh, Odisha, Tripura and Uttarakhand. See Annex 6 for details. 107 Strengthening Forest Fire Management in India Kerala, forest fires are identified in the state’s plan as DDMAs in Tripura have held discussions on forest fire a possible hazard in “forest bordering districts”. In management with community members and officials, Odisha, the Forest and Environment Department has organized special trainings for community volunteers prepared a departmental plan, of which a Forest Fire and disaster management team members on forest Management Plan is an integral part. The Madhya fire suppression techniques using locally available Pradesh SDMA is preparing a comprehensive action/ resources and green branches, produced audio- management plan which covers actions pre-, during video documentaries on handling fires and forest and post-fire events. Forest fires are expected to fires, and put on special awareness programmes by be included in the new state and district disaster the forest departments and fire service departments management plans under preparation, although a in fire prone areas. The SDMA and DDMAs in comprehensive forest fire management plan is not Uttarakhand also conduct awareness raising programs currently in place. In Tripura, a forest fire component in local communities that cover forest fires. In Madhya has reportedly been included in all district disaster Pradesh, the DDMAs have worked with the forest management plans as well as the state plan, and these department in creating informational materials for have been approved by the SDMA. village-level committees to conduct awareness-raising activities with people residing in and around forests. The role of the forest department within the In Kerala, outreach by the SDMA and DDMAs has institutional mechanisms for coordinating disaster been limited to fires preparedness in cities, towns, and planning and response under the SDMA and DDMAs other settlements. also varies considerably from state to state. In Madhya Pradesh, the forest department does not have standing As part of improving forest fire preparedness, disaster or ad-hoc representation in the SDMA or DDMAs, management authorities in Tripura and Uttarakhand though this reportedly will change with the issuance have also conducted mock drills involving the forest of the new state plan. The forest department also departments, disaster responders, members of the does not have standing representation in the SDMA public, and other entities. The largest was conducted in Odisha. In Kerala, the SDMA has instructed all in April 2017 in Uttarakhand, where they organized departments to appoint liaisons to coordinate with the a state-wide drill on forest fires to assess the efficacy authority. In Uttarakhand, the forest department has of integrating the preparedness and response nominated individuals to coordinate with the SDMA mechanisms of the SFD with those of the district in overseeing the response to large fire events that administration. The exercise was carried out across all may threaten life or property. 13 districts and was conducted in collaboration with the state government and agencies, including fire, With varying levels of integration in state/district forest, army, health, police, NDRF, SDRF and civil plans, and varying involvement in the state/district defence. Such trainings are especially important given coordination mechanisms for disaster response, the the inexperience of disaster responders and other point at which other agencies should be mobilized agencies outside the forest department in dealing with to assist with the forest department in forest fire forest fires. suppression remains unclear. The authority of the forest department to call on other assets in responding After forest fires occur, the disaster management to forest fires is also limited. However, as a SDMA authorities generally do not have much of a role representative from Kerala explained, in general, in recovery, restoration, or rehabilitation. It was unless there is a threat to life or property, resources asserted that the Kerala SDMA does have a role in from the National Disaster Response Fund or State case of civilian areas including areas of indigenous Disaster Response Fund may not be utilized to support population, but not in the case of notified forest the costs of suppressing forest fires. areas. In the case of Madhya Pradesh, it was highlighted that the forest department has a plan Beyond disaster planning and response, in some of for recovery after a fire event, and that the authority the states surveyed, the SDMA and DDMAs have would approve a need-based plan, if required. It cooperated with the forest department on public was further noted that, in general, the Odisha State outreach. In Tripura, for instance, the SDMA and Disaster Management Authority (OSDMA) does not Strengthening Forest Fire Management in India   108 have a role in recovery after a forest fire incident. interviews with members of forest communities were The Tripura SDMA, however, is said to have a role to also conducted in Jharkhand, Madhya Pradesh, play in post-fire recovery, whereas the Uttarakhand Odisha, and Telangana. Disaster Mitigation and Management Centre is said to have no role in recovery after a fire event. Forest officers in all the states surveyed acknowledged that communities living in and around forested areas play an indelible role in preventing and managing 3.2 ENGAGING WITH COMMUNITIES forest fires. Most (83 out of 101) rated the role of the local community as either very or extremely With some 150 million people living in or near important in managing forest fires. Yet, officers had forests (FSI 1999)—and with nearly all forest fires mixed views as to the effectiveness of the department’s being caused by people—local communities are the current engagement with the local community, lynchpin of effective FFPM in India. The engagement mostly rating it as fair, somewhat poor, or very poor of the forest department with local communities (figure 3.2). on FFPM was assessed through a survey of forest officials as well as consultations with community As surveyed officers explained, the forest department members. In-depth community appraisals, including has engaged with communities in a variety of ways. structured interviews and focus group discussions, The most common entry point is for the department were performed with forest-using communities to work with the JFMCs. The second most common in Uttarakhand and Meghalaya.84 Field visits and method named by surveyed officers is for the FIGURE 3.2: SELF-RATED EFFECTIVENESS OF THE FOREST DEPARTMENT’S ENGAGEMENT WITH THE LOCAL COMMUNITY IN PREVENTING FOREST FIRES Uttarakhand 14% 43% 21% 14% 7% Tripura 11% 33% 11% 11% 33% Telangana 29% 14% 29% 14% 14% Odisha 14% 14% 14% 29% 29% Meghalaya 27% 18% 27% 18% 9% Kerala 11% 33% 56% Jharkhand 10% 30% 50% 10% Himachal Pradesh 23% 23% 38% 8% 8% Chhattisgarh 67% 17% 17% Assam 33% 17% 33% 17% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Share of respondents Very poor Poor Somewhat poor Fair Somewhat good Good Very good Note: responding officers = 95 Source: World Bank survey of state forest department officers, April-August 2017 For details on the community consultations in Uttarakhand and Meghalaya, see Annex 3. 84. 109 Strengthening Forest Fire Management in India department to conduct awareness-raising activities, 3.2.1 Incentives to communities for example, by issuing public announcements or organizing meetings in local towns and villages. The Surveyed officers who were asked how engagement third most common method is for the department with the local community on FFPM could be improved to hire people from the local community as seasonal pointed to the need for incentives more than anything fire watchers. Other methods of engagement are else (figure 3.4). The importance of involving local illustrated in figure 3.3. communities through an incentive-based mechanism FIGURE 3.3: METHODS OF COMMUNITY ENGAGEMENT USED BY THE FOREST DEPARTMENT Public awareness raising (non-specific) Public announcements Meetings and workshops Public gatherings and performances School programs Joint Forest Management Committees Other community institutions Women’s groups Van Panchayat Gram Panchayat or Gram Sabha Firewatchers Other employment Wages for clearing fire lines Provision of incentives (not specified) Development projects/programs Monetary payments for prevention Public awareness raising Rights and concessions Working with community institution Prizes Employment Provision of equipment for firefighting Provision of Incentives Fire lines and controlled burning with community Joint implementation of FFPM Involvement in management and planning Monitoring, patrolling, or policing Fire response 0 5 10 15 20 25 30 35 40 45 50 Frequency (number of responses) Note: responding officers = 94; officers may indicate more than one method of engagement Source: World Bank survey of state forest department officers, April-August 2017 Strengthening Forest Fire Management in India   110 is also noted in state policies, such as Tripura’s State Monetary payments have generally been offered by Action Plan on Climate Change and the State Strategy the state forest departments to JFMCs or villages in and the Action Plan on Climate Change for Himachal exchange for fire protection services. In Kerala, for Pradesh.85 The need to strengthen incentives was example, the forest department has offered JFMCs further emphasized in a parliamentary committee (also known as vana samrakshana samithis, or VSSs) report presented to the Rajya Sabha in 2016 (Rajya INR 1,000 for each hectare of forest they protect. Each Sabha 2016). VSS typically manages 150-200 hectares, translating into INR 1.5-2 lakh per VSS.86 In Meghalaya, the state Incentives to communities for forest fire prevention forest department has offered payments to communities can take many forms and have commonly included in exchange for creating and maintaining community monetary payments/rewards, jobs, and concessions for reserved forests. In Odisha, the forest department forest minor produce. rewards the best-performing VSS in each district with FIGURE 3.4: HOW CAN ENGAGEMENT WITH THE LOCAL COMMUNITY BE IMPROVED? Provision of incentives Public awareness raising programs Employment Work more with community institutions Changes in village government/institutions Policy changes Provide extension services More focus on the youth Joint implementation of FFPM Response unclear or unspecific More systematic programming People just aren’t interested Improve communication with communities Monitoring, patrolling or policing Address human-wildlife conflicts Broader engagement More of the same 0 10 20 30 40 50 Frequency (number of responses) Note: responding officers = 91; officers may indicate more than one way to improve engagement Source: World Bank survey of state forest department officers, April-August 2017 85. State action plans on climate change are available from MoEFCC at http://www.moef.nic.in/ccd-sapcc. Ayyapan, R. 2017. “Wildfire threat looms large in Kerala.” Deccan Chronicle, February 28. http://www.deccanchronicle.com/nation/in- 86. other-news/280217/wildfire-threat-looms-large-in-kerala.html 111 Strengthening Forest Fire Management in India INR 20 lakh; the top VSS in the state receives a cash 3.2.2 Support for community institutions reward of INR 200 lakh and a certificate.87 Though not enough to completely defray the costs of fire As the JFMCs and other community-level institutions prevention by JFMCs, these monetary rewards offer have been the main entry point for the forest an important means of recognition and a behavioral department to engage with local forest users on nudge to encourage more active participation. FFPM, this engagement may be enhanced by investing in strengthening these institutions. In Meghalaya, One of the most common ways for the forest for instance, community members pointed to the departments to engage communities in FFPM by creation of Village Fire Control Committees (VFCCs) providing employment is through hiring seasonal fire as a good example of how the forest department can watchers. Most of the forest officers surveyed said the play a positive role in bringing communities together department hires fire watchers in their area. The hiring to protect community-owned forests. Additional of fire watchers from specific villages has sometimes, case studies of community institutions for forest but not always, been tied to those villages’ performance management and FFPM in Meghalaya are provided in Annex 3. on fire prevention. In some cases, the department hires fire watchers directly. In others, the JFMCs are With such a diversity of community-level institutions entrusted to select fire watchers. Due to out-migration for forest management across different states and and a reduced dependence on forests for livelihoods in regions of India, there is no universal formula for some areas, for example, in Uttarakhand, some forest strengthening forest fire management by these officers noted they had difficulty in finding enough institutions, and the creation of VFCCs may not people from the local villages to work as fire watchers. be appropriate for all areas. As some forest officers Delays in the payment of wages to fire watchers is also cautioned, it may be better to work with existing common88 and has hurt the ability of the SFD to hire community or village-level institutions than to more field staff. create new ones. Recommendations for creating new institutions or programs should consider local Allowing local community members to harvest NTFPs cultural, financial, and social constraints. in exchange for limiting unwanted fires has been an effective incentive in many areas, including in Madhya 3.2.3 Equipment and training Pradesh, where the forest department has incentivized local forest users to manually prune tendu trees Community members interviewed in several states, instead of burning to promote fresh shoots (box 3.2). including in Jharkhand, Madhya Pradesh, Meghalaya, Odisha, and Uttarakhand, said they are often called Forest officers who were surveyed also offered ideas for to help fight forest fires, though they typically do how incentives could be improved. Some recommended not receive any equipment or training to do so. that these incentives be made contingent on zero Odisha was the only state visited where community fires occurring in local forests. Others stressed that firefighters from the local JFMCs were provided with the responsibilities of the community must be clearly protective clothing, such as cotton drill uniforms, defined and that the provision of incentives should be helmets, goggles, or boots. The JFMCs were also regular and predictable. As one officer in Himachal provided with leaf blowers for clearing fire lines. In Pradesh noted, “Incentives have to be given at fixed Uttarakhand, community members said that trainings rates and without fail.” Several officers proposed that on forest firefighting were not regularly conducted incentives be provided via communal institutions such by the fire department in the villages. When trainings as the JFMCs. were done, they were usually very short and general. As an alternative, community members suggested that 87. T.A.K. Sinha, “Forest Fire Protection in Odisha,” presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017. 88. Of the surveyed officers who said fire watchers in their area are provided wages, about half (30 of 66) noted delays or shortages in payments. Strengthening Forest Fire Management in India   112 the department organize field-based trainings, which Jagran (awareness creation programs), public rallies, would include, for example, instruction on first aid and yatras on a regular basis, and has disseminated and the use of controlled burning in suppression. information to local communities through short films, banners, and brochures. Also, the department has 3.2.4 Public outreach and awareness raising planned sensitization activities for tourists and visitors throughout the yatra season in most fire-prone areas. According to forest officers surveyed and interviewed, Awareness campaigns involving the distribution of awareness raising is the one aspect of working with banners, posters, handbills and stickers are regularly communities on which the state forest departments organized as well. Puppet shows, street rallies, have probably focused the most effort. In padayatras89, oath ceremonies, lectures in schools, Uttarakhand, for example, the SFD has organized Jan signature campaigns etc. are also organized at the Box 3.2: Partnering with Communities in Madhya Pradesh to Prevent Burning for Tendu Leaves Madhya Pradesh considers itself a leader in developing and introducing new initiatives that improve fire management. Involving communities has been critical in Madhya Pradesh, as the forest department has worked with the Joint Forest Management Committees to ask local people for solutions and to promote alternatives to burning for collecting non-timber forest products such as tendu leaves. The state’s policy on tendu changed in 2004, when collection was shifted to cooperative societies. In supervising the tendu collection process, the SFD in Madhya Pradesh has worked with communities to promote manual pruning instead of fire. Pruning is done in February and March, during the peak fire season. Mature leaves may be collected around 45 days after pruning. Tendu leaf collection is carried out by the Tendu Patta Samiti - cooperative societies organized under the Madhya Pradesh State Minor Forest Produce Federation. The federation is tied administratively to the SFD - the managing director at the apex of the federation holds the position of Principal Chief Conservator of Forest (PCCF). All residents of villages with JFMCs may register as members of the federation. Residents are given cards/permits by the forest department allowing them to collect tendu leaves at a designated place. Residents deliver leaves to the forest department, which records how many bundles they have collected. Tendu traders then submit bids to the forest department to buy a certain amount of leaves, so the department acts as mediator in the supply chain. The SFD deposits any collection fees owed to the villagers plus profits from tendu sales into the account of the federation. Village residents then draw from the federation account to receive their collection wages plus a bonus. Wages are the same for all villages, as set by the federation, and were INR 1.25 per bundle of 50 leaves in 2016. Community members in the Raisen District that the World Bank team visited are allowed by the forest department to collect a variety of NTFPs, such as honey, fruit, medicinal plants, mahua flowers, and tendu leaves. Landless residents especially rely on these forest products for their livelihoods. Local people are also employed by the forest department for labor on plantations and nurseries and to help maintain check dams and fire lines in the forests, particularly during the lean season between planting and harvest. Plantation work sponsored by the department between sowing and harvest may earn them INR 200 per day. Residents say they have been trained by the forest department to remain with the fires they set, though some fires occasionally get out of control and spread into the forest. Unlike in other states, residents do not use fire to collect tendu leaves. Instead, the SFD has worked with communities to promote manual pruning instead of fire. Slashing or pruning is done in February and March about 6 weeks prior to harvesting. Source: World Bank Field Visit to Madhya Pradesh, January 2017; Agnihotri (2017) A padayatra or “journey by foot” is a foot pilgrimage by a political leader to interact directly with the community and raise awareness or 89. rally support for an issue 113 Strengthening Forest Fire Management in India Box 3.3: Scientific Research Organizations Studying Forest Fires in India - Forest Survey of India (FSI), Dehradun, Uttarakhand: As described earlier in this report, FSI carries out satellite-based detection of forest fires and disseminates alerts to State Forest Departments. It has also begun issuing pre-fire alerts and is moving towards creating a full-fledged FDRS. There are significant opportunities for states to collaborate with FSI to develop a full-fledged FDRS which meets their needs. FSI offers trainings to forest department officers aimed at exposing participants to new tools and technologies for forest fire monitoring and damage assessment. - Indian Council of Forestry Research and Education (ICFRE): founded in 1986, ICFRE is an autonomous entity under MoEFCC constituting the “apex body in the national forestry research system.”90 The Council has a nationwide presence with nine research institutes and five research centers representing the different bio-geographic regions of India. - Forest Research Institute (FRI), Dehradun, Uttarakhand: a research institute under the ICFRE, FRI’s work on forest fires includes creating awareness and providing knowledge to a range of stakeholders such as students, researchers, forest department personnel, and technologists on FFPM and forest resource management more broadly. FRI’s capacity-building programs for forest departments include modules on forest fire mitigation, and FRI has assisted departments in crafting FFPM strategies in working plans. FRI has also developed forest firefighting equipment kits for department personnel working in difficult terrain (Singh 2017). - National Remote Sensing Centre (NRSC), Hyderabad, Telangana: NRSC is an entity under the Indian Space Research Organisation (ISRO), responsible for remote sensing satellite data acquisition and processing, data dissemination, aerial remote sensing, and decision support for disaster management. The Decision Support Centre (DSC) established at NRSC in 2005 provides data and information services in near-real time to central government and state departments for a range of natural disasters, including forest fires. NRSC provides near-real time satellite data to FSI for the generation of active fire alerts. NRSC scientists have also performed nationwide estimates of burnt area and carbon emissions from above-ground vegetative biomass (Reddy et al. 2017a and Reddy et al. 2017b). - North Eastern Space Applications Centre (NESAC), Umiam, Meghalaya: NESAC has mapped fire incidents for states in the Northeast region and helped departments identify forest areas for management priority (see Chakraborty et al. 2014). In addition to providing active fire alerts to state forest departments, NESAC also provides geospatial inputs to the departments in preparing forest working plans. Inputs include information on land resources, forest types, forest density, species density and composition, and so on. - Indian Institute of Remote Sensing (IIRS), Dehradun, Uttarakhand: Another ISRO institute, IIRS specializes in capacity building on remote sensing, geo-informatics and their applications. IIRS has worked with Uttarakhand state officials to identify regions prone to forest fire and to conduct burnt area analysis in the Rajaji and Corbett Tiger Reserves (Rajya Sabha 2016). - Wildlife Institute of India (WII), Dehradun, Uttarakhand: WII offers training programs, academic courses, and advisory services on wildlife research and management and is engaged in research on biodiversity-related issues in India. WII has advised on preventative/remedial measures for forest fires, restoration of habitation, wildlife habitat improvement and post-fire restoration and rehabilitation. ICFRE, “History,” http://icfre.gov.in/history. 90.. Strengthening Forest Fire Management in India   114 - Ashoka Trust for Research in Ecology and the Environment (ATREE): ATREE has formed a network of researchers to study the ecological, socioeconomic and cultural causes and consequences of fire in Indian ecosystems to see whether there is a way to accommodate the differing forest management goals of a diverse range of stakeholders. Focuses of this research have included fire and invasive species, traditional controlled burning practices, and the effects of fire on regeneration (see chapter 1). - Centre for Ecological Sciences, Indian Institute of Science (IISc), Bengaluru: Researchers from IISc have conducted long-term studies of forest fires on forest ecology in the Western Ghats, including in Mudumalai National Park and Wildlife Sanctuary (Sukumar and Suresh 2017). Additional sources: excerpted and adapted from ATREE, “A Fiery History,” http://atree.org/fiery_history; FSI, “Training,” http://fsi.nic.in/details. php?pgID=mn_6; ICFRE, http://icfre.gov.in/; NESAC, “North Eastern Regional Node for Disaster Risk Reduction,” http://www.nerdrr.gov.in/; WII, http://www.wii.gov.in/ local level. Such campaigning is ongoing throughout addressing the challenge of frequent, unwanted the fire season. fires in India’s forests, especially in the context of a changing climate. The opportunities for training forest officials must also be tapped into for improving 3.3 COLLABORATING WITH FFPM outcomes on the ground. RESEARCH ORGANIZATIONS There is no dearth of excellent research organizations 3.4 SUMMARY in India that have been working on various aspects of FFPM. The Forest Survey of India (FSI) and Indian The state forest departments (SFDs), under the overall Council of Forestry Research and Education (ICFRE), guidance of MoEFCC, have primary responsibility for both headquartered in Dehradun, Uttarakhand, preventing and managing forest fires; however, they stand out as potential host institutions for a center do not operate in isolation. This chapter evaluated the of excellence that can provide guidance to SFDs and coordination between the state forest departments and develop new methods for preventing and managing the other stakeholders involved in FFPM, including the forest fires. Stronger collaboration of the SFDs with disaster management authorities, local communities research entities would enable the states to conduct of forest users, and research organizations. experiments and provide data to these institutes for further developing and refining their research in the The SFDs manage about 654,137 km2 of forest lands field, ultimately leading to better fire management contained in reserved and protected forests, plus outcomes on the ground. Indeed, FSI and FRI are much of the 113,881 km2 of unclassed forest. Together, already active in providing training and technical these lands comprise about 23 percent of India’s support to the state forest departments. geographical area (FSI 2018). Not all these areas are forest covered, and additional areas of forest cover Research organizations and others are also important exist outside the jurisdiction of the departments. In sources of knowledge on the long-term impacts of fire, practice, the SFDs often assume sole responsibility for which can help inform and guide the FFPM planning forest fires on these non-department lands. National process in the country. Box 3.3 provides a select list data on the forest fires on non-department lands is of these research entities that are actively engaged lacking, though data from Uttarakhand show that in studying FFPM-related issues. Strengthening the these lands accounted for about 35 percent of state- collaboration between forest departments and these wide burnt forest area in 2016. The threat of fire researchers working on FFPM is critical for efficiently on non-SFD lands is non-trivial, and fires started 115 Strengthening Forest Fire Management in India outside state forests may spread to state forests. Better fires in forests by local people is an unattainable goal. coordination with other forest land managers and Thus, the SFDs must strike a fine balance, working more clearly defined responsibilities (including for the with communities to make sure fire is used responsibly provision of funds) are needed. in a way that promotes forest health, while avoid damaging and out-of-control fires. Though large fires such as those observed in Uttarakhand in 2016 and Karnataka in 2017 do Forest officers interviewed and surveyed for this occur, forest fires are not typically treated as disasters, study agreed that more effective engagement with and the disaster management authorities have so far communities will hinge on better incentives. Existing played a minor role in FFPM. A survey of the state incentives have included monetary rewards, the disaster management agencies (SDMAs) revealed provision of jobs to community members, and access a wide variation in how forest fires are treated in to harvest NTFPs from state forests. The Joint disaster planning and how institutional mechanisms Forest Management Committees (JFMCs) have been have been set up for organizing the response to large the primary avenue through which the SFDs have or destructive fires. Thus, the point at which other offered such incentives. Monetary payments have agencies should be mobilized to assist the SFDs with not typically been enough to cover the costs of fire forest fire suppression remains unclear, and the prevention work by the JFMCs but rather have served authority of the forest department to call on other as a behavioral nudge. Seasonal firewatchers and assets in responding to forest fires is also limited. community volunteers are rarely provided equipment and training for FFPM. More effective coordination with local communities— the primary forest users in India—is essential. Researchers have been an underutilized part of the Strategies for FFPM should be founded on a clear FFPM community. Stronger collaboration between recognition of how local communities depend on the SFDs and research entities would enable states to forests for important goods and services and aim better monitor the ecological and economic impacts to ensure the delivery of these goods and services of fires, to develop robust protocols for gathering fire while also reducing damaging and unmanaged fires. data, and innovate new science-based management Although all forest fires are treated as an offense approaches for preventing fires and rehabilitating under existing laws, completely excluding the use of fire-affected areas. Strengthening Forest Fire Management in India   116 CHAPTER FOUR RECOMMENDATIONS MoEFCC has requested the World Bank to assess A first-order national policy on FFPM would establish the current system for FFPM in the country, identify the guiding principles and provide the framework constraints and gaps in how FFPM is implemented, for FFPM in India, beginning with a clear statement and make recommendations for how FFPM may of goals and priorities. The experience of Mexico be improved. The following chapter consolidates in revamping its national FFPM program in 2013 the findings of the assessment in the previous exemplifies the importance of goal-setting (box 4.1). chapters and puts forth recommendations to inform The country reoriented its national policy on FFPM national-level policy for FFPM. The chapter also away from the total suppression of fires toward a presents recommendations offered by stakeholders recognition of the ecological and social functions of in MoEFCC, NDMA, NITI Aayog, the state forest fire, acknowledging that some fires can be beneficial departments, the research community, and NGOs and seeking to maximize the benefits of fire while at an international workshop on FFPM organized minimizing the negative impacts. Mexico’s FFPM by MoEFCC and the World Bank in New Delhi in program has also aimed to improve coordination November 2017. between federal, state, and local agencies and to increase participation by communities. The United States has undergone a similar process in crafting the 4.1 POLICIES, PLANS, vision and principles for FFPM in its National Wildland REGULATIONS, AND FUNDING Fire Cohesive Strategy, issued in 2014. The strategy represents the culmination of a five-year process, 4.1.1 Formulation of a national policy or action plan with a focus on resilient landscapes, fire-adapted for FFPM communities, and safe and effective wildfire response. The vision of the strategy is to “safely and effectively The National Green Tribunal issued a ruling in August extinguish fire when needed; use fire when allowable; 2017 calling for MoEFCC to formulate a national manage our natural resources; and as a nation, to live policy or guidelines for FFPM in consultation with the with wildland fire” (US DOI-DOA 2014). In India, the states. The development of a national action plan for development of a strategic vision for FFPM through a FFPM is underway in MoEFCC.91 consultative process with the states would help build Per discussions with MoEFCC, as of the time of writing in early February 2018, a committee to prepare the plan had been formed, 91. though the plan had not yet been drafted. 117 Strengthening Forest Fire Management in India consensus on the purpose and priorities of FFPM and a process or mechanism for the provision of regular establish a direction for subsequent policies and plans. funding for FFPM to the states. Ideally, a national policy on FFPM should take the form of an inter- A national action plan would also offer an opportunity agency document that could clarify the roles and for MoEFCC to review and consolidate the existing responsibilities for FFPM horizontally across other policies and guidance on FFPM that it has issued agencies also responsible for managing forested lands over the years. This would include the guidelines for (such as the Revenue Department) and responding to FFPM issued by the Ministry in 2000, which are not forest fires (such as the National Disaster Management widely known and are no longer being implemented. Authority). Moreover, given that current Government of India policies to increase carbon sinks through forestry A national policy should also include guidelines for programs can also help to tackle forest fires, the the development of standard operating procedures national action plan should also reference relevant (SOPs) by the states for various aspects of FFPM, climate change policies. such as requirements for post-fire data collection and reporting. SOPs are an excellent way for the states to A national FFPM policy should clearly delineate the communicate the objectives, principles, and required respective roles and responsibilities of MoEFCC and actions for effective FFPM down to field staff in the the state forest departments, including establishing state forest departments. The SOPs also provide Box 4.1: Transforming Forest Fire Policy and Practices in Mexico Mexico has recently re-assessed its national policy on forest fires, from a policy of total suppression to a more integrated policy of fire management. This transition took place with a growing recognition that some fires are beneficial – ecologically, socially, and economically. Until 2012, Mexico’s national forest fire program focused on the complete suppression of fires by contracting helicopters to douse flames. In addition, state forest fire programs were weak and there was little institutional coordination. In 2013, it was recognized that the total suppression of fires was not enough, so the country set out to revamp its national forest fire program with the context of a changing climate. An institutional consensus emerged around the need to develop a public policy that recognized the ecological and social role of forest fires, acknowledging that some fires can also be beneficial. For instance, some ecosystems, such as pine forests, are adapted to fire, as fire releases seeds from cones and promotes regeneration. Some of the other benefits of forest fires that have been recognized relate to the control of pathogens, invasive species, maintenance of natural pastures, and improvements in habitat for wildlife. Achieving this shift in Mexico’s approach to FFPM took time and strong institutional, technical, scientific, and social leadership. The transition provided a unique opportunity to reform forestry policy while at the same time making improvements in operations under existing laws. Mexico has been able to improve its approach to FFPM without increasing the budget. Instead, the focus in Mexico has been on allocating resources more effectively and efficiently to strengthen the two fundamental pillars of fire management: better coordination between three levels of government, as well as greater participation by society. Some of the measures that have been implemented since 2013 include increasing community-based fire management and training for rural crews; establishing agreements between CONAFOR (the National Forestry Commission in Mexico) and federal, state and local agencies; constructing national, regional and state centers for forest fire control; increasing the number of forest firefighters from 5,000 to 22,000; upgrading personal protection equipment; acquiring tools, vehicles, and tanker trucks; improving the management of fuels; building the capacity of forest firefighters and technical staff; strengthening basic research (including on fire danger rating and fuel models); promoting public engagement; and bolstering international cooperation with the United States, Canada, the Dominican Republic, Colombia, and other Central and South American countries. Source: Alfredo Nolasco Morales, National Forestry Commission, Mexico, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017 Strengthening Forest Fire Management in India   118 Strengthening Forest Fire Management in India   119 a medium for the states to consolidate the various policy across the states that would allow states to create orders, instructions, and letters they have issued from new fires lines where necessary. While limited financial time to time on different aspects of FFPM. Guidelines resources were cited by surveyed officers as the main issued by MoEFCC may elaborate on: reason for not clearing fire lines, the stocktaking may also assess funding gaps. • Revision of working plan codes • Development and requirements for regular As with fire lines, nationwide information about the use updating of SOPs by the SFDs of controlled burning is lacking. Controlled burning • Implementing a common classification scheme for is not required for all forested areas, but where it is the causes of forest fires stipulated, it is not always performed. Most of surveyed • Standard protocols for post-fire reporting and forest officials who said controlled burning was data collection, including burnt area estimation, required in their area per forest department working investigation of suspected or probable causes of plans admitted that burning was not regularly done. fire, and damage assessment A policy on controlled burning would help establish • Incentivizing accurate post-fire reporting by field greater regularity to where, when, and how burning staff on fire occurrence, burnt area, and damages is done. Australia’s national guidelines for controlled burning (box 4.2), established through a consultative The process of formulating the national policy or process, provide a best-practice example of such action plan on FFPM would be just as important as a policy. The performance of controlled burning the policy or plan itself. The process should be open, should be monitored by the state forest departments, consultative, clearly defined, and time-bound. A core with MoEFCC playing a supervisory role at the group with the Director General of Forest, MoEFCC national level. and representatives from the SFDs, NDMA, NGOs, and research institutes should be established immediately 4.1.3 Review of MoEFCC’s Working Plan Code to initiate this process for the development of the national policy and action plan over the course of one Working plans set forth area-specific requirements to two years. The group would provide the mandate for FFPM among other aspects of scientific forest and overall work plan for the process and meet management. A review of working plans in 11 states regularly to monitor progress, resolve pending issues showed that the amount of detail contained in the or questions, and make recommendations for further plans for fire prevention and management varies action. greatly from area to area, without much consistency. 4.1.2 Coordinated policy for maintaining and The National Working Plan Code issued by MoEFCC creating fire lines and for controlled burning in 2014 suggests that working plans include details about the historic occurrence of fire, the area affected MoEFCC should oversee a stocktaking of fire lines by fire in previous years, the locations of fire lines, by the states to determine the location, length, and and fire protection work undertaken in previous functionality of existing lines. Half of the state forest years. The Code also suggests the part of the working department officers surveyed by the World Bank said plan focused on future management should include that required fire lines in their area were not cleared, a section on “Associated regulations and measures” and in only one of the 11 states surveyed had the forest that may describe fire protection work to be done. department mapped and digitized the locations of Beyond outlining the elements of a typical working fire lines. It is unclear how effective existing fire lines plan, however, the Code does not establish any specific are in preventing out-of-control fires and whether requirements for fire prevention or control as part new lines are required. MoEFCC should assess the of a plan, nor does the Code provide any guidance functionality of current fire line locations and the on what types of fire prevention and control actions need for additional lines, taking into account land should be required for different forests and areas. use changes and new roads, railways, power lines, and other infrastructure that have been built since MoEFCC should revisit the 2014 Working Plan Code the fire lines were first drawn. In parallel with the to determine whether more substantive guidance on stocktaking, MoEFCC should establish a coordinated FFPM is required. For example, in areas where the Strengthening Forest Fire Management in India   120 maintenance of fire lines is prescribed, the Code could be led by the Deputy Inspector General of should require that the coordinates, length, width, Forests (Forest Protection) or other senior MoEFCC and budget for fire lines are clearly stipulated in officer with representatives chosen from the state the working plan. Another suggestion would be for forest departments. After the Working Code is revised, the Code to establish a clear and empirically-based additional training may be needed by division-level method for fire risk zoning so that states and divisions field staff in carrying out the new requirements of can determine which fire-prone areas warrant the Working Plan Code for FFPM. This content may special attention for fire prevention measures. The be integrated into the standard training curricula for methods for risk zonation implemented by Odisha forest officers on FFPM (see 4.4.1 below). MoEFCC’s and Telangana offer a best-practice example for other Research and Training Division may provide states. Revisions to the Working Plan Code pertaining additional support. to FFPM may form part of the national policy or action plan for FFPM recommended in 4.1.1 above. 4.1.4 Clarify MoEFCC’s position on maintenance MoEFCC should lead this process in consultation with of fire lines, silvicultural operations, and other fire the state forest departments. prevention practices in areas where green felling is restricted As with the development of a national policy or action plan, the process for revising the Working Plan Code The creation of a national FFPM policy provides an should be formalized, with a clearly defined timeline opportunity for MoEFCC to make a clear statement and a core working group or committee. This group on the need to clear fire lines and conduct silvicultural Box 4.2: Using Fire to Prevent Conflagrations in Australian Forests Deliberate burning of forests and grassland in Australia dates back more than 40,000 years. The use of fire was central to the way of life of Aboriginal people throughout Australia. Some post-1788 settlers learnt from and tried to adopt aspects of burning practice from Aboriginal people, but they often used fire for different purposes. Nevertheless, there was universal recognition by both, the post-1788 settlers as well as Aboriginal people, about the value of using fire for risk mitigation in the context of FFPM. In response to high-consequence fire events occurring in the late 19th and early 20th centuries, government policies aimed at excluding fire were attempted, but these efforts failed. In response to the catastrophic 1939 “Black Friday” fires in Victoria, the Stretton Royal Commission recommended a strategic program of burning selected areas of forest in a controlled manner in spring and autumn. As a result, planned burning became an official fire management practice in Victoria. Prescribed burning for community and asset protection has been used by Australian public land management agencies since the 1970s, with early development of systematic approaches and techniques founded in the 1960s. Today, fire prevention practices in Australia include regular controlled burning. Australia has developed National Guidelines for Prescribed Burning Strategic and Program Planning. Keeping in mind the very wide range of operating environments and operational risk profiles that can be found in Australia these guidelines establish a logical, consistent and robust planning and works implementation process. In fact, the Australian Fire and Emergency Service Authorities Council (AFAC) and Forest Fire Management Group (FFMG) have released a National Position on Prescribed Burning AFAC and FFMG member agencies take the position that “Prescribed burning is an essential part of bushfire mitigation across the Australian landscape to reduce risk to communities and to maintain ecological health” (AFAC 2016: 3). The health and safety of firefighters and the public is a key priority for fire prevention, and fire managers have formed partnerships with and have closely involved Aboriginal communities in FFPM planning and operations. Source: AFAC (2016); Tim McGuffog, Forestry Corporation of New South Wales, Australia “Fire Prevention”, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017. 121 Strengthening Forest Fire Management in India interventions to reduce unwanted fires and promote working conditions, and incentives must be sufficient healthy, productive forests at elevations above 1,000 to ensure that these vacancies are filled in fire-prone m and in other areas that are currently affected by forest areas. Having boots on the ground is essential court-ordered restrictions on green felling. The policy for implementing all aspects of FFPM, including may stipulate that such interventions should only be preventing, detecting, and suppressing forest fires. permitted where they promote the stated management Staffing should be a top priority for increased funding goals for the affected forest. For example, in protected for FFPM. areas, selective thinning or prescribed burning may be beneficial to maintaining habitat for species of wildlife the reserves are designed to protect. Where protected 4.2 FIRE PREVENTION PRACTICES areas are threatened by encroaching lantana and other invasive species, controlled early-season burning 4.2.1 Continued development of systems for early in line with traditional practices may help keep these warning and fire danger rating invasives in check. In areas above 1,000 m, fire line clearance and other silvicultural operations may International experience has shown that early warning prevent more severe conflagrations that would kill and fire danger rating systems (FDRS) developed trees, strip soils, and harm downstream watersheds with input by local fire managers and tailored to local with increased erosion. conditions are more likely to be successful than systems that are imported directly from other contexts. For 4.1.5 Provide dedicated and regular funding for this reason, FSI’s effort to develop a locally-tailored FFPM, with a focus on optimizing the allocation of system for India is a worthwhile endeavor and should existing financial resources be supported. FSI should be encouraged to continue to test new elements of the system and modify it for In the near term, rather than seeking to increase local use in the variety of vegetation types, climates, total spending, states should examine existing budget and topographies that characterize India’s forests. resources to determine if enough is being allocated Initial versions of the FDRS should be robust but for FFPM. Emphasis should be on the adequate also easily understood by, and implemented by field- protection of existing forest resources before pursuing based forest fire practitioners. As expertise in both plantations and afforestation projects. the development of a useful model and its field implementation grows, more sophisticated versions CAMPA offers a potential source for funding for that account for particular forest types and climatic FFPM if the institutional obstacles can be resolved and zones should be encouraged. more of the ad-hoc CAMPA funds can be unlocked. Strengthening FFPM is in line with the stated As the experience in other countries has illustrated, the objectives of CAMPA. States may also seek provide development of an FDRS is a long-term project that greater financing for FFPM and forest protection requires the involvement of a variety of stakeholders. through increased revenue generation by increasing Canada, for example, began work on its FDRS in 1968 the productivity of public forests. and continues to refine its system today. At each stage of the process in developing the Canadian FDRS, the 4.1.6 Ensure adequate budgetary resources Canadian Forest Service (the lead agency) has worked and incentives to fill field staff vacancies in forest in conjunction with the provinces and territories92. departments In the case of India, FSI is well-positioned to continue Audits by the Comptroller and Auditor General (CAG) playing the leading role as the champion and chief have consistently found that state forest departments developer of systems for early warning and fire danger are understaffed and unable to fill vacancies, rating. The resulting systems are more likely to be especially at the field level. Budgetary resources, successfully adopted if the state forest departments Brian Simpson, “The Canadian Forest Fire Danger Rating System,” Presentation and discussion at the workshop on Forest Fire 92. Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017. Strengthening Forest Fire Management in India   122 also play an active role in the process of development. section 1 above, such interventions would benefit from The states could be involved through the creation a coordinated policy and SOP for where they should of a structured process for field testing, validating, be applied, how local forest users and communities and refining the systems. It is vital that the systems should be involved, and what measures should be put be empirically grounded and tested against field in place to ensure they are conducted safely. conditions. Other stakeholders should also be involved in this process, for example, research institutes such as In more fire-prone areas, the state forest departments IISC, FRI, and the Indian Meteorological Department. should consider planting fire-hardy species in planning new plantation areas, as was recommended So far, the development of FSI’s “pre-warning alert in the National Forestry Action Programme of 1999 system” has focused mainly on the data and methods (MoEF 1999). identifying areas of high fire danger. Less attention has been paid to communicating fire danger with the More systematic use of silvicultural practices for fire public and making the system actionable. Indeed, prevention is also needed in forested areas outside many of the field officers interviewed by the World those managed by the forest department. The risk of Bank team were unsure of the purpose or the utility unwanted and out-of-control fires in state-managed of the pre-warning alerts. Here, too, the state forest forests is increased by the lack of fire management on departments can play an invaluable role in conducting adjacent lands. In each of the 11 states surveyed, forest public outreach. officers said that other public and private entities are responsible for managing forests in their area. In most The need for the states to become more involved needs cases, these forests are not covered by working plans or to be emphasized. It is one thing to develop useful similar planning documents. The forest department indices but quite another to get the message through does not have formal jurisdiction over these lands. to the community. Unless all the states are onboard Though information on forest fire occurrence is and actively participate, uptake by communities is lacking for these non-department lands, data from likely to be limited. There is therefore a need for FSI Uttarakhand suggests that the risk of fire is significant, to establish a strong outreach and feedback system at least in some areas. Responsibility and management with the SFDs on utility and value of FDRS developed arrangements for silvicultural operations on non- by FSI, which in turn will support SFDs to bring in department lands will vary depending on their status. communities and other stakeholders for a better and Strengthening fire prevention on community-held effective public outreach. The states should also be lands is discussed in section 4.6 below. Coordination involved in the process of testing and refining the with other public agencies in managing forest lands is FDRS. dealt with in section 4.7. 4.2.2 More systematic use of silvicultural practices 4.2.3 Working with communities to modify how for fire prevention fire is used and prevent unwanted fires More systematic use of silvicultural practices such The total exclusion of fires from forests is not an as selective thinning, pruning, and early-season attainable or desirable goal for FFPM. Some fires controlled burning should be applied to reduce fuel can be beneficial, both from an ecological and social loads as part of the management norms for plantation point of view. There exists a fundamental tension areas. Limited silvicultural operations and controlled between the total prohibition on fire under current burning may also benefit other state-managed forests, law in India and the reality on the ground, as fire where the total exclusion of fire is still practiced and continues to be used as a landscape management tool where fire prevention measures have been mostly by communities of forest users across the country. A limited to the maintenance of fire lines. The need for more effective policy for FFPM may begin with the silvicultural interventions is especially apparent in the recognition that people will continue to use fire, that hill states in pine forests above 1,000 m, where fire lines some fire is desired, and that the goal of FFPM should have overgrown and where the rapid accumulation of be to minimize the ecological, social, and economic fuels presents a risk for more intense fires. As noted in impacts of fire while ensuring that the benefits reaped 123 Strengthening Forest Fire Management in India from fire may continue. From this starting point, fire done to systematically evaluate these data to see how managers may then work with communities to ensure the accuracy or utility of fire alerts may be improved. that fire is used responsibly in a way that promotes Existing algorithms and methods that are used to forest health, while seeking to avoid damaging and generate fire alerts from the satellite data cannot be out-of-control fires. modified or improved without field verification. Community institutions for forest management Third, integration between the FSI’s alert system such as the Joint Forest Management Committees, and the state- or regionally-developed systems can be collectives for NTFP harvesting, the Van Panchayat improved. The utility of the systems can be leveraged in Uttarakhand, and village governing bodies such by sharing user databases and ensuring greater as the Gram Panchayat offer a mechanism for the consistency with how the alerts are generated (e.g., forest department to engage with forest users. For with quality screening criteria or the definition of this interaction to be effective, however, community boundaries for areas of management concern). institutions will need to be strengthened. Fourth, the use of direct readout satellite data from the MODIS and VIIRS instruments would help reduce lag 4.3 FIRE DETECTION times and outages in the FSI active fire alert system. With additional training and support, FSI would have 4.3.1 Improving satellite-based alert systems the capability to become a direct readout site.94 Satellite-based remote sensing forms an indispensable part of forest fire detection in India, but the 4.3.2 Improving ground-based detection systems effectiveness of the fires alert systems developed by FSI and the states can be strengthened further. Even with the advent of new remote sensing technologies, ground-based detection will continue First, the digitization of management boundaries by to be essential. Greater funding for construction of the state forest departments should be completed. watchtowers and crew stations and for frontline officers Boundaries have been digitized down to the lowest and seasonal firewatchers to spot fires is needed, as administrative level (beats) in 17 states so far.93 Mapping most of the areas surveyed reported shortfalls and these boundaries is crucial for determining what fire field officers reported frequent delays in the payments locations to report and who should be alerted. This in to seasonal firewatchers. turn will require maintaining up to date rosters and contact information of field staff and exchanging this The utility of ground-based detection can be enhanced information with FSI. by integrating it with the satellite-based alert systems. Madhya Pradesh is already moving in this direction Second, ground verification data on satellite-based by experimenting with a new mobile app that would alerts should be collected by field staff, shared with allow field staff to send validation reports for fire FSI by the state forest departments, and analyzed. alerts and to submit reports for fires that were not A handful of states have built online platforms for detected by the satellite instruments. This would field-level personnel to submit ground verification allow the forest department to track whether fires are reports for fire alerts in their areas. Also, though FSI observed first by field staff or by satellite, the location has established a way for users registered with its fire and time of ignitions or detections, the time required alert system to provide feedback, it has only received for field crews to arrive on-site to verify alerts, and reports from a few states so far. In those states where other valuable information that can assist with fire ground verification data are collected, little has been management. E. Vikram, FSI, written comments to World Bank on draft report, February 2018. 93. See US NASA, “Direct Readout Laboratory,” https://directreadout.sci.gsfc.nasa.gov/?id=home. 94. Strengthening Forest Fire Management in India   124 4.3.3 Extending systems for detection to non-forest to adopt the most effective suppression technique at department lands their disposal and know when retreat is necessary. The type of training provided to firefighters should be As revealed in the India State of Forest Report 2017, the tailored according to their level of responsibility and area of tree cover outside forests expanded by 1,243 role in the command structure in responding to fires. km2 between 2015 and 2017 (FSI 2018). Satellite- Because close-in attack exposes firefighters to dangers based detection systems that have evolved to monitor from quickly changing fire behavior, crew leaders and active fires on department-managed lands should commanding officers must always be aware of and be be expanded to include other forest areas beyond able to react quickly to changing conditions. Thus, department jurisdiction. FSI already does this to some different levels of training are needed for crew leaders extent by providing active fire alerts for all forest- and fire bosses versus the crew members under their covered areas at the district level in states that have command.95 not digitized the management boundaries of forest department-controlled lands. Extending systems A decentralized and field-based “train the trainer” for satellite detection to non-department lands will system may be most appropriate for India. At the eventually require digitizing boundaries of community central level, a modern and standardized training and privately-held forests. The need to include non- curriculum should be developed by MoEFCC and department lands is especially acute in the Northeast, Directorate of Forest Education (DFE) together with where the forest department only manages a small the state forest departments. The training should form percentage of the overall forest estate. part of the curriculum of all state forestry training centers that train frontline staff such as forest guards. DFE is the agency responsible for coordinating 4.4 FIRE SUPPRESSION such training, and for providing refresher courses for field staff from time to time. By involving the 4.4.1 Training for field staff, firewatchers, and states, the curriculum should capture and utilize community firefighters local knowledge in developing a suite of fire training manuals, pitched at different levels. Other agencies In general, forest fire suppression relies heavily on involved in fire response, including NDMA, NDRF, dry firefighting techniques. Dry techniques include and the state disaster management authorities may be directly beating out the fire with hand tools to involved in a consultative role. The development of smother the flames (for very low-intensity fires) or a standardized curriculum is important for ensuring by separating the fuel in advance of the active fire, smooth operations across departments when large either by natural breaks in the fuel or by deliberately fires affect neighboring divisions or states. creating mineral earth breaks devoid of fuel. People on the ground are the key to effective fire suppression 4.4.2 Provision of equipment to firefighters using dry techniques. In spite of the availability of hi- tech equipment globally, the principal need is always Only a handful of field officers surveyed by the World to have a competent, trained, and equipped workforce Bank team agreed that firefighting equipment is on the ground, ready to respond and take immediate adequate and sufficiently available in their area. Many action. pointed to the need for basic safety equipment and clothing. Some called for additional hand tools and The need for greater training was almost unanimously transport vehicles for field staff. mentioned among the officers surveyed and interviewed. Training should be provided to field In principle, the focus for equipment should be on officers, seasonal firewatchers, and community providing hand tools, small motorized equipment, volunteers involved in firefighting. All these firefighters and protective clothing. Handtools should be should understand the basic principles of fire behavior robust but light enough to avoid overly fatiguing Ross Smith, “Safety and Equipment,” presentation and discussion at the workshop on Forest Fire Prevention and Management organized 95. by MoEFCC and the World Bank in New Delhi, November 2017. 125 Strengthening Forest Fire Management in India ground crews. Motorized equipment might include 4.4.4 Eventual implementation of Incident leaf blowers, chainsaws, off-road quad bikes, and Response System (IRS) motorbikes. Protective clothing is essential for forest firefighters and should be made of low-flammability India may eventually consider instituting standards material such as tight weave cotton drill. Clothing for multi-agency and cross-border forest fire response should be loose fitting with underarm and side pocket such as those that have been popularized in North slits, loose fitting trouser and sleeve cuffs to allow easy America and by the International Organization ingress and egress of airflow. Sturdy boots made from for Standardization (ISO).96 In the United States, leather or fire-resistant material, safety helmets, and the standard Incident Command System (ICS) leather work gloves are also recommended. Protective first emerged in response to large wildfires at the clothing is needed for field staff, seasonal firewatchers, wilderness-urban interface in California in the 1970s. and community firefighters. Odisha was the only The early development of the ICS was led by the U.S. state visited in which the forest department provided Forest Service and now forms a central part of the protective clothing to community firefighters. national system for multi-agency response to all kinds of disasters and emergencies (Stambler and Barbara Rather than prescribe what tools ought to be applied 2011). The ICS has since been implemented in other in different areas of India, a better approach would be to provide firefighters with a range of tools and countries for responding to forest fires, including seek their views about which tools have useful fire in Mexico (box 4.3). Implementing ICS in Mexico management roles in different geographical areas and was a multi-year process and required considerable different fuel types. The forest fire cells in the state investment.97 forest departments should take the lead in this process of identifying and providing firefighting equipment In 2010, the NDMA in India issued the National suitable to local needs, with MoEFCC or a MoEFCC- Disaster Management Guidelines on Incident delegated entity such as FRI playing a supportive, Response System (IRS). With the IRS framework as guiding role. It is only by trialing the various tools a base, NDMA has also released SOPs for managing available that their capabilities can be fully appreciated various disasters and guidelines for conducting by local firefighters and their use adopted. There is no mock drills. By organizing large-scale mock drills in “silver bullet” for forest firefighting – what will work Uttarakhand, such as those in April 2017, NDMA is best under forest fire conditions in India is what the beginning to hone IRS as a mechanism for forest fire local people develop or elect to use. response. 4.4.3 Development of a national SOP for forest fire IRS is useful when there are multiple suppression response agencies involved that have capability to undertake effective fire management operations (mostly, but Only a few states have developed SOPs or manuals not restricted to suppression), that they are on the on standardized forest fire response systems. Odisha same wavelength and use common systems and is a good example. MoEFCC may consider the terminologies. It is more urgent for India to effectively development of a national SOP for fire response, resource fire management through the provision of adapted to local conditions in consultation with the state forest department and communicated by the state tools and equipment and by training forest officials forest departments to field staff. The development of and communities in fire suppression practices. It is a standard training curriculum (4.4.1 above) would critical to get these and other basic building blocks in further this effort. The SOP may include procedures place (e.g., state-level SOPs for forest fire management) for the management of fires across state boundaries before adopting a full-blown IRS. and jurisdictions of multiple agencies. See ISO 22320:2011 (en), https://www.iso.org/obp/ui#iso:std:iso:22320:ed-1:v1:en. 96. 97. Alfredo Nolasco Morales, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017. Strengthening Forest Fire Management in India   126 Box 4.3: International Cooperation on Forest Fire Management in Mexico As part of its new approach to forest fire prevention and management described in Box (4.1), Mexico has been strengthening international cooperation related to forest fires since 2013. Mexico currently has agreements with Colombia, the United States of America (USA), Canada and Chile on a range of themes including research, technical exchanges, Incident Command System and Incident Management Teams, as well as support in areas of mutual assistance and natural disasters. Mexico has implemented an Incident Command System (ICS) for fire suppression, although this took significant time and training to institute. In particular, Mexico’s cooperation with the USA includes training courses, technical exchanges and research. In addition to research, Mexico and Canada also cooperate on forest fires in terms of international deployment. Chile is another country with which Mexico has an international deployment arrangement, in addition to cooperation in other areas, such as technical exchanges. Furthermore, Mexico cooperates with other countries in Latin America in areas such as training courses and technical exchanges. The FAO, in its report titled “Fire management: global assessment 2006” notes that the borders “between Mexico and the United States are covered by international agreements that authorize the exchange of firefighters and provide for assistance on fires that cross international boundaries.” Moreover, it is noted that national-level agreements and also local agreements exist between adjoining jurisdictions to address local needs. In addition, it is indicated that Mexico, the United States and Canada are able to work together on fire suppression because they have all adopted the ICS. The International Wild Land Fire Summit, held at Sydney in October 2003 led to the formulation of guiding principles for international cooperation with regard to forest fires (Satendra and Kaushik, 2014). India should consider enhancing inter-state and international cooperation on forest fires to improve fire management. A recent experience of inter-state cooperation between Karnataka and Kerala is a case in point - when Bandipur Tiger Reserve in Karnataka was being ravaged by fire in February 2017, the timely intervention of forest personnel from the bordering Wayanad Wildlife Sanctuary in Kerala is said to have been very helpful in controlling it (The Times of India, 2017). Source: Morales (2017); Alfredo Nolasco Morales, National Forestry Commission, Mexico, presentation and discussion at the workshop on Forest Fire Prevention and Management organized by MoEFCC and the World Bank in New Delhi, November 2017; Satendra and Kaushik (2014); Food and Agriculture Organization of the United Nations (2007); The Times of India (2017) 4.4.5 Fire response beyond forest department- 4.5 POST-FIRE MANAGEMENT managed forests 4.5.1 Training, resources, and incentives for Provision of training, equipment, and coordination accurate and complete reporting of forest fires should extend beyond state-managed forests to community institutions in regions such as the Post-fire reporting by field staff is hindered by Northeast, where communities are responsible insufficient resources, difficult terrain, and a lack of for managing most of the forest estate. Because connectivity in more remote areas. Underreporting of financing is already a perennial challenge for FFPM the area affected and damages caused by forest fires is in department-managed forests, extending additional common. Reasons for underreporting include human support to community institutions may require error (burnt area is almost always assessed by ocular additional funding sources outside the MoEFCC inspection only); a lack of standard reporting protocols budget. (officers may exclude areas burnt by ground crews 127 Strengthening Forest Fire Management in India Box 4.4: Indicators for Monitoring Progress and Measuring Results on Forest Fires Standard protocols for reporting the occurrence, burnt area, and damages from forest fires along with other data—and the collection of this information in a central database—will enable MoEFCC and the state forest departments to better monitor progress toward improving FFPM. Measured results are needed to justify the allocation of greater financial resources for FFPM and will in turn require the development of appropriate indicators for monitoring and evaluation of forest fire programs. As an illustrative example, the table below presents indicators used by the World Bank in Kazakhstan to measure the impacts of a forest protection project on fires. It is important to highlight at the outset that while indicators such as a reduction in the number of forest fires or burnt area may be used to track the impacts of certain interventions, these are also influenced by several other factors such as weather conditions, which cannot be controlled. TABLE B4.1: INTERVENTIONS FOR IMPROVING FFPM, IMPACTS, AND POTENTIAL INDICATORS BASED ON THE WORLD BANK KAZAKHSTAN FOREST PROTECTION AND REFORESTATION PROJECT Interventions Impacts Indicators Investments in infrastructure • This has improved the effectiveness • Land degradation (e.g., fire stations and of fire management in about 650,000 (specifically, deterioration or fire breaks), equipment, hectares of Irtysh pine forest, and area with lack of tree cover or staff training and public started a reversal of fire degradation other vegetative cover) awareness campaigns trends • Average number of fires in • During 2008-13, the number of fires project area (versus other compared to the five years before the comparable areas) project decreased by 20 percent, and the • Share of human-caused fires share of human-caused fires dropped from 60 percent to 35 percent (Arhipov • Area under improved fire and Arhipov 2015)98 management Investment in fire detection • Detection times are now quicker (2- • Average area of fires upon and information systems 25 minutes faster) leading to shorter detection based on automated smoke response times, and a decrease in the • Average response time after detection through optic average area of a fire incident (from detection sensors and surveillance 23.7 hectares during 2003-11 to 1.67 hectares in 2012-13 after installation). • Area under monitoring/ Moreover, larger areas can be surveillance monitored than is possible through human observation. Source: Authors, based on World Bank (2015) Other potential indicators to track improvements in FFPM may include: • Percentage of fires contained within 24 hours following detection • Proportion of large fires above a certain size out of the total number of forest fires detected • Number of trained and equipped field personnel deployed • Number of field staff receiving fire alerts and providing field verification reports • Reach and acceptance of fire prevention messages among the local community in comparison to the baseline (based on survey results) • Forest regeneration in the target area versus other comparable areas Where FFPM interventions include the provision of incentives to local communities, indicators should also measure outcomes related to forest livelihoods. In this case, it will be vital to collect good-quality baseline data to compare outcomes over time. Arhipov V.A., Arhipov, E.V., (2015) The Study of Forest Fires in Ribbon-Pine Forests of Priityshie, in Forest and Wildlife Committee 98. (2015). Strengthening Forest Fire Management in India   128 in setting counterfires as part of suppression); and collection of data on fires will be essential to measuring institutional disincentives (field officers who report the results of FFPM interventions (box 4.4). large fires may create additional work for themselves and their superiors in filing and prosecuting a Information technologies may further assist in forest offense, and the department may receive less improving field reporting. Madhya Pradesh, for financing). example, is currently exploring the development of a mobile app that could be used by field staff to send Merely strengthening vertical oversight by superior feedback on satellite fire alerts. Such an app could also officers within the department does not address be used to collect and report other information, such fundamental problem of incentives. First, to encourage as tracing burn scars. more accurate and complete reporting, department financing should be delinked from fire damages, 4.5.2 Common classification scheme for suspected and the reported incidence of fire within an officer’s or probable causes of fire jurisdiction should not be tied to the determination of job performance, monetary compensation, or More than just an administrative task, the purpose career advancement. Exempting forest fires from the of investigating the causes of forest fires is to gather reporting requirements to the Accountant Generals information to assist with planning and management Office was one of the key recommendations to be for FFPM. Currently, investigation of the causes of issued by MoEFCC after the National Workshop forest fires is limited. The availability of personnel on Forest Fires in 2007, though it has yet to be to conduct such investigations is a major constraint, implemented. Second, field officers should be held especially during the peak fire season. About one- accountable for the fulfillment of required prevention quarter of surveyed officers said the causes of forest and control activities, and, as recommended by the fires were investigated only partly or not at all in their National Forest Commission (2006), the performance area. of fire control duties should be included in the annual evaluations of field staff. However, staff should not be More useful information on the causes of fire could punished for the occurrence and reporting of fires in be gathered for planning and management purposes their jurisdiction unless such damaging or unwanted if field officers could report the probable or suspected fires are caused by negligence or poor management on cause of fire using a general classification scheme. The their part. Fires are a semi-natural occurrence and not need for “a uniform classification of forest fires by completely within the control of field personnel. Also, types and causes…evolved and adopted by the States” the complete exclusion of fires from forests is not the was also recognized by the National Commission aim of the department. Field-level personnel should for Agriculture in 1976 (NCA 1976: 45.2.3). Many be rewarded for providing accurate and thorough countries and regions have developed such schemes, data on fires, not punished. including Australia, Canada, the EU, New Zealand, Russia, and the United States. A common classification Standardized protocols and procedures are needed scheme for India would need to recognize the variety of to facilitate the reporting of the area affected by fire circumstances and uses of fire in the different regions and should be developed by MoEFCC or a delegated of the country. As a starting point for discussion, the research entity under ICFRE. Field reports should be classification scheme used in chapter 1 to analyze the cross-checked using GPS or remotely-sensed imagery common causes of fire cited by officers in the survey of burn scars. Additional resources may be required of 11 states is presented in Annex 4. Using such a to support the use of GPS units or development of classification scheme, a field officer could report what applications using GPS-enabled mobile phones to map the general cause of fire is and the degree of certainty the perimeters of burnt areas. Protocols for estimating with which the cause is known, ranging on a scale from and mapping fire-affected areas may be integrated with highly uncertain to certain (using four categories for guidelines for classifying the suspected or probable cause degree of certainty – certain, highly likely, uncertain, of fire (4.5.2 below). The accurate and standardized no idea/unknown, for example). To aid officers in 129 Strengthening Forest Fire Management in India making such a determination, MoEFCC would need developed by the UN Economic Commission for Latin to develop standard protocols and training materials. America and the Caribbean (UN ECLAC 2014). The Uniformity across states in investigating and reporting UN ECLAC methodology follows a damage-plus- the suspected causes of fire is essential to allow cross- loss approach. Damages represent the destruction of state comparisons and the aggregation of statistics on physical assets; losses are a flow-based concept and fire incidents. represent changes in economic activity following a fire. The valuation of losses due to forest fires in 4.5.3 Strengthening the assessment of the economic Indonesia in 2015 (box 2.10) provides an example of impacts of fire how the UN ECLAC methodology may be applied to large forest fires. Without good data on the impacts and costs of fires, it At the strategic level, economic valuation may also be is difficult to convince political leaders and the public used in conducting periodic reviews to evaluate FFPM that forest fires are a priority. Good data on impacts programs by the states and MoEFCC. In the United and costs will also allow MoEFCC to track progress States, for example, the Forest Service is required to over time in reducing damages from fires if policies perform such evaluations regularly to support budget and programs for FFPM are successful. requests for fire management programs to the U.S. Congress.99 A strategic evaluation may be used to As a starting point, MoEFCC should develop a determine the efficient level of spending on FFPM, to guidance note and standard set of methods that support a request for additional financial resources if could be used for performing field assessments a shortfall is found, and to weigh the costs and benefits of damages due to forest fires. The development of investments in FFPM. of such a toolkit for impact assessment could be done through a facilitative process involving the 4.5.4 Silvicultural practices for restoring and states (some of which already have standards) and rehabilitating fire-degraded forests is an ideal entry point for involving the research community in India (see 4.7.3). The process could be High-ranking officers surveyed and interviewed in led by the Indian Council of Forestry Research and the state forest departments noted that restoration Education (ICFRE) or one of the research institutes and rehabilitation of fire-affected forests is limited. under ICFRE. The guidance note should also clarify Although damages from individual surface fires are when assessments should be performed and for what usually minimal, the occurrence of repeated fires purposes. over short intervals may lead to degradation. Again, establishing protocols for post-fire restoration activities More detailed assessments of economic impacts may as part of the Working Plan Code is one approach to standardize it. be appropriate for large fires or for fires affecting areas of significant natural or cultural value (e.g., a national The identification of severely-affected areas requiring park). The need to conduct an impact assessment restoration and rehabilitation should be integrated into might also be triggered by the formal declaration of forest working plans at the division level. At the state a disaster in the event of a fire. The National Forest and national level, FSI may assist in identifying areas Commission of 2006, for example, suggested that all of highest priority as part of its regular nationwide fires that burn an area larger than 20 km2 should be burn scar assessment. Funding from CAMPA and declared a state disaster. FPM Scheme may support ecological restoration and rehabilitation activities. These areas may be further For the assessment of economic impacts due to forest monitored for progress through the e-Green Watch fire disasters, MoEFCC may draw on the methodology mechanism.100 See U.S. Forest Service, “Budget & Performance,” https://www.fs.fed.us/about-agency/budget-performance. 99. 100. See MoEFCC and NIC for Transparent and Responsive Governance, “e-Green Watch,” http://egreenwatch.nic.in/Portal.aspx Strengthening Forest Fire Management in India   130 4.6 ENGAGING WITH COMMUNITIES firewatchers, or protecting an area from fire. However, the JFMCs have also drawn criticism for being top- 4.6.1 Institutional support for communities as down, with decision-making powers and management managers of their forests authority concentrated in the forest department, and increasingly low levels of investment and participation Aside from strained department resources, challenges on the part of communities. If effective community with public engagement were cited by surveyed forest involvement is to be garnered, it is essential to work officers as the biggest obstacle to preventing forest with communities and give them a voice in the fires. Although there is no universal formula for how decision-making process. If they have that, they will the forest department may improve its outreach on more likely feel included and be an effective part of FFPM, at the root level, the success of such outreach will the partnership. Micro-plans and working schemes depend in large part on the nature of the partnership providing for FFPM have been implemented in fewer with communities, which can range from awareness than half the JFMCs, and growth in the number of raising to co-management based on an understanding JFMCs has stalled (Bhattacharya et al. 2010). of the needs of the community. To reinvigorate the JFMCs, the state forest departments In many parts of the country, the Joint Forest will need to provide more meaningful incentives and Management Committees (JFMCs) will continue support for undertaking FFPM activities (see 4.6.2), to function as the primary entity for FFPM at the grassroots level in forest-dependent communities. and to shift the focus of their relationship with the The forest department may engage with the JFMCs JFMCs toward a more cooperative engagement on the in a variety of ways, for example by providing managed use of fire. For example, engagement could payments to the JFMCs for clearing fire lines, hiring be based on allowing space for traditional practices Box 4.5: Testing Community Incentives for Preventing Forest Fires in Indonesia In support of the Government of Indonesia’s “Grand Design” for forest fire prevention announced by the Coordinating Minister for the Economy in December 2017, researchers are testing new models for providing economic incentives to local villages to limit their use of fire for clearing and other purposes. The experiment, led by Stanford’s Center on Food Security and the Environment together with TNP2K (Indonesia’s Agency on Poverty Reduction within the Vice President’s Office) and (Bogor-based) Daemeter Consulting Company, will be conducted from 2018 to 2019. A total of 400 comparable villages will be drawn randomly from fire-prone areas of three provinces: West Kalimantan, Riau, and East Kalimantan. Of these, half of the villages will be used as the baseline for measuring what happens if there is no intervention by project personnel; 100 villages will receive instruction on fire prevention at the village level, plus Rp 25 million (USD 1,900) in grant funds at the beginning of experiment; and 100 villages receiving instruction, the Rp 25 million, AND a conditional payment of Rp 150 million (USD 11,500) at the end of the year if the village is successful in eliminating fires. The project brings together environmental and economic research that improves the wellbeing of villagers while at the same time protecting Indonesia’s natural resources. The central focus is on independent smallholders, who are often quite poor and often outside the developmental activities of both private companies and government programs. Rigorous empirical evidence gathered by the project will feed into the government’s “Grand Design” policy initiative and provide empirical evidence for the incentive scheme that the government eventually plans to launch as part of this initiative. Source: Project Summary, “Fire Prevention in Indonesia,” February 28, 2018, Stanford Center on Food Security and the Environment, Stanford, CA. 131 Strengthening Forest Fire Management in India involving controlled burning to be planned and forest tenure, resource rights, and sharing revenues executed by communities in fire-adapted forests (with from commercial products such as teak, sal, and the supervision of the forest department) to ensure bamboo where allowable. the provision of fire-associated environmental goods and services for local forest users while also mitigating Regardless of the form that incentives take, their the negative impacts of overly-frequent fires and the provision should adhere to several guiding principles. dangers of unsupervised burning or burning when First, the roles and responsibilities of the state forest fire danger is too high. department and community institutions for FFPM should be clearly defined. Second, the provision of In parts of the country where the JFMCs have not the incentive should be directly linked to fulfillment taken root, MoEFCC and the state forest departments of the stated roles and responsibilities for FFPM. will need to identify other community institutions that Unconditional investments by the forest department, can take on a greater role as the focal point for FFPM for example in building roads or schools, in fire- and invest in strengthening them. The Gram Sabha affected areas may not spur communities to cooperate or Gram Panchayat may offer another avenue for on FFPM unless the link between the investment and the forest department to refocus its engagement on implementation of FFPM is clear. Third, especially FFPM, particularly in Scheduled Areas inhabited by in the case of monetary payments, the provision of tribes and other traditional forest dwellers. In these incentives should be regular and predictable. areas, the Forest Rights Act (FRA)101 grants the Gram Sabha overall authority to administer customary forest States should be encouraged to experiment with new lands and the minor forest produce harvested on those and creative ways of providing incentives. New research lands. In Meghalaya, engaging with communities from Indonesia may serve as a useful reference for through Village Fire Control Committees (VFCCs) has MoEFCC and the state forest departments in designing provided a good example how the forest department, scientifically-based incentive schemes (box 4.5). with little ownership of forest resources, can still play a positive role in bringing communities together to MoEFCC may explore whether payments for protect community-owned forests from unwanted ecosystem services to communities via REDD+ can be fire. Meghalaya is also planning to orient communities scaled up to other parts of the country to incentivize toward natural resource management activities protecting forests from damaging fires. The Khasi (including FFPM) through utilizing MNREGA funds. Hills Community REDD+ Project in the uplands of Meghalaya, India’s first community-based REDD+ 4.6.2 Incentivizing communities initiative (case study 1 in Annex 3), offers a good initial case study. How the experience in the Khasi Hills can Forest officers who were interviewed and surveyed be replicated to other parts of the country where the pointed to the need for greater incentives as the most landscape of community institutions and the causes of important way for the forest department to increase forest fires are much different remains a question. the effectiveness of its engagement with communities on FFPM. Many noted that the department already 4.6.3 Public education and awareness raising provides incentives to communities in their area. These incentives have taken a variety of forms, Public education has been one of the most common including wage labor, small cash rewards, and public forms of engagement by the state forest departments recognition for outstanding performance. However, with local communities. During the fire season, forest in many parts of the country, current incentives have departments have distributed banners, posters, not been enough to mobilize communities as partners handbills, and stickers; they have aired public service in FFPM. Stronger incentives may include securing announcements on local television and radio; and 101. The Scheduled Tribes and Other Traditional Forest Dweller’s (Recognition of Forest Rights) Act 2006, also commonly referred to as the Forest Rights Act (FRA). Strengthening Forest Fire Management in India   132 they have organized puppet shows, street rallies, lands. Other agencies to involve include the Ministry padayatras, oath ceremonies, school programs, of Agriculture (MoA), given that escaped agricultural and signature campaigns, working with the Gram burning is a prevalent cause of forest fires, as well as Panchayats and other community institutions. Such the Ministry of Rural Development (MoRD), given its outreach efforts should continue. To change attitudes substantial budget and program under the Mahatma toward forest fires over the long run, emphasis Gandhi National Rural Employment Guarantee Act should be placed on integrating knowledge about (MNREGA) for soil and water conservation works. forest fires and their environmental impacts as part of environmental education in school curricula. A coordination mechanism should be put in place at the national and state level to clarify responsibility and organize emergency response to large forest fires. 4.7 COORDINATION WITH OTHER Currently, it is unclear at what level of fire and in AGENCIES AND ENTITIES what areas the NDMA, the NDRF, and the state-level authorities should be expected to respond and assist 4.7.1 Clarifying the roles and responsibilities for with fire suppression. The NDRF and SDRF joined the forest department and other agencies in the response to the severe fires in Uttarakhand in 2016 but have not been active elsewhere. At the state Forest management agencies globally have generally and local level, the involvement of other agencies in been the most successful agencies in dealing with fire response has also been primarily on an ad hoc land and forest fires. MoEFCC and the state forest basis. A coordination mechanism at the national level departments should continue to play the leading role could take the form of a core group led by MoEFCC in implementing FFPM in India. While FFPM will with representatives from the NDMA, Ministry continue to be the remit of the forest department, of Home Affairs (MHA), Ministry of Tribal Affairs other agencies and entities should be involved in (MoTA), Ministry of Road Transport and Highways both the prevention and suppression of fires. These (MoRTH), Ministry of Railways (MoR), the military, other agencies, including the disaster management and other concerned agencies. At the state level, such authorities, disaster response force, police, home a group would be led by the state forest department guards, military and paramilitary are often called upon with representatives from the line agencies, state to assist the state forest departments in responding to police, state fire service, home guards, military, and especially large or damaging forest fires. paramilitary. The core group would be activated upon declaration of a fire disaster at the state or national At the national level, a coordinating policy is needed level and would coordinate with the district authorities to clarify the role and responsibility of MoEFCC in implementing FFPM on forest lands administered to organize the deployment of people and assets on by other agencies. These other agencies include, for the ground in response. example, the Department of Revenue, which manages about 13 percent of the forested area of Uttarakhand. At the more local level, forest fires should be written The joint management of these areas should be formally into the district disaster management plans for areas established through an inter-agency agreement, where forest fires are a perennial risk and where with additional funding to MoEFCC if MoEFCC vegetation types, weather patterns, or local topography continues to fulfill its primary role in carrying out fire have created conditions for potentially severe fire prevention and suppression activities on these lands. behavior. Districts with large populations or important Similar agreements may be needed with the Ministry infrastructural assets sited in fire-prone areas should of Tribal Affairs and the Administrative District also include fire in their disaster management plans. Councils in the Northeast, as field officers surveyed Technical support on best practices for fire disaster and interviewed noted that the forest department is planning may be provided by MoEFCC in consultation often called to respond to fires on community-held with NDMA and the state-level authorities. 133 Strengthening Forest Fire Management in India 4.7.2 Joint trainings organized to facilitate and state governments and other entities, may serve as coordination during a fire event a good example of an institutional model for bringing together public land managers and members of the In April 2017, the NDMA conducted a first-of-its-kind research community.102 state-level mock exercise on forest fire in Uttarakhand in order to assess the efficacy of integrating the 4.8 FOREST FIRE SCIENCE, DATA, preparedness and response mechanisms of the SFD with those of the district administration. The KNOWLEDGE SHARING, AND exercise was carried out across all 13 districts and was TRAINING conducted in collaboration with the state government and agencies including fire, forest, army, health, 4.8.1 Creation of a national forest fire information police, NDRF, SDRF and civil defense. Such exercises database should be replicated in other areas where there is the potential for severe fires, and gradually extended to Currently, nationwide information on forest fires interstate exercises. in India is limited to satellite-based remote sensing data. The creation of a common classification Members of JFMCs, Van Panchayat, and other scheme for the causes of fire, standard reporting community institutions interviewed also expressed protocols, and standard methods for assessing the need for field-based training with the forest burnt area would facilitate the creation of a national department on fire suppression. Forest officers forest fire information database incorporating field- surveyed by the World Bank were unanimous in reported data. The database should also capture citing the need for trainings for seasonal firewatchers. information on fire lines, controlled burning, watch With the development of new training materials towers, firefighting assets (and their locations), and and decentralized training via a “train the trainer” communications infrastructure. Such a database model (4.4.1 above), trainings should be extended would be instrumental for assessing longer-term to community members and fire responders in other trends across states and regions and for planning fire concerned agencies. prevention and response. As noted by the National Forest Commission (2006), creating a database would 4.7.3 Involving other entities, particularly the include establishing a mechanism for ensuring data research community quality and cross-checking figures reported by local field staff. Field-level officers in the state-level forest The still-limited knowledge about fire ecology in departments should have access to the database as different forest types and climates, the longer-term well. FSI is a good agency to develop and maintain impacts of fires on forest degradation in India, such a database. and methods for assessing such impacts signals the need for greater involvement of the country’s 4.8.2 Definition of a national forest fire research research community on FFPM. This would include agenda public institutes and agencies, universities, and NGOs. The definition of a national research agenda Research priorities for forest fires should be for forest fires (see 4.8.2 below) and provision of determined by MoEFCC in consultation with the state funding opportunities for scientific research would forest departments and members of India’s research be instrumental in bringing these entities together. community in creating a national research agenda Australia’s Bushfire and Natural Hazards Cooperative for forest fires. The National Forest Research Plan of Research Centre (CRC), funded by the national 2000, crafted by the ICFRE, may serve as a starting government with matched support from territorial point and should be updated. The plan should reflect 102. See Bushfire and Natural Hazards Cooperative Research Centre, http://www.bnhcrc.com.au/. Strengthening Forest Fire Management in India   134 the priorities contained in the national policy or action could also serve as a forum to collect information plan on FFPM. in a consistent and standardized manner (e.g., fire statistics). Although MoEFCC already organizes The National Research Priorities to 2020 and Beyond annual meetings of PCCFs from the states, there is a set by Australia and New Zealand’s Forest Fire need for a formalized knowledge-sharing mechanism Management Group provides a good example of how or forum that is expressly focused on FFPM. such a priority-setting exercise may be conducted in collaboration between forest managers, the scientific 4.8.4 Creating a Center of Excellence for FFPM community, and traditional forest users (FFMG 2013). The goal of such an exercise is to identify crucial The creation of a Center of Excellence should advance policy-relevant research with a focus on FFPM. knowledge gaps and priorities for forest managers Such a center should bring together other agencies and to ensure they are supported by policy-relevant and institutes with a stake in FFPM and disaster scientific research. management, including FSI and NDMA. ICFRE, with data and technology support from FSI, could develop 4.8.3 Formalized mechanism for sharing knowledge such a center of excellence. In fact, the Government across states of India is considering setting up a National lnstitute of Forest Fire Management with satellite centers in India could however benefit from the development of different parts of the country to bring the latest forest a mechanism to allow useful exchange between states. fire fighting technologies to lndia through proper There is a real need for a suitable forum where state research, training of personnel, and technology representatives can regularly meet and swap ideas and transfer on a long-term basis. information. Presently, each state forest department seems to operate in isolation from other states. There are excellent initiatives developed by individual states 4.9 SUMMARY AND PRIORITIZATION that could be easily transferred to and adopted by other states. Examples include the use of leaf blowers Table 4.1 below presents a summary of for creation and maintenance of fire lines as well as recommendations, which have been ranked according direct attack on fires, schemes to incentivize villages to priority. Many of these recommendations will to display greater care with fire use, smarter use of involve a multi-year process that should start now. At funding models (e.g., CAMPA funds), eliminating or the central level, immediate priorities for MoEFCC controlling fire use for harvesting NTFPs (e.g. tendu include the updating of FFPM guidelines to provide slashing in lieu of burning), training of villagers in basic greater clarity and consistency, initiating the process firefighting techniques, and the provision of safety for developing a national action plan, and establishing clothing. Useful and productive activities developed an inter-agency coordination mechanism for forest in a state are therefore not shared for the benefit of fire response. At the state level, immediate priorities other states. This is a critical shortfall in the system include ensuring that adequate field-level staffing and because those actions that are effective and which have financial resources are in place to carry out forest fire been developed in response to local conditions meet prevention activities as per the forest working plans, as the necessary cultural, social and financial constraints. well as providing field staff with basic safety equipment They are thus far more likely to be acceptable, easier to and hand tools. Implementing these recommendations transfer and more effective than a solution imported will require increased, dedicated financing for FFPM. from elsewhere. As noted in 4.1.5, emphasis should be on leveraging existing sources of funding and ensuring the optimal There is need for a mechanism to identify and allocation of funding for forest protection first before transfer important initiatives between states. This increasing total spending. 135 Strengthening Forest Fire Management in India TABLE 4.1: SUMMARY OF RECOMMENDATIONS AND PRIORITIES Recommendation Lead Implementer Priorities and Timing FFPM Guidelines to cover: MoEFCC (in consultation MoEFCC to begin drafting these with relevant stakeholders) immediately, and to finalize them in • Revised Working Plan Code consultation with relevant stakeholders. • Development of Standard Operating Procedures (SOPs) by the SFDs (see below) • Fire lines, siting and maintenance Controlled burning • Silvicultural practices (prevention and post-fire restoration or rehabilitation) • Common classification scheme for the causes of forest fires • Standard protocols for post-fire reporting, the investigation of fire causes, and standard methods for assessment of damages • Incentivizing accurate reporting by field staff on fires occurrence, burnt area, and damages Ensuring adequate funding and field SFDs In the near term, states should examine staffing existing budget resources to determine if enough is being allocated for FFPM. CAMPA offers a potential source of funding. In the longer term, states should seek to increase funding by increasing productivity of forests and thereby the revenue generated from the sector. A top priority is for SFDs to fill vacancies for field staff and community firewatchers in fire-prone areas. Boots on the ground are essential for all aspects of FFPM, including prevention, detection, and timely response to fires. Training in fire suppression DFE (training curriculum) There is a real need for this, and this (prevention, detection, and post-fire to be rolled out in activity must begin immediately with reporting) for field staff coordination with SFDs the development of a curriculum for all forest guards and other field-level officers in the SFDs. Provision of equipment for field staff SFDs in coordination with There is a real need for this, and this FRI activity must begin immediately. The focus should be on basic hand tools, safety gear and other equipment for ground crews that are appropriate and suited to local needs and conditions. Strengthening Forest Fire Management in India   136 Recommendation Lead Implementer Priorities and Timing Establishment of coordination MoEFCC at the national This process should also begin mechanism, at national, state, level, and SFDs at the state immediately, both to define the and district levels, between and district level, working coordination mechanism and also to forest departments and disaster with relevant disaster establish it. MOEFCC and NDMA should management agencies management agencies take the lead and provide guidance for the state-level mechanisms Development and deployment of FSI with SFDs FSI to continue the development of Fire Danger Rating System FDRS in collaboration with SFDs, with the recognition that this is a long-term process. The immediate priority is to formalize this process and create a mechanism for SFDs to provide input to the FDRS and field data/feedback for testing the FDRS. Continued improvement of satellite- FSI with SFDs FSI has a well-functioning nationwide based fire detection system satellite-based fire detection system in place. This system can be refined as new technologies and detection algorithms come available, and both FSI and SFDs should work toward this. The immediate priority is to improve two- way communication between FSI and SFDs and strengthen the process by which field-level forest officers provide feedback to both SFDs and FSI on the accuracy of the alerts. National Policy or Action Plan (which MoEFCC Core group with Director General of would also clarify role of other Forest and representatives from SFDs, agencies) NDMA, NGOs, and research institutes to be established immediately to initiate a consultative process for the development of the national policy and action plan over the course of one to two years. Incentivizing communities SFDs working with There is a real need for this, and this communities and local activity must begin immediately, although NGOs it will entail a longer-term process Standard Operating Procedures SFDs in consultation with SFDs to begin development once relevant agencies MoEFCC issues guidelines. Defining a national research agenda ICFRE ICFRE, as part of its mandate, has (with funding) developed a National Forestry Research Plan for 2000-2020. FFPM research needs can be defined as part of this on- going process. Formal mechanism for knowledge MoEFCC MoEFCC organizes annual meetings of sharing between states PCCFs and one of these meetings can focus on forest fires. 137 Strengthening Forest Fire Management in India Recommendation Lead Implementer Priorities and Timing National Forest Fire Information FSI While such a database will serve many Database needs, it can be developed over the coming years once the underlying processes to collect the necessary data have been established. National Center of Excellence ICFRE in coordination with While these is need for such a Center of FSI Excellence, this too can be developed over the coming years, once the underlying processes have been established. Acronyms: DFE = Directorate of Forest Education; FDRS = fire danger rating system; FRI = Forest Research Institute; FSI = Forest Survey of India; ICFRE = Indian Council for Forestry Research and Education; MoEFCC = Ministry of Environment, Forest and Climate Change; NDMA = National Disaster Management Authority; NGO = non-governmental organization; SFD = state forest department Strengthening Forest Fire Management in India   138 ANNEX 1 DATA AND METHODS FOR GEOSPATIAL ANALYSIS OF FOREST FIRES103 The analysis of forest fire trends and characteristics the ground, making them invisible to the satellite- relies primarily on observations of thermal anomalies by based instrument. Second, due to the coarse spatial the Moderate Resolution Imaging Spectroradiometers resolution of the sensor, MODIS may not be able to (MODIS) aboard the Aqua and Terra satellites. detect low-intensity surface fires under canopy cover. Active fire data from MODIS are available starting Also, fires on lands adjacent to forests may be detected in November 2000 for the Terra satellite, and June as occurring within forested areas. Third, though 2002 for the Aqua satellite. The latest fire detections MODIS has a relatively short return period (around used for the analysis were for December 2016. Each 6 hours between overpass), it will not detect fires that detection of an active fire by MODIS represents a 1-km are started and extinguished before satellite revisit. by 1-km pixel containing an anomaly. One or many fires may be burning within or even nearby a pixel to signal an anomaly. Fires do not need to reach a size of 1. O  VERALL PATTERNS AND 1 km2 to be detected. MODIS can detect fires as small TRENDS IN FOREST FIRES as 50 m2 depending on the intensity of the fire and its visibility from space. References to the number of fire The analysis of fire occurrence per district and detections made in this report refer to counts of fire- region from 2003 to 2016 is performed using district containing pixels, not individual ignitions or events. boundaries as defined in 2012.105 The one exception is for districts within the present-day area of Telangana, MODIS data on active fires for this analysis were which became a state in 2014. Districts in Telangana processed and provided by Forest Survey of India are designated as belonging to Telangana and not (FSI) using the MODIS Collection 6 monthly Andhra Pradesh. Regions are defined by FSI based standard science-quality data product for active fires on physiography and similarities in forest use and are (MCD14ML).104 FSI screens forest fires by clipping the classified as per table A1.1 below. MODIS data to include lands under forest department management. The boundaries of forest department lands have been mapped and digitized down to the The MCD14ML active fire data provided by FSI are lowest administrative level (beats) in 10 states. For further screened by including only high-confidence states where forest boundaries have yet to be digitized, detections in the analysis. High-confidence detections FSI screens the MODIS data for areas with forest cover are defined as those with a confidence score of at least per the latest India State of Forest Report. Also, FSI 50 on a 100-point scale.106 only includes observations for the months of January to June, the peak fire season for most of the country. The MCD14ML data product provides information about the frequency or occurrence of active fires; The MODIS-derived data on active fires have inherent however, it does not give the area affected by fire. limitations that are worth noting at the outset. First, Thus, for the analysis of burned forest area per cloud cover and heavy smoke may obscure fires on district and region, a different data product is needed. 103. Analysis by Christopher Sall, World Bank, csall1@worldbank.org 104 The archived MODIS Collection 6 data product MCD14ML for active fires is available from NASA’s Fire Information for Resource Management System (FIRMS) at https://firms.modaps.eosdis.nasa.gov/download/. 105 See Open Government Data (OGD) Platform India, “Number of Districts/DRDAs/Blocks/Villages in the Country – State for 2012,” https://data.gov.in/catalog/number-districts-drdas-blocks-villages-country 106 For an explanation of confidence scores, see Giglio (2015). User’s guides are available at University of Maryland, “MODIS Active Fire and Burned Area Products User Guides,” http://modis-fire.umd.edu/pages/manuals.php. 139 Strengthening Forest Fire Management in India TABLE A1.1: REGIONAL DEFINITIONS Central Northeast North South Western West Himalayas Chhattisgarh Arunachal Bihar Andaman & Himachal Pradesh Dadra & Nagar Pradesh Nicobar Haveli Jharkhand Assam Chandigarh Andhra Pradesh Jammu & Kashmir Daman & Diu Madhya Pradesh Manipur Delhi Goa Uttarakhand Gujarat Maharashtra Meghalaya Haryana Karnataka Rajasthan Odisha Mizoram Punjab Kerala West Bengal Nagaland Uttar Pradesh Lakshadweep Sikkim Puducherry Tripura Tamil Nadu Telangana Source: Classifications by Forest Survey of India (FSI) based on physiography and similarities in forest use Fire-affected area is estimated using the standard in state s during time period t (annual, 7-day, 14-day, science-quality data product for monthly burnt area or 30-day period); Y is the year; and εs,t is a state and (“MCD45A1”) provided by NASA and the University period-specific error term. The coefficient of interest of Maryland (United States), which is derived from is β1, which can be interpreted as the percent change MODIS and has a spatial resolution of 500 m. Fire- in F per year. Regressions were repeated for each affected area includes any area that was under forest state and time period. The estimated values for β1 are cover in 2000 (at least 10-percent canopy cover) and presented in table A1.3 of chapter 1. which was affected at least once by fire between 2003 and 2016. Data on forest cover in 2000 came from Hansen et al. (2013).107 2. FACTORS INFLUENCING FIRE POTENTIAL AND BEHAVIOR Year-on-year trends in active fire locations per state from 2003 to 2016 are assessed using the MCD14ML data product, screened for high-confidence detections. 2.1 Weather Regression analysis is performed to estimate the average year-on-year change in the total number of 2.1.1 Forest fire seasonality fires per year as well as the number of fires during the peak 7-, 14-, and 30-day period of the forest fire The violin plots in figure 4 of chapter 1 illustrate the season. The peak period is defined as the running seasonality of forest fires by showing how fires in each period during which the greatest number of fires is state are distributed across the months of the year.108 detected. Because it is defined on a running basis, The figures are constructed using the MCD14ML data the timing of the peak period is allowed to vary from product, screened for high-confidence detections. year to year. An increase in the number of fires during Because the MCD14ML data provided by FSI are only the 7-, 14-, or 30-day peak period would suggest an for the months of January to June, they are not used intensification of the fire season or a trend toward for the analysis of fire seasonality. Instead, active fires larger, more severe fire events. The annual percent for all months from 2003 to 2016 are extracted for change in the number of fires is given as: forested areas using the Hansen et al. (2013) forest cover data for 2000. States and UTs with fewer than lnFs,t=β0+β1 Y+εs,t, 400 total active fire detections in forested areas from 2003 to 2016 are excluded from the analysis. These where Fs,t is the number of active forest fire detections states/UTs include Andaman and Nicobar Islands, 107 See Hansen et al. (2013). Data available from, http://earthenginepartners.appspot.com/science-2013-global-forest . 108 See Hintze and Nelson (1998). Strengthening Forest Fire Management in India   140 Chandigarh, Dadra and Nagar Haveli, Goa, Haryana, Mean monthly SPI values were estimated for each Puducherry, Rajasthan, and Sikkim. of the country’s 647 by overlaying gridded SPI data obtained from Columbia University’s International In figure 4, the lengths of the violins show the Research Institute for Climate and Society (IRI) on continuous period between September 1 and August the district boundaries and calculating zonal statistics. 31 the following year in which 80 percent of all fires Above- or below-normal monsoon rainfall was defined are concentrated. The widths of the violins represent according to the 3-month SPI data for June, July, and the kernel density of detections, binned into 7-day August (JJA) as well as July, August, and September periods—the wider the area, the more fires have (JAS). The effect of post-monsoon rainfall was also occurred around that week of the year. Bars within the tested using the 6-month SPI for June to December violins show the interquartile range of observations— and the 3-month SPI for October, November, and the shorter bar, the more concentrated the fire season. December combined with the 3-month SPI for JAS. 2.1.2 Monsoon rainfall and fire season severity Descriptive statistics for fire occurrence, monsoon and post-monsoon SPI, and forest area are given in table District-level statistical analysis was performed to A1.2 below. Each observation in the table represents a evaluate how monsoon precipitation can influence the district and a year. severity of coming fire season. Fire season severity is indicated by the number of fires detected in a district The outcome variable in the analysis is a count of from January to May in the following year (the peak fire detections per district. Though most districts fire season before the arrival of the monsoon rains experience only a handful of fires each year, there in June-July). Fires per district were calculated for is a long tail of districts with many hundreds of fire district boundaries as of 2012 using the MCD14ML detections (figure A1.1). Because of over-dispersion data product provided by FSI, with additional in the count of fire detections (variance > mean), the screening for high-confidence detections. Above- or number of districts and years with zero fires is under- below-normal monsoon precipitation was measured predicted by a Poisson regression model. A negative according to the monthly Standardized Precipitation binomial regression (NBR) model that allows the Index (SPI). The SPI, as elaborated by Guttman conditional variance of fire detections to exceed the (1999), is a unitless index equal to the number of mean provides a better fit. Because the NBR model standard deviations that precipitation differs from still under-predicts the occurrence of zero fires, a the long-term average over a specified time scale.109 zero-inflated NBR model is also evaluated, with a TABLE A1.2: DESCRIPTIVE STATISTICS FOR DISTRICT-LEVEL ANALYSIS OF MONSOON RAINFALL AND FOREST FIRES Variable Observations Mean Std. deviation Minimum Maximum Fire detections, Jan-May (count) 6930 45.18658 121.5167 0 1618 3-month SPI for JJA 6930 -0.01031 0.833507 -2.76078 2.758469 3-month SPI for JAS 6930 0.080339 0.784677 -3.09023 3.090236 6-month SPI for JASOND 6930 0.105011 0.96672 -3.09023 3.090236 3-month SPI for OND 6930 -0.00678 0.896613 -3.09023 2.572988 Forest cover in 2000 (km2) 6930 933.6713 1473.945 0.008326 10097.75 Notes: SPI = Standardized Precipitation Index; JJA = June, July August; JAS = July, August, September; OND = October, November, December; JASOND = July, August, September, October, November, December Sources: SPI data are from the International Research Institute for Climate and Society, Columbia University, https://iridl.ldeo.columbia.edu/ SOURCES/.IRI/.Analyses/.SPI/?Set-Language=en; forest cover data from Hansen et al. (2013) SPI data are from the International Research Institute for Climate and Society, Columbia University. The data have a spatial resolution 109 of 0.5° x 0.5°. For references and to download the SPI data, see the IRI/LDEO Climate Data Library https://iridl.ldeo.columbia.edu/ SOURCES/.IRI/.Analyses/.SPI/?Set-Language=en. 141 Strengthening Forest Fire Management in India logit model as the link function. In the zero-inflated per district. Districts with a very small area of forest model, the “excess” number of zero fires in the sample are expected to be much more likely to have zero fires. of districts and years (2,166 of 6,930 observations) is Control variables in the analysis include forest cover assumed to be influenced by the area of forest cover per district, state-level fixed effects to account for unexplained differences in fire incidence across FIGURE A1.1: FIRES DETECTED PER states, and the year of observation to separate out DISTRICT FROM JANUARY the unexplained effect of year-on-year trends in fire TO MAY, 2003-2016 occurrence. 80 Parameter estimates and diagnostic information for the NBR and zero-inflated NBR models are provided Percent of observations in tables A1.3 and A1.4 below. According to table A1.3, 60 districts with monsoon rainfall that is one standard deviation above the long-term average for JJA and JAS typically experience about 7-12 percent fewer 40 fires the following year (models 1 and 2). If rainfall continues to be one standard deviation above average over the longer period of July to December, then 20 the average district will be predicted to experience about 21 percent fewer fires (model 4). Furthermore, separating the effects of monsoon during JAS and 0 0 500 1000 1500 post-monsoon rainfall during OND, it emerges that Fires detected per district, Jan-May JAS rainfall is more influential in determining fire Sources: MODIS monthly data product for active fires (MCD14ML), season severity (model 3). provided by Forest Survey of India (FSI) TABLE A1.3: NEGATIVE BINOMIAL REGRESSION RESULTS FOR MONSOON RAINFALL AND THE NUMBER OF FIRES OBSERVED JANUARY-MAY THE FOLLOWING YEAR Model 1 Model 2 Model 3 Model 4 Outcome variable Count of fires per district from Jan-June next year Variables of interest 3-month SPI for JJA -7.206* 3-month SPI for JAS -11.23*** -10.44*** 3-month SPI for OND -5.980** 6-month SPI for JASOND -20.61*** Control variables District forest area Yes Yes Yes Yes Year-to-year trends in fires Yes Yes Yes Yes State-level fixed effects Yes Yes Yes Yes Diagnostics Observations 6930 6930 6930 6930 ln(alpha) 145.5*** 144.9*** 144.6*** 140.9*** BIC 48856.7 48843.2 48844.8 48744.6 AIC 48829.3 48815.8 48810.6 48717.2 Notes: * Significant at 90-percent level; ** significant at 95-percent level; *** significant at 99-percent level; coefficients represent the percent decrease in the predicted count of fires for each one-unit increase in the SPI; each unit increase in the SPI represents one standard deviation from the long-term average for rainfall. Strengthening Forest Fire Management in India   142 After accounting for excess zeros in the data for fire The Bayesian Information Criterion (BIC) and Akaike detections per district, table A1.4 shows that the Information Criterion (AIC) scores in the tables are predicted effect of above- or below-normal monsoon measures of parsimony and goodness of fit; lower BIC rainfall with the zero-inflated NBR model is only and AIC scores indicate an improved model. The slightly larger than with the NBR model. If JJA and zero-inflated NBR model offers lower BIC and AIC JAS rainfall are one standard deviation above normal, scores and is thus preferred. Differences in parameter the typical district will be expected to see about 10-12 estimates between the NBR and zero-inflated NBR percent fewer fires the following year. If rainfall for models are small, though. JASOND continues to be one standard deviation above normal, the district will experience about 21 percent 2.1.3 El Niño/Southern Oscillation (ENSO) and fire fewer fires. The parameter estimates for district forest season severity cover in the “certain zero” logit model in table A1.4 show that for each additional km2 of forest cover, the Exploratory statistical analysis was done to test the odds of a district having zero fires decreases by about relationship between ENSO and fire season severity 2 percent. using several different indices of ENSO that capture departures in mean monthly sea surface temperatures The diagnostic information in tables A1.3 and A1.4 from the climatological average for different equatorial confirms the choice of the zero-inflated NBR model regions of the Pacific Ocean. These indices include: over the NBR or Poisson models. The ln(alpha) statistic (1) the Niño 3 Index, which reflects temperatures in the tables test for over-dispersion. Statistically in the eastern Pacific (5°N-5°S, 150°W-90°W); (2) significant ln(alpha) values suggest over-dispersion the Niño 4 Index for the western Pacific (5°N-5°S, and reject the use of a Poisson model as an alternative. 160°E-150°W); and (3) the Niño 3.4 Index for the TABLE A1.4: ZERO-INFLATED NEGATIVE BINOMIAL REGRESSION RESULTS FOR MONSOON RAINFALL AND THE NUMBER OF FIRES OBSERVED JANUARY-MAY THE FOLLOWING YEAR Model 1 Model 2 Model 3 Model 4 Outcome variable Count of fires per district from Jan-June next year Explanatory variables of interest in the full model 3-month SPI for JJA -9.519*** 3-month SPI for JAS -12.37*** -11.82*** 3-month SPI for OND -6.311** 6-month SPI for JASOND -20.51*** Explanatory variables in the “certain zero” logit model District forest area -2.160** -2.160** -2.145** -2.158** Control variables in the full model District forest area Yes Yes Yes Yes Year-to-year trends in fires Yes Yes Yes Yes State-level fixed effects Yes Yes Yes Yes Diagnostics Observations 6930 6930 6930 6930 ln(alpha) 44.84*** 44.39*** 41.25*** 44.06*** BIC 47379 47372.6 47257.7 47361.7 AIC 47337.9 47324.7 47209.8 47313.8 Notes: * Significant at 90-percent level; ** significant at 95-percent level; *** significant at 99-percent level; coefficients represent the percent decrease in the predicted count of fires for each one-unit increase in the SPI; each unit increase in the SPI represents one standard deviation from the long-term average for rainfall. 143 Strengthening Forest Fire Management in India central Pacific (5°N-5°S, 170°W-120°W).110 The Niño In table A1.5, the only states for which a statistically 3.4 Index—and the standardized Oceanic Niño Index significant relationship between ENSO and fire season (ONI) that is derived from the Niño 3.4 Index—is the severity exists is Arunachal Pradesh, where La Niña most commonly used measure for determining the years have been followed by more severe fire seasons, existence of El Niño (or La Niña) events,111 defined and Odisha, where La Niña years have been followed as five consecutive 3-month running means of sea by more severe fire seasons. However, the correlation surface temperatures that are above (or below) the for both these states is weak and significant only at climatological average by at least 0.5°C. Research by the 90-percent level. The results indicate that the Kumar et al. (1999, 2006) has further suggested that relationship between ENSO and fire season severity is El Niño events marked by warmer seas in the central not straightforward, and there is insufficient evidence equatorial Pacific are more likely to produce drought to suggest a meaningful link that could be used for in India than events with warming concentrated in planning purposes. the eastern Pacific. The Niño3.4 Index reflects sea surface temperatures in the central equatorial region To further test the hypothesized mechanism by which most closely associated with drier monsoons during ENSO supposedly influences fire season severity, El Niño years. For each of the indices, the analysis regression analysis was performed at the state and was run separately using values averaged for June- district level. In the first stage of the analysis, the September (JJAS) and June-December (JJASOND) to Niño 3.4 Index was used as an instrument to predict capture the monsoon and post-monsoon months. monsoon rainfall. In the second stage, predicted monsoon rainfall is then related to fire detections, Spearman’s rank-order correlation coefficients were such that: calculated to determine if the number of active fire locations detected per state during peak fire season (January to June) has varied systematically with ln firest+1 = β0+β1 ln precipt +ε, ENSO. Spearman’s rank-order correlation is a non-parametric test of the strength and direction of ln precipt =γ0+γ1 ninot+ω variation between two variables. A coefficient of -1 indicates a perfectly monotonic inverse relationship where fires is the total number of fire detections per state between fire occurrence and ENSO, suggesting in or district during January-May in year t + 1; ninot is the this case that strong La Niña episodes (cooler sea average Niño 3.4 Index value for June-September or surface temperatures) are strongly correlated with June-December in year t; precipt is predicted rainfall more severe fire seasons. A value of 1 indicates a for June-September or June-December in year t; and ε perfectly monotonic positive relationship, suggesting and ω are error terms. The coefficient of interest is β1, that strong El Niño episodes (warmer sea surface the percent change in fire detections for each percent temperatures) are followed by bad fire seasons. A change in monsoon rainfall attributed to ENSO. In a value of 0 indicates no relationship. The Niño 3, variation of the district-level analysis, ln precipt was Niño 4, and Niño 3.4 Index scores JJA, JAS, and alternatively replaced with SPIt , the Standardized JASOND were compared against the total number of Precipitation Index value for June-September or fire detections from January to May for the following June-December. The coefficient of interest is β1, the years using the MCD14ML data provided by FSI, percent change in fire detections for each unit change which was further screened for high-confidence fire in ln precipt or SPIt . The coefficient β1 was not found detections. The analysis was run for individual states to be statistically significant in any of the variations, as well as the entire nation. Results are presented in though γ1 in the first-stage regression was significant, table A1.5 below. reinforcing the link between ENSO and the monsoon. 110 Monthly ENSO data for 1950-2017 are from the Climate Prediction Center, National Weather Service, National Oceanic and Atmospheric Administration, US Department of Commerce, “Monthly Atmospheric and SST Indices,” http://www.cpc.ncep.noaa.gov/ data/indices/. 111 See Trenberth, Kevin & National Center for Atmospheric Research Staff (eds), “The Climate Data Guide: Nino SST Indices (Nino 1+2, 3, 3.4, 4; ONI and TNI),” February 2016, https://climatedataguide.ucar.edu/climate-data/nino-sst-indices-nino-12-3-34-4-oni- and-tni. Strengthening Forest Fire Management in India   144 TABLE A1.5: SPEARMAN’S RANK-ORDER CORRELATION COEFFICIENTS, ENSO INDEX VALUES FOR JUNE-DECEMBER AND THE NUMBER OF FIRES DETECTED JANUARY-MAY THE NEXT YEAR, 2003-2016 Niño 3 Index Niño 4 Index Niño 3.4 Index Area JJAS JJASOND JJAS JJASOND JJAS JJASOND National -.02 -.04 -.2 -.19 -.13 -.13 Andaman & Nicobar .24 .05 -.06 -.02 0 .05 Andhra Pradesh -.18 -.14 -.2 -.21 -.13 -.16 Arunachal Pradesh -.23 -.36 -.53* -.53* -.43 -.46* Assam .02 -.11 -.35 -.32 -.3 -.24 Bihar .19 .04 -.16 -.09 .05 -.04 Chandigarh -1 -1 -1 -1 -1 -1 Chhattisgarh .28 .23 .01 .05 .13 .1 Dadra & Nagar Haveli -.19 -.26 -.09 -.09 -.06 -.17 Delhi .71 .71 .71 .71 .71 .71 Goa .11 .17 .35 .3 .26 .28 Gujarat .19 .18 .22 .2 .33 .22 Haryana .28 .15 .16 .19 .18 .15 Himachal Pradesh .38 .24 .14 .18 .25 .19 Jammu & Kashmir .2 .03 -.13 -.12 -.04 -.06 Jharkhand .36 .19 -.01 .03 .15 .09 Karnataka -.25 -.11 -.08 -.14 -.03 -.09 Kerala -.05 -.05 -.17 -.16 -.07 -.11 Madhya Pradesh -.22 -.26 -.34 -.34 -.26 -.32 Maharashtra .05 .03 -.13 -.1 -.02 -.07 Manipur -.1 -.04 -.14 -.15 -.07 -.07 Meghalaya .2 .19 -.04 -.04 -.02 .06 Mizoram .29 .28 .16 .14 .25 .23 Nagaland -.06 -.11 -.32 -.31 -.23 -.21 Odisha .55* .48* .25 .29 .45 .39 Punjab .13 .01 .01 .05 .06 0 Rajasthan -.2 -.18 -.38 -.32 -.35 -.33 Sikkim -.11 -.23 -.37 -.36 -.35 -.29 Tamil Nadu -.28 -.1 -.2 -.22 -.13 -.15 Telangana -.08 -.16 -.27 -.25 -.25 -.24 Tripura -.12 -.02 .11 .06 .05 .07 Uttar Pradesh .23 .22 .12 .15 .16 .16 Uttarakhand .11 .08 .13 .12 .13 .11 West Bengal -.03 -.15 -.33 -.31 -.29 -.25 Notes: JJAS = June-September; JJASOND = June-December; * = significant at 90-percent level; positive coefficients indicate positive relationship between ENSO and fire detections; negative coefficients indicate inverse relationship 145 Strengthening Forest Fire Management in India 2.1.4 Summer weather conditions and fire potential months. Fires were detected in 12,920 of 38,610 of the district-months from 2003 to 2015. District-level regression analysis was performed to quantify the relationship between weather conditions A logistic regression model was employed to quantify during the fire season and the odds of fire occurrence. how changes in mean monthly weather conditions Weather variables tested include mean temperature, influenced the chances that a forest fire would be precipitation, and wet day frequency. Monthly detected in a district. Additional control variables weather data were obtained from the University of were introduced to account for differences in forest East Anglia’s Climate Research Unit University (CRU) area, state-level fixed effects, and unexplained year- of East Anglia.112 The gridded weather data (with to-year variation in fire frequency. The basic form of a resolution of 0.5° x 0.5°) were overlaid on district the equation used to estimate the odds of a fire being boundaries (as of 2012) to calculate the monthly detected in a district was: average for each district in India over the years from 2003 to 2015, the latest available year of data. Monthly fire detections per district were summarized where fired is the probability of a fire being detected from the MCD14MCL data product, provided by FSI in district d during a given month m; W is the and further screened for high-confidence detections. monthly weather variable (mean temperature, total Analysis was restricted to the months of January to precipitation, or total wet days); forestd is the area in May, the height of the fire season before the arrival of the district that had at least 10% tree canopy cover in the monsoon rains. 2000; states is a binary dummy variable equal to 1 if districtd is in state/UT s and zero otherwise; monthm is a dummy variable equal to 1 if the observation is in Descriptive statistics for each of the weather variables month m and zero otherwise; and εd,i is the error term. are reported by month in table A1.6 below. Each The coefficient of interest is β1, the increase in the observation in the table represents an individual log odds of fire for each unit increase in W. Lagged district and a month. The table shows tremendous values for W in months m - 1, m - 2, and m - 3 and variability in weather conditions in districts across the monsoon rainfall in year y - 1 (precipitation and wet country, with mean monthly temperatures ranging days during June-September of the previous year) from -14°C to 35°C and monthly precipitation ranging were also introduced to test the lingering effects of from 0 mm to 1,136 mm during the peak fire season weather in previous months on the odds of fire. TABLE A1.6: MEAN MONTHLY WEATHER CONDITIONS AND FIRE DETECTIONS BY DISTRICT, 2003-2015 Month Observations Mean tempera- Precipitation Wet days Fire detections ture (°C) (mm) January 6,435 17.4 (-14.8-28) 10.0 (0-231.8) 1.1 (0-11.1) 1.3 (0-114) February 6,435 20.0 (-13.9-28.6) 14.6 (0-225.5) 1.3 (0-11.3) 6.3 (0-298) March 6,435 24.3 (-6.7-31.5) 18.2 (0-431.7) 1.8 (0-12.9) 26.4 (0-1292) April 6,435 28.0 (-2.6-34.8) 37.3 (0-1135.5) 2.5 (0-22.2) 12.9 (0-963) May 6,435 30.0 (-.6-36.7) 52.4 (0-943.7) 3.6 (0-21.5) 4.0 (0-190) Notes: mean values by district are weighted by district area; values in parentheses represent the range in the sample Source: MODIS monthly data product for active fires (MCD14ML), provided by FSI; monthly weather data from CRU TS v4.00 gridded time- series dataset, available at https://crudata.uea.ac.uk/cru/data/hrg/ 112 See CRU TS v4.00 gridded time-series dataset, available at https://crudata.uea.ac.uk/cru/data/hrg/ Strengthening Forest Fire Management in India   146 Regression results are presented in tables A1.7 and percent. That means an additional cm of rainfall A1.8 below. The coefficients reported in the tables would lower the odds of fire detection by 2.7 percent have been transformed as percent changes in the odds [(1 - .00272)^10 * 100% - 100%], and two more wet of a fire detection. The tables show that precipitation days would reduce the odds of fire detection by 22.1 in the current month reduces the odds of fire, while percent [(1 - .1181) ^2 * 100% - 100%]. Fire potential higher temperatures raise the odds of fire. Each is even more sensitive to temperature. Each 1°C additional mm of precipitation to fall within the past increase in mean temperature during the past month month reduces the odds of a fire being detected in raises the odds of fire by 16.6 percent. the average district by 0.3 percent. Each additional wet day within the current month (a day with more The signs on the coefficients in tables A1.7 and A1.8 than 0.1 mm of precipitation) reduces the predicted for monsoon rainfall and wet-day frequency in months odds of fire detection in the average district by 11.8 m – 1, m – 2, and m – 3 are all positive. This suggests TABLE A1.7: DISTRICT-LEVEL REGRESSION FOR MONTHLY PRECIPITATION (MM) AND ODDS OF FIRE DETECTION Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Weather variables Precipitation 0.0343** 0.0428** -0.409*** -0.638*** -0.494*** -0.467*** -0.386*** -0.273** -0.272** (mm, month t) Precipitation -0.00657 0.0169 0.0516** 0.0435* (mm, t - 1) Precipitation 0.0235 0.0448 0.0248 (mm, t - 2) Precipitation 0.035 0.0137 (mm, t - 3) Monsoon 0.0313 precipitation (mm) Temperature 2.491*** 7.419*** 9.582*** 13.06*** 19.48*** 18.37*** 16.61*** 16.59*** (C, month t) Temperature -5.942** 5.328* 5.043* 5.155* (C, t - 1) Temperature -11.65*** 5.011 4.621 (C, t - 2) Temperature -17.64*** -17.79*** (C, t - 3) Additional controls District forest area Yes Yes Yes Yes Yes Yes Yes State-level fixed Yes Yes Yes Yes Yes Yes effects Month-of-year Yes Yes Yes Yes Yes fixed effects Diagnostics Observations 32110 32110 32110 32110 32110 32110 32110 32110 32110 BIC 42689.5 42435.9 35790.6 33080.7 31628.6 31615.4 31472.6 31142.8 31110.6 AIC 42672.7 42410.7 35757.1 33047.2 31553.2 31523.2 31363.7 31017.1 30968.2 Pseudo R2 9.39E-05 0.00628 0.162 0.226 0.261 0.262 0.266 0.274 0.275 Notes: * p < .10, ** p < .05, *** p < .01; coefficients are expressed as percent change in odds of fire detection in district per unit increase in explanatory variable 147 Strengthening Forest Fire Management in India TABLE A1.8: DISTRICT-LEVEL REGRESSION FOR MONTHLY WET-DAY FREQUENCY (DAYS PER MONTH WITH > .01 MM PRECIPITATION) AND ODDS OF FIRE DETECTION Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Weather variables Wet days (month t) 0.307 1.498*** -11.96*** -17.42*** -16.38*** -15.62*** -14.10*** -11.64*** -11.81*** Wet days (t - 1) -0.783 -0.995 0.829 0.583 Wet days (t - 2) 2.079 2.545* 2.007* Wet days (t - 3) 1.443 0.365 Wet days during 1.209 monsoon Temperature 2.583*** 7.011*** 9.930*** 10.07*** 14.75*** 14.34*** 12.97*** 13.03*** (C, month t) Temperature -4.531* 2.914 3.525 3.643 (C, t - 1) Temperature -7.712* 7.340** 7.105** (C, t - 2) Temperature -15.75*** -15.96*** (C, t - 3) Additional controls District forest area Yes Yes Yes Yes Yes Yes Yes State-level fixed Yes Yes Yes Yes Yes Yes effects Month-of-year fixed Yes Yes Yes Yes Yes effects Diagnostics Observations 32110 32110 32110 32110 32110 32110 32110 32110 32110 BIC 42692.9 42428 35520.4 32693.5 31332.1 31346.2 31261.9 30999.8 30954.8 AIC 42676.1 42402.8 35486.9 32660 31265.1 31254.1 31153 30874.2 30820.7 Pseudo R2 1.46E-05 0.00647 0.169 0.235 0.268 0.268 0.271 0.277 0.278 Notes: *p < .10, ** p < .05, *** p < .01; coefficients are expressed as percent change in odds of fire detection in district per unit increase in explanatory variable that more rainfall in earlier months without a marginal quantified by the Keetch-Byram Drought Index increase in rainfall in the current month can lead to higher (KBDI), which measures the deficit of moisture in odds of fires occurrence. The mechanism for this shift the upper soil or duff layer of a forest. Higher KBDI in the direction of influence is unknown, but one values indicate a lack of available water, leading to the hypothesis is that higher precipitation several months increased flammability of fine fuels such as dried-out earlier may stimulate the growth of grasses and other grasses and decaying organic material in the ground such as buried roots or wood, and signaling the vegetation and increase the availability of fine fuels. potential for more intense fire behavior (Keetch and If, on the other hand, rainfall continues to be higher Byram 1968). As originally formulated, the KBDI is than normal into the current month, then the odds of calculated on an 800-point scale, where each point forest fires will decrease. represents 1/100 inch of additional rainfall necessary to restore soils back to a saturated state. In metric 2.1.5 Drought conditions and fire potential units, the index is calculated on a 200-point scale, with each point representing 1 mm of rainfall. Inputs to The potential for more intense fire behavior under the KBDI include daily rainfall, mean annual rainfall, conditions or warmer and drier weather is further and daily maximum temperature. Strengthening Forest Fire Management in India   148 District-level logistic regression analysis was performed to evaluate the relationship between the KBDI and forest fire occurrence. Gridded daily temperature and precipitation data were obtained from the where fired,i is the probability of a fire occurring on that Physical Sciences Division of the Earth Systems day and being detected by MODIS, the odds of fire are Research Laboratory at NOAA. The daily data from expressed as fired,i / (1 - fired,i); KBDId,i is the daily KBDI NOAA combine reports from weather stations with value; forestd is the area in the district that had at least additional information from satellite monitoring and 10% tree canopy cover in 2000; states is a binary dummy forecasting.113 The KBDI was calculated for each 0.5° variable equal to 1 if district d is in state s and zero x 0.5° grid cell for each day from 2011-2016, with a starting value of zero assumed for July 1, 2011, otherwise; monthm is a binary dummy variable equal to corresponding with the monsoon season when it was 1 if day i is in month m and zero otherwise; and εd,i is assumed that soils would be saturated. Daily KBDI the error term. The states captures unobserved state- values were estimated for each district in the country, level characteristics that are thought to influence the and these daily KBDI values were then overlaid with likelihood of fire and are time-invariant on the scale of daily satellite observations of active fire locations the years covered in the analysis. The monthm variable from MODIS. Daily fire detections per district were captures seasonal trends which are not reflected in the summarized from the MCD14MCL data product, daily drought index values. provided by FSI and further screened for high- confidence detections. The analysis was performed The regression results in table A1.10 below support for the months of January – June. Table A1.9 below the hypothesis that the KBDI is a significant predictor shows the percent of districts and days for which a of fire danger, as measured by the odds that an active forest fire was detected when the KBDI was within a given range. forest fire will be detected in a particular district on a particular day. Coefficients in the table are expressed To determine the relationship between KBDI and as odds ratios, or the factor by which the daily odds of forest fires, the odds that a fire would be detected in fire detection are multiplied for each unit increase in district d in states on day i in month m were estimated KBDI. A one-unit increase in the KBDI is predicted as: to raise the odds of fire detection in a district by a TABLE A1.9: SHARE OF DAILY OBSERVATIONS WITH FIRES DETECTED, BY STAGE OF DROUGHT AS MEASURED BY THE KEETCH-BYRAM DROUGHT INDEX (KBDI) AND BY REGION, FEBRUARY 1 TO MAY 31 (2012-2016) Share of forest fire detections by region (% of districts and days for which a forest fire was detected when KBDI was in the KBDI Drought stage given range) Central North Northeast South West W. Himalaya 0-99 0 0.00 0.00 1.55 0.02 0.14 6.52 (saturated soils) 100-199 1 0.01 0.00 2.06 0.02 0.00 13.38 200-299 2 0.00 0.19 3.44 0.35 0.00 5.13 300-399 3 0.02 1.08 3.74 2.01 0.00 5.73 400-499 4 0.11 4.26 5.93 5.88 0.00 9.79 500-599 5 0.60 9.60 13.62 13.64 3.24 15.91 600-699 6 7.47 19.71 24.31 24.02 14.19 26.23 700-800 7 91.79 65.16 45.34 54.05 82.43 17.31 (severe drought) Notes: regional averages are constructed by weighting districts by size of forest area See NOAA, “CPC Global Unified Gauge-Based Analysis of Daily Precipitation” and “CPC Global Daily Temperature,” NOAA/OAR/ 113 ESRL PSD, Boulder, Colorado, United States, http://www.esrl.noaa.gov/psd/. 149 Strengthening Forest Fire Management in India factor of 1.005. A 100-unit increase in KBDI on the stage of drought to the next, as compared to the odds 800-unit scale of the index raises the odds of fire by a of fire in stage 0. The coefficients for Model 1 suggest factor of 100.5. that the odds of fire jump as KBDI goes from 0 to 100 and jump even more as KBDI passes 700. In Model 2, In table A1.10 above, KBDI is treated as a continuous two additional categorical variables are introduced for variable, and the relationship between KBDI and the drought stages 0 and 7 to better capture these jumps. log odds of fire is assumed to be linear. Alternatively, BIC and AIC scores are lower for Model 2 than for the KBDI may be specified as a categorical variable. A continuous model in table A1.10, and likelihood-ratio series of binary dummy variables were created for each tests agree that the categorical model is preferable. stage of drought, as defined by Keetch and Byram and The linear model understates the effect of changes as depicted in table A1.9. in drought at either extreme of the KBDI scale and overstates the effect at the middle of the scale. Treating KBDI as a categorical variable relaxes the assumption that the relationship between KBDI and 2.2 Topography the log odds of fire occurrence is perfectly linear. For example, it may be that drought progressing from Previous research in other parts of the world has KBDI 200 to KBDI 300 does not increase the log odds found that human-caused ignitions of forest fires tend of fire occurrence as much as going from KBDI 600 to to occur more along road networks.114 Analysis was KBDI 700; or it may be that drought does not affect conducted to explore the distribution of forest fires in the odds that a forest fire will occur until conditions India in proximity to built-up areas and roads. reach a certain level of severity (e.g., KBDI 600). India does not maintain a national database of reported Table A1.11 shows the results for the categorical ignitions of forest fires, so observations of thermal model. Coefficients are expressed as odds ratios and anomalies by the MODIS and VIIRS instruments in can be interpreted as the factor by which the odds of areas that had forest cover as of 2000 were used as fire increase (are multiplied) as KBDI rises from one a proxy indicator for fire occurrence.115 Raster layers TABLE A1.10: REGRESSION RESULTS FOR ANALYSIS OF KBDI AND ODDS OF FIRE DETECTION, TREATING KBDI AS A CONTINUOUS VARIABLE Model 1 Model 2 Model 3 Model 4 KBDI 1.003* 1.004* 1.006* 1.005* Controls Forest area Yes Yes Yes State-level fixed effects Yes Yes Month-of-year fixed effects Yes Diagnostics Observations 458035 458035 456221 456221 BIC 149516 129259.3 115887.9 107183.5 AIC 149493.9 129226.2 115557 106797.4 Pseudo R2 0.0269 0.159 0.247 0.305 Notes: * significant at 99-percent level; coefficients expressed as odds ratios (factor by which daily odds of fire detection increase for each unit increase in KBDI) 114 As an example, from North America, see Narayanaraj and Wimberly (2012). 115 The MODIS monthly science-quality data product (MCD14ML) and VIIRS 375 m data near-real-time data product (VNP14IMGTDL_ NRT) were used. Data are available from NASA, Fire Information for Resource Management Systems, “FIRMS Fire Archive Download for MODIS Collection 6 and VIIRS 375 m,” https://firms.modaps.eosdis.nasa.gov/download/. As noted elsewhere in the Annex, only observations with a confidence score of 50 or higher were used. Forested areas are those with at least 10-percent canopy cover in 2000, as per Hansen et al. (2013). Strengthening Forest Fire Management in India   150 TABLE A1.11: REGRESSION RESULTS FOR ANALYSIS OF KBDI AND ODDS OF FIRE DETECTION, TREATING KBDI AS A CATEGORICAL VARIABLE REPRESENTING THE STAGE OF DROUGHT Model 1 Model 2 Drought stage KBDI range Odds ratio KBDI range Odds ratio 0-49 1 Stage 0 0-99 1 50-99 6.3* Stage 1 100-199 5.6* 100-199 13.1* Stage 2 200-299 6.8* 200-299 16.2* Stage 3 300-399 9.1* 300-399 21.5* Stage 4 400-499 13.0* 400-499 30.8* Stage 5 500-599 17.6* 500-599 41.9* Stage 6 600-699 26.4* 600-649 55.7* 650-699 71.3* Stage 7 700-800 60.1* 700-749 103.7* 750-800 196.2* Observations 456221 456221 BIC 107586.2 106739.4 AIC 107100.8 106221 Pseudo R2 0.303 0.309 Notes: * significant at 99-percent level; logistic regression with dependent variable = 1 if a fire occurs in a district on a given day and 0 otherwise; odds ratios for each drought stage express the factor by which the daily odds of fire detection increase compared to drought stage 0; KBDI categorical variables = 1 if the KBDI is in the shown range and 0 otherwise; models 1 and 2 also include controls for forest area per district, state-level fixed effects, and month of year (seasonal effects); drought stages are as proposed by Keetch and Byram (1968), with sub- stages introduced in model 2; observations are for January to June, 2012 to 2016 depicting the number of high-confidence detections A similar analysis was performed to explore the by MODIS for 2014-2016 and VIIRS for 2016 were relationship between fires and distance to the nearest overlaid on a surface representing the Euclidian settlement. The extent of built-up areas in India was distance of areas with forest cover to the nearest mapped using the Global Human Settlement Layer, road.116 which is derived from Landsat imagery for 2014 (EC JRC 2016). The results in figures A1.2 and A1.3 show most forested areas in India are within 2-3 kilometers The results for built-up area are shown in figures to the nearest road. Forested pixels in which active A1.4 and A1.5 below. The median distance to the fires were detected by MODIS or VIIRS tend to be nearest built-up area for forested pixels where a fire only slightly farther away from the nearest road. The was detected by MODIS and VIIRS was 7.9 km and median distance for forested pixels with fires detected 7.4 km, respectively, compared to 5.8 km and 5.9 by MODIS and VIIRS was 3.4 km and 3.8 km, km for forested pixels without any detected fires. By respectively, versus 2.9 km for forested pixels without performing Kolmogorov-Smirnov and nonparametric any fires. Kolmogorov-Smirnov and nonparametric K-sample tests, the null hypotheses that the K-sample tests confirm the statistical significance of distributions and medians of fire and no-fire pixels this disparity in the distributions and medians of fire can be confidently rejected. The results of these tests versus no-fire pixels. suggest that fires tend to occur in more rural areas (i.e., Data on India’s road network as of April 2017 are from Open Street Map and are provided by Geofabrik, a GIS consulting firm, at 116 http://download.geofabrik.de/asia/india.html. 151 Strengthening Forest Fire Management in India FIGURE A1.2: DISTRIBUTION OF FORESTED AREAS WITH AND WITHOUT FIRES BY DISTANCE TO NEAREST ROAD, USING MODIS DETECTIONS FOR 2014- 2016 20% 18% MODIS fire detections in forest 16% 14% Forest with no fires detected Percent of observations 12% 10% 8% 6% 4% 2% 0% 0 2 4 6 8 10 12 14 16 18 20 Distance to nearest road (km) Data sources: MODIS monthly data product for active fires (MCD14ML); Open Street Map data from Geofabrik; forest cover data from Hansen et al. (2013) FIGURE A1.3: DISTRIBUTION OF FORESTED AREAS WITH AND WITHOUT FIRES BY DISTANCE TO NEAREST ROAD, USING VIIRS DETECTIONS FOR 2016 20% 18% 16% VIIRS fire detections in forest 14% Forest with no fires detected Percent of observations 12% 10% 8% 6% 4% 2% 0% 0 2 4 6 8 10 12 14 16 18 20 Distance to nearest road (km) Data sources: VIIRS near-real-time active fire data product (VNP14IMGTDL_NRT); Open Street Map data from Geofabrik; forest cover data from Hansen et al. (2013) Strengthening Forest Fire Management in India   152 FIGURE A1.4: DISTRIBUTION OF FORESTED AREAS WITH AND WITHOUT FIRES BY DISTANCE TO NEAREST BUILT-UP SETTLEMENT, USING MODIS DETECTIONS FOR 2014-2016 10% 9% MODIS fire detections in forest 8% Forest with no fires detected 7% Percent of observations 6% 5% 4% 3% 2% 1% 0% 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Distance to nearest built - up area (km) Sources: MODIS monthly data product for active fires (MCD14ML); built-up area data from EC JRC (2016); forest cover data from Hansen et al. (2013) FIGURE A1.5: DISTRIBUTION OF FORESTED AREAS WITH AND WITHOUT FIRES BY DISTANCE TO NEAREST BUILT-UP SETTLEMENT, USING VIIRS DETECTIONS FOR 2016 10% 9% VIIRS fire detections in forest 8% Forest with no fires detected 7% Percent of observations 6% 5% 4% 3% 2% 1% 0% 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Distance to nearest built - up area (km) Data sources: VIIRS near-real-time active fire data product (VNP14IMGTDL_NRT); built-up area data from EC JRC (2016); forest cover data from Hansen et al. (2013) 153 Strengthening Forest Fire Management in India areas farther from built infrastructure) than would be respectively. The predominant forest types in areas expected if fires were just randomly distributed across above 1,000 m in these two states include subtropical areas with forest cover. As in the case with roads, there pine and montane moist temperate forests. In both are several possible explanations for this disparity. It states, fires typically occur at elevations between could be that forest fires in more rural areas are not 300 m and 2,000 m. The number of fires declines suppressed as quickly, and thus are more likely to be precipitously for areas above 2,000 m in both states. detected upon satellite overpass. It could also be that Fires below 1,000 m occur primarily in areas with moist people in more rural areas tend to rely more on fire as deciduous forest. Fires above 1,000 m occur mostly in a land management tool, for example, in stimulating subtropical pine and montane moist temperate forests. the growth of fresh grasses as fodder for livestock or to aid in collecting certain non-timber forest products. 2.3 Fuels The elevation profiles and terrain characteristics Average fire return intervals (FRI) for forests in of fire-affected forests were also examined. For this various regions and forest types were estimated by analysis the MODIS-derived MCD14ML data product comparing the number of times that forested areas provided by FSI was used, with further screening for burnt from 2003-2016, using the MODIS-derived high-confidence detections. Elevation and terrain data product for burnt area (MCD45A1) overlaid data were derived from the 90-m (3 arc-second) void- on data for the extent of forest cover in 2000 from filled digital elevation model from the Shuttle Radar Hansen et al. (2013). Forest type data are from Reddy Topography Missions (SRTM).117 The elevation and et al. (2015).118 Because of the short time frame of terrain profile of fire detections during January-June the MODIS data, these estimates of FRI are highly from 2003-2016 were analyzed. tentative and should be checked against longer-term historical data where available. Terrain ruggedness scores for forested pixels with MODIS fire detections were calculated following Riley et al. (1999). Scores of around 100 or less indicate 3. HUMAN-CAUSED FOREST level or nearly level ground; scores of around 100-250 FIRES, AND SOCIAL FACTORS indicate gently hilly terrain; scores of 250 to 500 are for moderately rugged terrain, and scores above 500 INFLUENCING FIRES are for highly rugged mountainous terrain. About 90 percent of forest fires detected by MODIS in India MODIS fire detections were overlaid with district-level occurred at elevations below 1,200 m. poverty data to test whether a spatial correlation exists between areas with more forest fires and a higher More than half of all detected fires occurred in areas incidence of poverty. where the terrain was moderately or highly rugged. States in which forest fires tended to be observed in A variety of statistical tests were performed to compare the highest and most rugged areas include Himachal fire density (number of fires per 100 km2 of forested Pradesh, Jammu and Kashmir, Manipur, Nagaland, area) between poorer and better-off districts. The Tamil Nadu, and Uttarakhand. The ruggedness of analysis focused on rural forest districts, defined as fire-affected areas in these states presents a challenge those districts with a population density of less than for effective fire suppression. 1,000 people per km¬2 and with at least 10 percent of the total area under forest cover in 2000, per the Himachal Pradesh and Uttarakhand are among states Hansen et al. (2013) forest data. The number of fire that have large areas of forest in rugged terrain at detections per district for each year from 2003 to elevations above 1,000 m. Forests at 1,000 m or higher 2016 was normalized in terms of detections per km2 accounted for 68 percent and 72 percent of forest of forest area in the district (assuming forest cover cover in 2000 in Himachal Pradesh and Uttarakhand, for 2000). Districts were then sorted into quantiles See USGS, “Shuttle Radar Topography Mission (SRTM),” https://lta.cr.usgs.gov/SRTM. 117. Forest type data are available from the National Remote Sensing Centre (NRSC), Bhuvan, http://bhuvan.nrsc.gov.in/. 118. Strengthening Forest Fire Management in India   154 by poverty headcount ratio, which is the percent of Further regression analysis was done to see if the the population in a district living below the national observed correlation between higher poverty rates poverty line. Poverty data were for 2011, based on and forest fire density holds up if other environmental household survey results.119 and social factors are also considered. Average fire density (MODIS detections per km2 forest in the Differences in the average fire density were first district) for 2009-2013 was regressed on the poverty tested for two groups (those above and below the headcount ratio for 2011 and a variety of control 50th percentile for poverty headcount ratio). T variables that were found to be relevant in influencing tests for equal and unequal variances were done to fire potential. Descriptive statistics for included test for the equality of means in the number of fire variables are provided in table A1.12 below. detections between the two groups. Non-parametric Mann-Whitney U tests were also performed, relaxing District-level fire density was related to the poverty the assumption that the number of fire detections is rate as: normally distributed. From the results of these tests, the hypothesis that fire density is the same in poorer and less-poor districts can be confidently rejected at the 99-percent level. where Fire is average fire density, Pov is the poverty headcount ratio, Pop is population density in Districts were then grouped into quartiles by poverty 2011, Tmp is average temperature during the fire headcount ratio. Districts with the lowest poverty rates season (January-June) for 2009-2013, Pre is average were sorted into quartile 1, while districts with the precipitation during the fire season for 2009-2013, highest rates were grouped into quartile 4. To compare Pret-1 is average precipitation during July-December fire density across the poverty quartiles, one-way ANOVA of the prior year, F is a binary dummy variable equal tests were performed, weighting and unweighting the to 1 if the predominant forest type in the district sample of districts by total forest-covered area. From (the forest type with the greatest area) is type f and these tests, too, the hypothesis that districts in the 0 otherwise,120 R is a binary dummy variable equal to different poverty quartiles experience the same number 1 if the district is in region r and 0 otherwise, and ε is of forest fires per unit area of forest can be rejected at an error term.121 Observations were weighted by the the 99-percent level. Thus, higher rates of poverty are district’s area of forest cover. Results are presented in stronger correlated with higher rates of forest fires. table A1.13 below. TABLE A1.12: DESCRIPTIVE STATISTICS FOR DISTRICT-LEVEL FOREST FIRE DENSITY AND POVERTY RATES, 2009-13 Variables Obs Mean Minimum Maximum Std dev Fire detections (count), 2009-2013 436 288.0688 0 5311 635.3861 Forest cover in 2000 (km2) 436 1024.915 0.008326 8629.139 1434.03 Fires per km2 forest 436 27.95653 0 1696.843 125.8306 Poverty headcount ratio, 2011 436 24.55459 0 78.6 17.42344 Population density, 2011 436 429.1292 1.126142 24968.25 1206.568 Average temperature, Jan-Jun, 2009-13 436 24.63276 -3.64333 30.3425 6.010288 Average precipitation, Jan-Jun, 2009-13 436 360.3335 40.67571 1881.9 323.5565 Average precipitation, Jul-Dec, previous 436 890.337 163.6829 2308.62 353.8306 year Source: MCD14ML data product, provided by FSI; forest cover from Hansen et al. (2013); World Bank subnational poverty and population data; weather data from CRU TS v4.00, https://crudata.uea.ac.uk/cru/data/hrg/ 119. World Bank, “World Bank subnational poverty data,” unpublished data set compiled by the Environment and Natural Resources and Poverty Global Practice Groups, December 2017. 120. Forest type data are from Reddy et al. (2015). 121. R accounts for unexplained differences in environmental and social characteristics that could affect fires. 155 Strengthening Forest Fire Management in India TABLE A1.13: REGRESSION RESULTS FOR FOREST FIRE DENSITY AND POVERTY RATES, 2009-2013 Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Outcome variable ln (fire detections per sq. km forest cover) Social/economic factors Poverty headcount ratio 0.0269*** 0.0267*** 0.0176** 0.0180** 0.00745 0.00101 Population density -0.0231 -0.332 -0.338 -0.544** -0.431* Weather conditions Temperature, Jan-Jun 0.0863** 0.124*** 0.0168 0.120*** Precipitation, Jan-Jun 0.707 0.433 -0.602 Precipitation, Jul-Dec, -1.053 0.453 1.259 year t-1 Dominant forest types Wet evergreen 0 0 Semi-evergreen 1.405*** 0.830*** Moist deciduous 1.712*** 0.836*** Dry deciduous 2.368*** 1.689*** Littoral/swamp/mangrove 0.127 -0.594 Thorn 4.481*** 3.124*** Subtropical broadleaf 0.128 -0.136 Subtropical pine 0.728* 0.458* Montane moist temperate -0.846 -0.598 Montane dry temperate -2.345*** -1.004 Sub-alpine -0.473 -0.423 Regions Central 0 North 1.525*** Northeast 1.838*** South -0.888*** West -0.0042 W. Himalaya 1.799*** Constant 0.299 0.417 0.241 2.353 -3.958 -5.964 Observations 393 393 393 393 393 393 BIC 1340.1 1345.8 1302.2 1302.7 1254.3 1152.9 AIC 1332.1 1333.9 1286.3 1278.8 1194.7 1077.4 Adjusted R2 0.138 0.136 0.237 0.255 0.41 0.565 Notes: * p < .10, ** p < .05, *** p < .01; coefficients are changes in ln (fire density) per unit increase in the variable Strengthening Forest Fire Management in India   156 ANNEX 2 SURVEY OF STATE FOREST DEPARTMENT OFFICERS An online survey of forest department staff was World Bank team, each nodal officer was asked to designed by the World Bank team to gather information select up to 17 forest department officers per state on forest fire prevention and management (FFPM) in to take part in the survey. These 17 respondents various Indian states. Issues of focus covered by the were to include 2 higher-level officers in the state survey included: forest department headquarters at the rank of Chief • Causes and characteristics of forest fires; Conservator of Forest (CCF) or above and up to • Plans, policies, and procedures for FFPM 15 territorial officers from fire-affected circles and implemented by the forest department; divisions at the rank of Conservator of Forests (CF) or • Coordination with other public agencies and below. Nodal officers were provided with two options departments on FFPM; and for the section of territorial/field-level respondents • The role of the local community in FFPM and and given independence in choosing which option to avenues for improving public engagement. implement: Two different versions of the survey were tailored for 1. Department selection: The nodal officer was asked different categories of respondents: (A) higher-level to identify the 5 forest divisions in that state that officers in the state forest department headquarters, have experienced the greatest number of fires and (B) territorial forest officers working at the circle during the previous 5-10 years. The nodal officer level or below. was then asked to identify the Divisional Forest Officers (DFOs) responsible for these divisions, The survey was created using SurveyMonkey, an online at least one Range Officer (RO) or Deputy Range survey platform. Officers could complete the survey Officer (DRO) under the DFO, and at least one using a computer, tablet, or smart phone. Respondents officer at the CF or Assistant CF level in the chain who had difficulty connecting to the SurveyMonkey of command above the DFO. website also had the option of responding to the survey by filling out a Word document and emailing it 2. Random selection: The nodal officer could also to the World Bank team. choose from 5 fire-affected districts selected at random in each state. A list of randomly selected A. SAMPLING districts was provided to each nodal officer by the World Bank team. The probability that a district A sample of 11 states was identified by the World was selected was weighted by the total number of Bank team in consultation with the Ministry of satellite-detected fire locations reported by Forest Environment, Forest and Climate Change (MoEFCC). Survey of India (FSI) for that district. Thus, a The selection of states took into consideration forest district where fire occurs twice as frequently as area, frequency and extent of fires in recent years, and another was twice as likely to be selected in the the requirements of the ministry. sample. Because forest divisions do not perfectly overlap with districts, and many states have not In each of these states, MoEFCC appointed a nodal digitized the boundaries of their forest divisions, officer to assist the World Bank team with the collection the nodal officer was asked to assist the team in of data. The nodal officer was tasked with identifying identifying 5 forest divisions that roughly intersect respondents in that state, disseminating the survey, with those randomly-chosen districts. The process and assisting the World Bank team in following up of further selecting CCFs/CFs and field staff was with respondents. Per instructions provided by the then the same as in the first option above. 157 Strengthening Forest Fire Management in India For each of these options, the rationale behind Any answers you provide will be anonymous. Your choosing CCFs/CFs, DFOs, and field staff working identifying information will not be shared with in the same area was to provide some verification of anyone outside the team of World Bank researchers trends and practices noted by individual respondents completing this study. for that area. The selection of territorial officers at different working level could also provide insight into For any other suggestions, questions, or comments, how certain challenges are viewed by staff at those please email [...] different levels. Thank you for your participation! Altogether, the sampling plan for the survey aimed to provide at least 180 possible respondents in total Basic information [For sorting respondents to the from the 11 states. The number of completed surveys correct version of the survey] expected to be received from these respondents was 100-130, or at least 10 completed surveys per state. *1. Please select your State or Union Territory. *2. What is your designation in the forest department? B. IMPLEMENTATION A. PCCF [Skip to 4] The survey was designed in February and March B. Addl. PCCF [Skip to 4] 2017 and tested with 5 current and former forest C. CCF [Skip to 4] department officers in Uttarakhand in early April D. CF [Skip to 48] 2017. Nodal officers were identified, and survey instructions were sent to the states by MoEFCC the E. Addl. CF [Skip to 48] same month. The first online survey responses F. DFO [Skip to 48] were received in May 2017. The survey was closed in August 2017. Altogether, 101 useable responses G. SDO [Skip to 48] were received, including 92 online responses and 9 H. RO [Skip to 48] completed survey forms emailed to the World Bank team. I. DYRO [Skip to 48] J. Forester [Skip to 48] C. SCRIPT OF ONLINE SURVEY K. Forest Guard [Skip to 48] L. Other [Skip to 3] Welcome page *3. Are you a territorial officer working in the field at the circle level or below? This survey will collect information about the prevention, occurrence, and management of forest A. Yes [Skip to 4] fires in your area. The questions are part of a study B. No [Skip to 48] by the World Bank for the Ministry of Environment, Forest and Climate Change. *47. In which circle/division do you work? [Skip to 48] Your input is very important to the findings and recommendations of the study. We appreciate your time and thought in responding. Strengthening Forest Fire Management in India   158 SURVEY VERSION A: State forest department – PCCF, Addl. PCCF, or CCF Section 1: Framing questions and general causes of fire 4. How much of a concern are forest fires for your state? Not at all Slightly concerning Moderately Very concerning Extremely concerning concerning concerning 5. How important is the role of the forest department in managing forest fires? Not at all important Slightly important Moderately Very important Extremely important important 6. How important are other government agencies in managing forest fires? Not at all important Slightly important Moderately Very important Extremely important important 7. How important is the local community in managing forest fires? Not at all important Slightly important Moderately Very important Extremely important important 8. About what percent of fires in your state would you say are caused by natural versus human sources? (Please enter a whole number for each, rounding to the nearest percent. Natural sources might include lightning, friction, etc. Human sources would include any accidental, negligent, deliberate, or other use of fire.) 9. About what percent of fires in your state would you say are caused by known versus unknown sources of ignition? (Please enter a whole number for each, rounding to the nearest percent.) 10. In your view, what are the 6 most common causes of forest fires in your state? Please rank in order, beginning with the top cause. (For this question, causes refer to the source of ignition. If fewer than 6 are applicable, please enter “NA”.) Section 2: Legal issues 11. Are there any purposes for which burning in forest areas under the forest department is permitted? If so, are these stipulated in the forest working plans or management plans? [Comment box] 12. What restrictions exist for burning in areas classified as forest but which are not under the forest department (e.g., communal or revenue forest)? [Comment box] 13. What restrictions exist for burning on agricultural lands adjoining forests? [Comment box] 159 Strengthening Forest Fire Management in India Section 3: Fire prevention and preparedness 14. How would you rate the following? Very poor Poor Somewhat Fair Somewhat Good Very good poor good Overall level of prevention and preparedness for forest fires in your state Effectiveness of early warning or fire danger rating systems currently used 15. To the best of your knowledge, are all the fire lines stipulated in the working plans for the forest department in your state currently cleared and maintained? Yes [Skip to 18] No 16. About what portion of the fire lines under the forest department in your state are currently maintained and functional? None Half All 17. Please comment on why some fire lines in areas manage by the forest department are not cleared or maintained. [Comment box] 18. Excluding fire lines, is controlled burning required on any other forest areas managed by the forest department? Yes No [Skip to 22] 19. Is controlled burning done annually on all the areas under the forest department where required? Yes [Skip to 22] No Strengthening Forest Fire Management in India   160 20. About what percentage of the annual area prescribed by working plans or other management plans for controlled burning is actually treated? None Half All 21. Please comment on why controlled burning is not performed on some of these areas as required. [Comment box] 22. In your state, are there any forested lands that are not managed by the forest department? Yes No [Skip to 26] 23. Who manages these other forested lands (those not under the forest department)? [Comment box] 24. Are these other forested lands covered by working plans or similar planning documents? Yes No 25. What fire prevention measures are required for these other forested areas (e.g., fire lines or controlled burning)? And what role does the forest department have in fire prevention for these areas? [Comment box] 26. In your view, what are the biggest challenges to the effective prevention of forest fires in your state? [Comment box] Section 4: Public engagement 27. How would you rate the following? Very poor Poor Somewhat Fair Somewhat Good Very good poor good Effectiveness of communication to the public in your state about the danger or likelihood of fire Effectiveness of local Community engagement in your state in preventing forest fires 161 Strengthening Forest Fire Management in India 28. What are some ways in which the forest department engages with the local community in forest fire prevention? [Comment box] 29. How can engagement with the local community on forest fires be improved? [Comment box] Section 5: Fire response 30. What are the main techniques used for suppression of unwanted forest fires in your state? [Comment box] 31. What equipment is typically used for fire suppression in your state? 32. Is safety equipment provided to field staff for fire suppression (special clothing, boots, helmets, etc.)? [Comment box] 33. Is equipment for fire suppression adequate and sufficiently available? Yes [Skip to 35] No 34. What additional equipment is needed? [Comment box] 35. Are there multiple agencies ever involved in responding to forest fires in your state? If so, how are they coordinated? Who determines the coordination processes? [Comment box] 36. How are fires that cross jurisdictional boundaries managed? In these cases, who funds suppression activity? [Comment box] 37. What reporting is required from field staff if a forest fire occurs in your area? What information does the report contain? And to what office or person is the report sent? [Comment box] 38. To what extent are the causes of forest fires in your state investigated? How is this done? [Comment box] Section 6: Fire recovery 39. Is there any formal process to assess impact and commence recovery operations? Who does it? How is it funded? [Comment box] 40. Do communities receive any assistance in restoration of their losses after fires occur? [Comment box] Section 7: Research 41. Is any research undertaken about impact of unwanted fire in your state? If so, who does this? Who funds it? [Comment box] 42. Is there any scientific research that has been done or is currently being done on how fires behave that can aid fire predictions for your state? If so, can you please describe it (e.g. fire danger rating systems, drought indices, fuel accumulation in different forest types)? [Comment box] Strengthening Forest Fire Management in India   162 43. Has there been any research or evaluation of the efficacy of prevention programs? If yes, what are the conclusions? [Comment box] Section 8: Wrapping up 44. Do you have any other comments, questions, or suggestions that have not been covered? [Comment box] 45. May we contact you if we have any other questions about forest fires in your state? Yes No [End of survey] 46. Please provide your contact information. [Name, email, phone] Your name and contact information will not be used for any other purpose or shared with anyone outside the team of World Bank researchers completing the assessment without your consent. Should you have any questions or concerns, please contact the study team at [...] SURVEY VERSION B: State forest department (territorial officers working at circle level or below) Section 1: Framing questions and general causes of fire Note: Questions that ask about “your area” refer to the specific territory for which you are responsible (circle, division, or range). 48. How much of a concern are forest fires for your area? Not at all Slightly concerning Moderately Very concerning Extremely concerning concerning concerning 49. How important is the role of the forest department in managing forest fires in your area? Not at all important Slightly important Moderately Very important Extremely important important 50. How important are other government agencies in managing forest fires in your area? Not at all important Slightly important Moderately Very important Extremely important important 51. How important is the local community in managing forest fires in your area? Not at all important Slightly important Moderately Very important Extremely important important 52. About what percent of fires in your area would you say are caused by natural versus human sources? (Please enter a whole number for each, rounding to the nearest percent. Natural sources might include lightning, friction, etc. Human sources would include any accidental, negligent, deliberate, or other use of fire.) 163 Strengthening Forest Fire Management in India 53. About what percent of fires in your area would you say are caused by known versus unknown sources of ignition? (Please enter a whole number for each, rounding to the nearest percent.) 54. In your view, what are the 6 most common causes of forest fires in your area? Please rank in order, beginning with the top cause. (For this question, causes refer to the source of ignition. If fewer than 6 are applicable, please enter “NA.”) Section 2: Fire prevention and preparedness 55. How would you rate the following? Very poor Poor Somewhat Fair Somewhat Good Very good poor good Overall level of prevention and preparedness for forest fires in your area Effectiveness of early warning or fire danger rating systems currently used 56. Are all the fire lines stipulated in the working plan of the forest department for your area currently cleared and maintained? Yes [Skip to 59] No 57. About portion of the fire lines under the forest department in your area are currently maintained and functional? None Half All 58. Please comment on why some fire lines managed by the forest department in your area are not maintained or clear. [Comment box] 59. Excluding fire lines, is controlled burning required on any other forest area under the forest department? Yes No [Skip to 63] Strengthening Forest Fire Management in India   164 60. Is controlled burning done annually on all the areas under the forest department for which it is required? Yes [Skip to 63] No 61. What portion of the annual area prescribed by working plans or other management plans for controlled burning is actually treated? None Half All 62. Why is controlled burning not performed on all the required areas? [Comment box] 63. In your area, are there any forested lands that are not managed by the forest department? Yes [Skip to 67] No 64. Who manages these forested lands (those not under the forest department)? [Comment box] 65. Are these other forested lands covered by a working plan or similar planning document? Yes No 66. What fire prevention measures are required for these other forested areas (e.g., fire lines or controlled burning)? And what role does the forest department have in fire prevention for these areas? [Comment box] Section 3: Specific uses and causes of fire 67. Do people in your area ever graze their animals in the forest or collect fodder from the forest? Yes No [Skip to 69] 68. Do they use fire to promote the growth of grass and fodder? Yes No 165 Strengthening Forest Fire Management in India 69. Do people in your division or area use fire in gathering any non-timber forest products (NTFPs) from the forest? Yes No [Skip to 73] 70. What NTFPs do they collect by using fire or burning? When, how, and why is the burning done? [Comment box] 71. What is done before burning to make sure the fire does not spread? Is the forest department required to be onsite to supervise? [Comment box] 72. Do people in your area do burning in the forest for any other reason? Yes No [Skip to 74] 73. What are some other reasons why local people in your area set fire in the forest? [Comment box] 74. Are there any other restrictions on where, when or how local people may do burning in the forest in your area? Yes No [Skip to 76] 75. What other restrictions are there? [Comment box] 76. Are escapes of agricultural fires set on adjoining lands a cause of forest fires in your area? Yes No 77. In your view, what are the biggest challenges to the effective prevention of forest fires in your area? [Comment box] Strengthening Forest Fire Management in India   166 Section 4: Community engagement 78. How would you rate the following? Very poor Poor Somewhat Fair Somewhat Good Very good poor good Effectiveness of communication to the local community in your area about the danger or likelihood of fire Effectiveness of engagement with the local community in your area in preventing forest fires 79. What are some ways in which the forest department engages with the local community in your area in preventing and managing forest fires? [Comment box] 80. What are some ways that people in the community can become more effectively involved in managing forest fires in your area? [Comment box] Section 5: Fire response 81. What are the main techniques used for the suppression of unwanted fires in your area? [Comment box] 82. What equipment is most typically used for fire suppression in your area? 83. Is safety equipment provided to field staff for fire suppression (special clothing, boots, helmets, etc.)? Yes No 84. Is equipment for fire suppression in your area adequate and sufficiently available? Yes [Skip to 86] No 85. What additional equipment is needed? [Comment box] 86. Does the forest department maintain any fire watchtowers or crew stations in your area? Yes No [Skip to 90] 167 Strengthening Forest Fire Management in India 87. How many watchtowers or crew stations are there? [Text box] 88. Are they all functioning properly? Yes No 89. Are additional watchtowers or crew stations needed? Yes No 90. Does the forest department employ any seasonal fire watchers from the local community in your area? Yes No [Skip to 98] 91. How many have been employed this fire season? [Text box] 92. Are more fire watchers needed in your area? Yes No 93. Do fire watchers receive any equipment or training from the forest department? Yes No 94. Is additional equipment or training for fire watchers needed? Yes No [Skip to 96] 95. What additional equipment or training for fire watchers is needed? 96. Are seasonal fire watchers provided payment for their services? Yes No [Skip to 98] Strengthening Forest Fire Management in India   168 97. Were there any delays or shortages of funding last year that prevented fire watchers from being paid in full and on time? Yes No 98. What reporting is required from field staff if a forest fire occurs in your area? What information does the report contain? And to what office or person is the report sent? [Comment box] 99. To what extent are the causes of forest fires in your area investigated? How is this done? [Comment box] 100.Do communities receive any assistance in restoration of their losses after fires occur? [Comment] Section 6: Wrapping up 101. Do you have any other comments, questions, or suggestions that have not been covered? [Comment box] 102. May we contact you if we have any more questions about forest fires in your area? Yes No [End of survey] 103. Please provide your contact information. Your name and contact information will not be used for any other purpose or shared with anyone outside the team of World Bank researchers completing the assessment without your consent. Should you have any questions or concerns, please contact the study team at [...] 169 Strengthening Forest Fire Management in India STATE-WISE TABLES OF SURVEY RESULTS TABLE A2.1: BIGGEST CHALLENGES TO EFFECTIVE FOREST FIRE PREVENTION IDENTIFIED BY RESPONDING OFFICERS IN EACH STATE Uttara- Tripura Telangana Odisha Meghalaya Madhya Kerala Jharkhand Himachal Chhattis- Assam khand Pradesh Pradesh garh Public engagement 2 1 1 1 3 1 5 1 (non-specific) Lack of awareness/ 8 1 3 4 1 3 4 3 5 2 education Traditional practices 1 7 4 3 2 2 2 2 1 of local community Conflict or animosity 2 2 2 1 toward forest depart- ment Negligence 1 Missing or perverse 1 1 1 incentives Environmental factors (non-specific) Inaccessibility and 2 2 1 1 5 1 5 1 3 difficult terrain Weather or climate 1 1 conditions Forest structure or 3 1 1 3 1 species composition Invasive species 1 Water availability 2 Insufficient financial 4 1 1 2 1 3 1 1 resources Labor shortage 5 1 1 7 3 1 Lacking equipment, 2 3 1 4 2 3 6 1 tech, infrastructure Knowledge gaps (ex- 1 1 3 pertise, training, etc.) Poor coordination 2 1 1 2 1 1 between agencies Strengthening Forest Fire Management in India   170 Uttara- Tripura Telangana Odisha Meghalaya Madhya Kerala Jharkhand Himachal Chhattis- Assam 171 khand Pradesh Pradesh garh Problems with poli- 1 2 2 1 cies, laws, regulations Weak implementa- 1 1 2 tion of existing plans/ policies Lack of management 1 plan Low priority given to 1 1 fire prevention Illegal/criminal activity 2 Illegal felling 1 Land encroachment 1 1 Poverty 1 Lack of agricultural 1 land/capital Lack of alternative 1 livelihoods TABLE A2.2: PRINCIPAL TECHNIQUES USED TO SUPPRESS FOREST FIRES Uttara- Tripura Telangana Odisha Meghalaya Madhya Kerala Jharkhand Himachal Chhattis- Assam khand Pradesh Pradesh garh Dousing 2 1 6 3 1 Manual beating or 12 3 4 7 10 2 11 5 11 4 6 smothering Surface clearing 5 5 1 3 1 3 Use of fire in control 9 0 1 0 3 0 6 5 9 2 0 Creation of fire lines 4 3 3 4 2 2 3 4 2 5 Strengthening Forest Fire Management in India TABLE A2.3: ADDITIONAL EQUIPMENT NEEDS MENTIONED BY OFFICERS Uttara- Tripura Telangana Odisha Meghalaya Madhya Kerala Jharkhand Himachal Chhattis- Assam khand Pradesh Pradesh garh Other manual hand 1 1 tools Beaters 1 1 3 1 Rake 2 1 1 1 1 Axe 1 Cutting tools 1 1 Spade 1 Leaf blower 2 1 2 7 1 3 4 1 Dousing equipment 1 (non-specific) Lightweight fire 1 3 1 1 1 1 3 1 1 2 extinguishers High-power sprayers 1 Water tanker (truck 1 1 1 1 or tractor) Bucket 1 Hydrant accessories 1 Portable sprayers 2 1 Chemical retardants 1 Other equipment 2 1 1 1 1 1 2 3 2 1 and clothing Gloves 1 1 1 1 1 1 Fire-resistant uni- 3 5 3 3 4 1 3 2 5 1 form Boots 2 5 1 4 3 1 3 2 Helmet 3 4 2 2 1 2 3 2 Dust mask 1 First aid kit 1 1 Torch/flashlight 1 1 2 Protective eyewear 1 1 1 1 Strengthening Forest Fire Management in India   172 Uttara- Tripura Telangana Odisha Meghalaya Madhya Kerala Jharkhand Himachal Chhattis- Assam 173 khand Pradesh Pradesh garh Transport vehicles 4 3 1 1 3 1 2 3 Tractor 1 Fire engine 1 1 Helicopter or other 2 aircraft Communications 1 2 1 1 devices or GPS Food and water 1 1 Nothing 2 1 1 1 1 Additional manpow- 1 er more useful than equipment Simple firefighting tools 1 in sufficient quantity Light equipment that is 1 1 easy to carry Drones 1 Strengthening Forest Fire Management in India TABLE A2.4: METHODS OF COMMUNITY ENGAGEMENT USED BY THE FOREST DEPARTMENT Assam Chhattis- Himachal Jharkhand Kerala Madhya Meghalaya Odisha Telangana Tripura Uttara- garh Pradesh Pradesh khand Public awareness 2 1 4 2 7 2 3 2 2 raising (non-specific) Public gatherings 2 1 3 and performances Public announce- 2 2 1 1 2 2 3 ments School programs 1 1 2 Meetings and work- 3 2 1 5 shops Joint Forest Manage- 3 4 4 7 4 2 5 6 4 6 ment Committees Van Panchayat 2 Gram Panchayat or 2 Gram Sabha Other community 3 2 1 1 4 1 2 institutions Women's groups 3 Firewatchers 1 3 1 7 1 1 4 2 5 5 Wages for clearing 2 2 1 1 1 1 fire lines Other employment 2 2 1 1 1 2 Monitoring, pa- 1 trolling, or policing Fire response 1 Provision of incen- 1 1 2 1 1 tives (not specified) Rights and conces- 1 sions Prizes 1 Development proj- 3 1 1 ects/programs Monetary payments 1 2 for prevention Strengthening Forest Fire Management in India   174 Assam Chhattis- Himachal Jharkhand Kerala Madhya Meghalaya Odisha Telangana Tripura Uttara- 175 garh Pradesh Pradesh khand Fire lines and con- 1 1 trolled burning with community Provision of equip- 2 1 1 ment for firefighting Involvement in 1 management and planning TABLE A2.5: HOW CAN ENGAGEMENT WITH THE LOCAL COMMUNITY BE IMPROVED? Assam Chhattis- Himachal Jharkhand Kerala Madhya Meghalaya Odisha Telangana Tripura Uttara- garh Pradesh Pradesh khand More of the same 1 Public awareness 2 3 5 5 2 2 5 2 2 4 2 raising programs Work more with 1 2 2 2 2 2 3 community institu- tions Employment 2 2 5 1 2 2 3 Monitoring, pa- 1 trolling, or policing Provision of incen- 1 3 9 4 5 2 6 5 3 6 3 tives Joint implementation 1 1 1 of FFPM Response unclear or 1 2 unspecific Broader engagement 1 More systematic 1 1 programming More focus on the 1 1 2 youth Address human-wild- 1 life conflicts Policy changes 1 1 3 Strengthening Forest Fire Management in India ANNEX 3 COMMUNITY CONSULTATIONS AND CASE STUDIES Structured community appraisals involving site sampled districts were chosen to include all the three visits, interviews, and focus group discussions were indigenous tribes of Meghalaya (Khasi, Jaintia, and performed in Meghalaya and Uttarakhand in Garo) and to represent the various fire management August 2017. Additional consultations and field visits approaches being practiced by these groups. Further, were performed with forest-using communities in the selected districts also cover various forest types Jharkhand, Madhya Pradesh, Meghalaya, Odisha, found in the state, including clan forest, community Telangana, and Uttarakhand in January-May 2017 reserves, sacred groves, mining affected areas, areas of to gather community members’ views on the causes, jhum cultivation, and a REDD+ project area. prevention, and management of forest fires. The in- depth appraisals in Meghalaya and Uttarakhand are Uttarakhand described below. Case studies from the Meghalaya The community appraisal focused on 10 villages in 3 appraisal are also provided. districts that were affected by forest fires during 2015 and 2016. The appraisal attempted to cover majority Meghalaya of fire affected forest types in the state of Uttarakhand and a variety of community institutions, such as Interviews and focus group discussions were Van Panchayat, Mahilla mandals, JFMCs, or other performed with 41 respondents in 5 districts, including institutions. The appraisal also covered communities East Khasi Hills, West Jaintia Hills, Ri-Bhoi, North residing in the periphery of protected areas, including Garo Hills and West Garo Hills (figure A3.1). The biosphere reserves and wildlife sanctuaries. FIGURE A3.1: DISTRICTS VISITED IN MEGHALAYA FOR THE COMMUNITY APPRAISAL Tikrikilla Umling North Garo Hills Ri Bhoi Resubelpara Kharkutta Jirang Dadenggre Selsella Umsning Songsak Rongjeng Mairang West Garo Hills East Garo Hills Thadiaskein Rongram Mawshynrut West Khasi Hills Mylliem Betasing Mawryngkneng Laskein Samanda Nongstoin Mawthadraishan South West Garo Hills Mawphiang West Jaintia Hills Gambegre Laitkroh Mawkynrew Zikzak Chokpot Mawkyrwat East Khasi Hills Baghmara Dalu South Garo Hills South West Khasi Hills Saipung Gasuapara Rongra Ranikor Pynursia Amiarem Mawsynram Khliehriat Shella Bholaganj East Jaintia Hills Strengthening Forest Fire Management in India   176 CASE STUDY 1: MAWPHLANG (MEGHALAYA) COMMUNITY BASED REDD+ PROJECT Renowned globally for its sacred grove, Mawphlang FIGURE A3.3: REDD+ PROJECT AREA IN is one of the key hubs of Khasi culture in the state. MAWPHLANG The block is located approximately 25km from the State capital, Shillong and is owned by 184 villages.122 Mawphlang’s Khasi Heritage Villages along with sacred grove are a key tourist attraction in the State. A key achievement of Mawphlang is its REDD+ project or Reducing Emission from Deforestation and (Forest) Degradation, which is a mechanism under which communities can earn income through carbon credits. The project is being implemented by a consortium of the 10 Himas123 in the region, the “Ka Synjuk Ki Hima Arliang Wah Umiam Mawphlang Welfare Society”. The project aims to conserve the forest areas in the region, including the sacred groves and water sheds, and to increase tree cover in the surrounding areas. The project is spread over an area of 27,000 hectares covering 10 ‘Hima’s or local governments and 62 villages. FIGURE A3.4: VIEW OF THE REDD+ FIGURE A3.2: MAWPHLANG SACRED PROJECT AREA UNDER GROVES MAWPHLANG BLOCK 2011 Census 122 Region/ kingdom with local governance in Khasi hill governed by a traditional leader 123 177 Strengthening Forest Fire Management in India Additionally, the project seeks to provide sustainable FIGURE A3.5: CHARCOAL MAKING AS alternatives and solutions to current practices that are ONE OF THE LIVELIHOOD leading to degradation of forests, land and water. As ACTIVITIES IN on date, more than 80 thousand tonnes equivalent MAWPHLANG of carbon credits have been generated and sold to countries in Europe such as Italy, Sweden and Belgium, and USA. Each carbon credit is sold between USD $5 to 6. The project is expected to mitigate 3,18,427 tonnes of carbon dioxide between 2010 and 2021. Apart from carbon credits, the project has also brought significant changes to the socio-economic and ecological condition of the entire region. There has been significant increase in wildlife in the region because of the project. Certain species of fauna that were thought to have been extinct in the region have been rediscovered and are recovering. There has also been an increase in the amount and variety of extractable NTFPs which has provided local residents with a source of income. FIGURE A3.6: COMMUNITY MEMBERS CREATING A FIRE LINE IN THE REDD+ PROJECT AREA Strengthening Forest Fire Management in India   178 Fire management under REDD+ project: The forest community events under the project. Every household in the region traditionally comprises of broadleaf participates in the activity guided by coordinators tree species but have been invaded by Khasi pine from the consortium. Planning is done by the project that occupies large tracts of forest land. This species team following standard state forestry norms as is highly flammable, when dry making the forest issued by the Government. The project arranges for vulnerable to forest fires during the dry seasons. refreshments which are served to mark the end of the operation. As on August 2017, 27 fire lines measuring While certain agricultural practices such as the locally 88.5 kilometers have been created. practiced Bun Cultivation124, and charcoal making can sometimes lead to forest fires, the biggest cause To reduce fuel wood collection from forest, adoption is still man made, accidental or intentional ignition of fuel efficient stoves is promoted and supported of dry forest matter. The Joint Forest Management by the project. Quarrying has been banned in the Committees that were constituted by the Forest project area. These have reduced the amount of land Department in the region to manage forest in the degradation to a large extent in the project area. area are non-functional. However, the community engages volunteers under the project to patrol key In terms of socio-economic interventions, the project project areas to reduce intentional ignition of forest has taken several initiatives to spread awareness on by miscreants. Between 2010 and 2016, forest fires fire prevention and safety. For community members have devastated about 488 hectares of land. A total that are dependent on livelihood activities with high of 16 most vulnerable fire points responsible for risk of fire hazard (bun cultivation, charcoal making 80percent of fire incidences have been identified and fire management interventions are implemented. etc.), the project advocates for alternative means of livelihood and facilitates people with capacity building Within the project area, fire is managed with the help and support for establishing a new livelihood activity. of fire lines. Controlled pre-burning is avoided due Several alternative livelihood activities have been to the high slopes. Fire lines are made by community introduced including poultry and livestock rearing, members twice in a year through participatory charcoal briquette making, home-based nursery, etc. FIGURE A3.8: IMPROVED FUEL WOOD STOVES TO REDUCE FIGURE A3.7: FIRE LINES IN PROJECT CONSUMPTION AREA Bun cultivation is a traditional process of anaerobic burning of dry nitrogenous plant matter beneath a thin layer of soil to release 124 nutrients into the soil. Unlike Jhum cultivation, here the same plot of land can again be reused season after season 179 Strengthening Forest Fire Management in India FIGURE A3.9: DISTRIBUTION OF LPG FIGURE A3.10: AWARENESS PROGRAMME CONNECTION TO REDUCE ON FOREST FIRE FUEL-WOOD COLLECTION PROTECTION MEASURES FROM FORESTS A Community Development Grant has been set aside FIGURE A3.11: AREA AFFECTED BY from the revenue earned from the sale of carbon FOREST FIRES IN HA. credits. This is used to fund various other development 20.0 activities including distribution of cookers and Liquefied Petroleum Gas (LPG), installation of poly houses, distribution of farm inputs such as seedlings, 9.1 saplings, piglets, chicks etc. 4.0 1.7 1.5 0.4 2.3 Because of above measure forest fires have reduced considerably in the region as shown in the graph 2010 2011 2012 2013 2014 2015 2016 below. Source: Ka Synjuk Ki Hima Arliang Wah Umiam Mawphlang Welfare Society Strengthening Forest Fire Management in India   180 CASE STUDY 2: FIRE MANGEMENT IN JIRANG REGION, MEGHALAYA FIGURE A3.12: JIRANG FOREST IN RAINY SEASON Jirang is located in the Ri Bhoi District of Meghalaya FIGURE A3.13: FOREST FIRE IS RAMPANT and is 36 Km from the district headquarter, Nongpoh. IN JIRANG DUE TO THE Its thick forest cover is home to a rich biodiversity of PRACTICE OF JHUM wildlife and myriad forms of flora and fauna. This CULTIVATION IN FOREST area is prominent for its expensive hardwood timber like sal (Shorea robusta), teak and bamboo. Due to the AREAS remoteness of the area, the local community are still heavily dependent on forests for daily sustenance. Agriculture is the main stay of the people in Jirang and most of them still practice the conventional method of slash and burn cultivation or jhum cultivation to grow food crops such as rice and ginger, as well as cash crops such as broom grass and horticultural crops. Forest fires are rampant in the region with jhum cultivation being one of the main contributors to these fires since it is practiced inside the forest. Cattle grazers in the area are another source of forest fires, often responsible for setting ablaze grazing areas to clear land for new shoots to grow. These are generally uncontrolled burning which often spread to non- grazing areas, damaging forests. Lastly, irresponsible disposal of cigarettes butts and lit matchsticks is the third cause of forest fires. 181 Strengthening Forest Fire Management in India FIGURE A3.14: DEPENDENCE OF PEOPLE FIGURE A3.15: VILLAGE FIRE CONTROL ON FOREST FOR FUEL COMMITTEE MEMBERS WOOD IN JIRANG IN THE JIRANG AREA REGION fund for spreading awareness on the importance of controlling forest fires and management techniques. The real incentive came from the ability of organized VFCCs to protect their habitation from forest fires as most tribals live inside forests. Similar to Mawphlang, forest fires in Jirang are managed through the use of fire lines which are inexpensive yet effective. The Committees still lack Fire management: Traditionally Dorbar (local village equipment and training and rely on makeshift tools government) issued notices for forest fire protection to create fire lines. The fire lines are created through measures but offenders were seldom penalized. Due inter-village collaboration after formation of VFCCs to the remoteness of the forest area, it was very difficult since most forest is co-owned by multiple villages. to monitor and control forest fires. Since the forests are Fire lines are made twice in a year, once at the end jointly owned by 15 villages, there was no incentive for of the monsoon season, and then at the onset of the any particular village to take extra measures. In 2015, windy spring season. Community volunteers patrol with the intervention of the state Forest Department, 15 vulnerable areas, scouting for fires before they spread. Village Fire Control Committees (VFCC) were formed Since the formation of the VFCC, the communities under key villages in the block. Each VFCC was given have reported drastic reduction in number of forest Rs. 10000 by the Forest Department as an operational fires in the region. FIGURE A3.16: A FIRE LINE COVERED IN THICK FOLIAGE WHICH GREW OVER THE MONSOON Strengthening Forest Fire Management in India   182 CASE STUDY 3: KHLOO BLAI SEIN RAIJ TUBER COMMUNITY RESERVED FOREST, MEGHALAYA Thick fire lines have been created by the community FIGURE A3.17: A VIEW OF THE KHOO members, along the entire perimeter of the forest BLAI SEIN RAIJ TUBER which isolates the forest from others in the vicinity. COMMUNITY RESERVED These forest lines are maintained on a regular basis by FOREST community volunteers. The primary reason for the low fire incidence is , however, the status of the forest itself. Being a sacred forest, community members take great care to avoid causing any damage to the forest out of fear of divine and social repercussions. Hence, issues of miscreants and accidental ignition of fire in the forest are almost non-existent as compared to non-sacred forests. FIGURE A3.18: SIGNBOARD WITH RULES AND REGULATIONS INSTALLED AT The Khloo Blai Sein Raij Tuber community reserved ENTRANCE OF THE forest is owned and managed by the community COMMUNITY RESERVE members through a consortium of traditional heads known as ‘The Sein Raj Tuber’ which comprises 27 village from the region. The sacred forest is spread over an area of 16.5 hectares and is located in Tuber Kmaishnong village of Khliehriat Block in East Jaintia Hills district. Members of the Sein Raij Tuber perform various religious rites and rituals in the forest including the famous Chad Sukra which is a dance festival of the community that is performed in the forest every year before the sowing season. FIGURE A3.19: PROMINENT AND WELL- Fire management: There has been zero recorded MAINTAINED FIRE LINES instances of fire in the Sein Raij Sacred Forest due ALONG THE ENTIRE to various measures being implemented in the forest PERIPHERY OF THE area. FOREST Being a sacred location, the Seij Raij has laid down a set of rules and regulations that are strictly enforced by the village dorbar. Large visible sign boards are installed at the entrance of the forest on which all the rules and regulations are clearly stated for people who wish to enter the forest. Lighting of fire in and around the sacred forest is strictly prohibited. People who enter the forest are not allowed to leave in the forest anything that does not belong in it nor are they allowed to take anything from the forest. 183 Strengthening Forest Fire Management in India CASE STUDY 4: FOREST FIRE MANAGEMENT IN GARO HILLS, MEGHALAYA The Garo Hills constitute the western parts of the state forest fires is lesser when compared to the Khasi Hills of Meghalaya. The region is inhabited by the Garo due to the presence of less flammable broad leaf tree tribe and is characterized by mountainous features in species. However, forest fires damage undergrowths, the northern parts and plain areas towards the south wildlife and seeds that can affect the health of the and southwestern parts. In the Garo Hills, land is held forest in the long run. by the Nokma (village Headman) who then allocates it to residents of the village for their settlement and use. Fire management: Again, similar to the Khasi Hills The Garo Hills is one of the richest places in terms and Ri-Bhoi area, a large population of the people of biodiversity with much of the region untouched in the region practice slash and burn cultivation but by man and is host to one of the global biodiversity the main cause of forest fire is human negligence hotspots. The region alone has major biodiversity and intentional burning of fire by miscreants. The reserves viz., the Nokrek National Park, Selbagre frequency and scale however are lesser that in other Hoolock Gibbon Reserve, Balpakram National Park parts of the state. and the Baghmara Reserve Forest. Similar to the Khasi Hills, the Garo Hills receives Unlike in the Khasi Hills where management of heavy rainfall during the monsoon seasons. During community reserves is a joint effort of multiple that time, the forest is less vulnerable to forest fires. villages, the community reserved forests in the Garo Even during the dry season, the intensity and scale of Hills are managed by individual villages, in whose area FIGURE A3.20: TAPIOCA IS COMMONLY USED AS AN EFFECTIVE FIRE BARRIER AROUND JHUM AREAS Strengthening Forest Fire Management in India   184 FIGURE A3.21: JHUM CULTIVATION IS A COMMON PRACTICE IN THE GARO HILLS the forest falls, under the leadership of the Nokma. A Another interesting practice in the Garo Hills is majority of the reserves have been created through the that during the jhum period, when land has to intervention of the Forest Department which provides be cleared, youth volunteers would stand watch monetary incentives to the community for creating whenever controlled burning is being carried out. and maintaining such reserves. The purpose of their The responsibility to keep the fire under control formation is to conserve important resources such as within the jhum area lies with the farmer. In case of water sources and important plant species. a fire breakout, the youth volunteer alerts and rallies community members from the village to douse the Just as practiced in Khasi, Ri-bhoi and Jaintia Hills, flame. The members use twigs and branches to beat the common practice to prevent and control fires is by and douse the flames. There is no penalty to a farmer the use of fire lines. Additionally, community members for such fire accidents. volunteer to act as sentries, alerting people whenever Lately, however, a number of Jhum lands are being there is a fire breakout anywhere around their village. gradually converted into cash crop plantations which A unique practice that can be found in the Garo Hills do not require intermittent land clearing. Farmers are is the use of high resistant plants as fire barriers. A slowly shifting from growing vegetables to growing commonly used plant is Tapioca. areca nut and rubber trees due to higher revenue. 185 Strengthening Forest Fire Management in India CASE STUDY 5: RONGRAM, MEGHALAYA Rongram block is located in West Garo Hills district monitored and controlled by the community. Farmers and comprises 173 villages125. The district is far who wish to clear their land through jhum have to behind the Khasi, Ri-Bhoi and Jaintia Hills in terms intimate the Nokma who in consultation with the of development. Most people depend primarily on village council and the people sets the date for the agriculture for livelihoods and on forest products for activity. The day is chosen such that several youths daily sustenance. Cultivation, as is in most parts of and adult members of the community are available in Meghalaya, involves controlled slashing and burning the village to support in case of a fire breakout. Prior of forests to clear land for cultivation. to burning, pre-control burning is done along the perimeter of the jhum area to create a fire line. On Geographically, the region is hilly in terrain with jhum day, a number of youth volunteers stand watch dense forests which are untouched and a plethora of to alert community members in case of unintentional flora and fauna. spread of fire which has gone beyond the control of the farmer and the volunteers. Fire is doused using Fire management: One of the primary reasons for the branches of trees. In other non- jhum areas, the lower rate of fire incidences despite the prevalence of community relies on fire lines as prescribed by the jhum cultivation, is the way in which jhum is being Forest Department of the Government. Patrolling of practiced. Unlike in other districts, jhum cultivation inRongram and other parts of Garo Hills is tightly forest areas is done by departmental staff. FIGURE A3.22: FORESTS ADJOINING AGRICULTURAL FIELDS DECLARED AS COMMUNITY FORESTS - MANAGED AND PROTECTED BY THE COMMUNITY Census 2011 125 Strengthening Forest Fire Management in India   186 FIGURE A3.23: A FIRE LINE CREATED which in turn acts as a source of water for the Water AROUND THE JHUM Conservation Pond setup by the Soil and Water AREAS Conservation Department for recharging ground water and supplying irrigation water to agricultural fields. Each village frames its rules and regulations for the reserve often prohibiting felling trees and gathering of forest produce from such reserves. However, there is a lack of coordination between villages that are carrying out conservation works. Since the forest owned by each village is relatively small, impact would be better felt if efforts are coordinated. FIGURE A3.25: A FIRE LINE AROUND THE COMMUNITY Villages in Rongram also have VFCCs comprising RESERVE AT SANCHAGRE Nokmas, senior community members and youth, VILLAGE which manage forest fires and raise awareness on fire safety and protection. Each VFCC is given a fund of Rs. 10,000 by the Forest Department for their operational needs. Rongram block has 10 VFCCs while Dalu Block and Chokpot Block have 20 and 10 VFCCs respectively. Another protective measure prescribed by the Forest Department is the creation of community reserves under each village. These reserves are generally created to conserve important resources of the village. The villages are required to contribute land for the creation of such reserves and are given a fund of Rs. 30,000 per reserve to manage the reserve. These reserves are often helped through departmental schemes for holistic development. For example, in the villages of Rangwal, Sanchagre, and Misimagre, the community reserves protect the catchment area FIGURE A3.26: A WATER CONSERVATION POND RECHARGED FROM COMMUNITY FOREST FIGURE A3.24: A PANORAMIC VIEW OF RESERVE IN RANGWAL A COMMUNITY RESERVE VILLAGE AT RANGWAL VILLAGE SUPPORTING PADDY FIELD 187 Strengthening Forest Fire Management in India ANNEX 4 CLASSIFICATION OF THE CAUSES OF FOREST FIRES India does not currently have an official classification Indian context. The categories are hierarchical such scheme for causes of forest fire. The need for a that level-3 categories may be generalized to level 2 or “uniform classification of forest fires” to be “evolved level 1. and adapted by all the States” as part of the “collection and compilation of forest statistics” was recognized Responses by the forest department officers surveyed as early as the 1976 by the National Commission in the 11 states (see Annex 2) as to the main causes of on Agriculture (NCA 1976: 343). An example of a fire in their states were categorized using the proposed possible classification scheme has been devised below. scheme in table A4.1. Table A4.2 presents additional The scheme was adapted from the Fire Database of the information on non-timber forest products (NTFPs) European Forest Fire Information System (EFFIS),126 commonly collected by local forest users in India with with modifications to make it more relevant for the the aid of fire. TABLE A4.1: CATEGORIZATION OF CAUSES OF FOREST FIRES Level 1 Level 1 Definition Level 2 Level 2 Definition Level 3 Level 3 Definition Code Code Code 100 110 Unknown cause Unknown 120 Unspecified or response not clear 200 Natural: forest fire without 210 Lightning direct human involvement or 240 Other natural influence 300 310 Electric power equipment (e.g., sparks from power lines) Accident: forest fire indirectly 320 Railways caused by human actions or 330 Vehicles presence of infrastructure in 340 Works (e.g., road repair) forested area, not by negli- gent use of fire or glowing 350 Firearms, explosives objects 360 Self-ignition 370 Other accident Andrea Camia, Tracy Durrant, and Jesús San-Miguel-Ayanz, “Harmonized Classification Scheme of Fire Causes in the EU Adopted 126 for the European Fire Database of EFFIS,” European Commission Joint Research Centre, EUR 25923 EN (2013), http://effis.jrc. ec.europa.eu/media/cms_page_media/42/LB-NA-25-923-EN-N.pdf. Strengthening Forest Fire Management in India   188 Level 1 Level 1 Definition Level 2 Level 2 Definition Level 3 Level 3 Definition Code Code Code Vegetation management (including con- trolled burning, clearing pine needles, 411 removal of weeds, etc., but not including forest resource collection as reclassified un- der 700 or land use practices under 900) Negligent use of Agricultural burnings (including pasture, 410 fire 412 but not including land use practices as Negligence: forest fire unin- under 900) tentionally caused by humans 400 413 Waste management (non-agricultural) using fire or glowing objects in and around forested areas 414 Campfires, cookfires, or recreational fires 415 Other negligent use 421 Fireworks Negligent use of 422 Cigarettes 420 glowing objects 423 Hot ashes 424 Other glowing objects (torches, etc.) 511 Interest or profit (e.g., encroachment or illicit felling) 512 Conflict or revenge 513 Vandalism Voluntary: forest fire caused 510 Responsible (arson) 514 Excitement (incendiary) 500 by intentional or malicious 515 Crime concealment use of fire 516 Extremist 517 Motivation unknown 521 Mental illness 520 Irresponsible 522 Children 600 Reignition 711 Mahua flowers 712 Tendu leaves 713 Charcoal or ash Forest resource collection: 714 Mushrooms forest fire caused intentionally 710 NTFP collection by 715 Honey 700 or unintentionally to obtain people non-timber forest products 716 Tree resin (NTFPs) and services 717 Gum 718 Seeds 719 Other NTFPs 720 Use of fire for stimulating growth of grass and other fodder for livestock 810 Burning to deter wildlife (including to prevent disease carried by wildlife) 800 Wildlife management 820 Enhancement of wildlife habitat 830 Hunting 910 Shifting cultivation (e.g., jhum and podu) 900 Traditional land use practices 920 Other traditional cultural practices not listed elsewhere Notes: Adapted from classification scheme for Fire Database of the European Forest Fire Information System (EFFIS) (Camia et al. 2014); new categories 700, 800, and 900 have been added and would have be classified in the EFFIS scheme as belonging to category 400. 189 Strengthening Forest Fire Management in India TABLE A4.2: NON-TIMBER FOREST PRODUCTS (NTFPS) COLLECTED USING FIRE, ACCORDING TO SURVEYED FOREST OFFICERS NTFP Frequency How, why, and when fire is used Mahua 20 Feb-Apr (Chhattisgarh) To “clear the surface under the tree” and “make the flowers noticeably visible,” Feb-Apr (Jharkhand) “Instead of sweeping dry leaves they resort to burn dry leaves to clear the ground,” peak sum- mer (Odisha) “The leaf litter is burnt to clear the ground for collection of Mahua flowers,” Apr-May (Telan- gana) Tendu 16 For “better crop of tendu leaves,” Feb-Apr (Chhattisgarh) Burning “by local people through tendu patta contractor” (Chhattisgarh) To “get new flush of leaves of good quality instead of bush cutting” (Odisha) “Burning the whole forest area ground cover for production of new leaves,” peak summer (Odisha) “Tendu coppice growth / shrubs are burnt instead of pruning during the months of February to March for growth of flush new Tendu leaves” (Telangana) Honey 16 “To scare away bees,” Feb-May (Kerala) “Careless handling of fire torches during honey collection results in fire,” Feb-May (Kerala) “During night hours, to disturb honeybees” (Himachal Pradesh) Medicinal plants 7 Including ginger, garcinia, flemingia, and mahul Fodder 5 “ Local people just spread the fire in the forest. With the burning of debris collected in the forest the grass sprouts effectively,” Apr-May (Himachal Pradesh) Seeds 5 Sal seeds collected “by burning leaves and bushes below the trees,” (Jharkhand) Dammar and 5 “Smoke base of the tree to increase the maximum oozing of dammar”, Dec-May (Kerala) resin Materials 4 Materials include thatch grass and broomsticks (Meghalaya) To get quality broom sticks (Odisha) Bamboo shoots 2 Feb-Mar (Tripura) Tubers and 2 “People living in the fringes areas burn the forest for collection of wild tubers,” Nov-Apr (Tri- rhizomes pura) Fruits 2 Clearing undergrowth to make collection easier, Dec-May (Kerala) Nuts 2 “To get rid of thorny weeds underneath cashew plants,” Feb-May (Odisha) For various nuts (Madhya Pradesh) Vegetables 1 Feb-Mar (Tripura) Charcoal 1 "After burning the forest, they used to collect charcoal" (Tripura) Mushrooms 1 In sal forests (Chhattisgarh) Note: Frequency refers to the number of respondents mentioning the collection of that NTFP as a cause of forest fire Sal flower 1 "They burn the forest floor for collection of leaf", peak summer (Odisha) Strengthening Forest Fire Management in India   190 ANNEX 5 EQUIPMENT FOR FOREST FIREFIGHTING IN INDIA127 1. HAND TOOLS FIGURE - A5.1: FIRE RAKE WITH 1.1 Rakes BAMBOO HANDLE IN ODISHA A commonly used tool is a rake. Specialist fire rakes have been developed with longer tines to allow a reasonable “payload” of litter to be maneuvered by the rake as control lines are cleared of loose fire fuels. Longer teeth are required for fire rakes and often a rake may have a multi- purpose head with a cutting edge opposite the rake teeth. A good example of such a tool is a McLeod Tool, also termed “Rake hoe”. This is suitable for grass and forest fuel types. It has a wooden or synthetic handle about 4 feet in length. The use of a locally produced fire rake by Vana Samrakshan Samiti (VSS) members in Odisha was observed by the World Bank team. It is a simple iron rake with long tines and a bamboo handle. Importantly, it is a light tool with a wood or bamboo handle. It is a 1.2 Fire Beater or Swatter simple but effective tool. This is useful for “beating” and “swatting” fires FIGURE - A5.2: FIRE BEATER in fine fuels such as grass to smother the flames. Typically, fire beaters in many countries have been manufactured from flexible material such as a section of broad conveyor belt, perhaps slit into 3 or 4 flaps. This material more readily conforms to whatever shape it is beaten on and effectively “smothers” flaming combustion without generating significant displacement of burning firebrands. 1.3 Pulaski Tool This is a combination cutting and digging tool favored in North America. The tool can also be used as a lever to assist in moving very heavy debris. This is useful in the case of forest fires but is of little value in the case of grass fires. Background note by Ross Smith, World Bank consultant 127. 191 Strengthening Forest Fire Management in India FIGURE - A5.3: PULASKI TOOL FIGURE - A5.5: FIREBUG TORCH 2. PORTABLE POWERED TOOLS 2.1 Leaf Blower 1.4 Knapsack Spray or Backpack Sprayer Landscape-grade leaf blowers have already been Typically, this comprises a 16-20-liter container with successfully used in India (and other countries). a double action pump to ensure a constant stream of Relying upon a sustained and powerful air blast, water. This is not to be confused with an agricultural they are useful in lighter fuels in broadleaf forests. or horticultural spray unit that delivers a fine mist at They have the capacity to quickly remove fuel from a ultra-low volumes. It is useful for mop-up operations proposed fire line - either a control line intended for or direct attack against very small fires. It is usually use against an actual fire or “fire lines” that are planned worn as a back pack, so it is of more limited use in in advance and regularly maintained. Blowers have steep rugged terrain. also been successfully deployed in direct attack against low intensity fires, whereby the operator can create a FIGURE - A5.4: KNAPSACK SPRAYER mineral earth break by forcing leaves and other litter directly into the fire while a fire line is being cleared. 2.2 Chain Saw Chain saws are invaluable for removing downed trees from roads and trails, for quick and effective break up of heavy fuels such as hollow logs and for felling trees close to the fire edge. Some parts and accessories should be regarded as mandatory, including a chain catcher, chain brake, anti-vibration handle, ear muffs and safety goggles or helmet attached face shield. It must be emphasized that the use of chain saws demands training and achievement of minimum standards of competence. Firefighters should never be asked to use, or be provided with chain saws, unless they have undergone appropriate training and hold 1.5 Firebug Torch the necessary accreditation for tasks they are requested to complete. This is a tool for lighting back-burns along a fire control line. The canister is filled with the recommended fuel 2.3 Small Motorized Plant which drips onto the burner head. The flow of fuel can be controlled by the operator who aims to place Two small plant items worthy of consideration for several drops of burning fuel onto fine surface fuel, use in fire investigation and suppression activities thereby igniting it. This is a very quick and efficient are off-road bikes (or trail bikes) and Quad bikes tool for back-firing operations. Fire can be lit at about for speedy access for one or two firefighters to the same pace as an operator can walk. undertake reconnaissance and initial response. Trail Strengthening Forest Fire Management in India   192 bikes can be effective for efficient patrol of fire lines always be avoided. Forest firefighting is different from with minimal personnel resources. Likewise, Quad structural firefighting, so it does not follow that the bikes can offer the ability to carry a small quantity of same type of safety equipment is applicable129. hand tools such as chainsaw, backpack pumper and several handheld tools to a fire scene. They are also Safety clothing for forest firefighters should be made useful for transporting food and water for firefighters of low flammability material such as tight weave cotton when access is minimal or other forms of transport drill and that clothing should be loose fitting with are limited. It is important to note that these items underarm and side pocket slits, loose fitting trouser and require adequate training and strict observance of sleeve cuffs to allow easy ingress and egress of airflow. operational limits for safe application, as well as the Typically, forest firefighters should always leave some use of mandatory safety equipment such as approved bare skin exposed to act as a signal for whether or not helmets and clothing. conditions are suitable for continued work. Safety boots are important when working on fires - sturdy boots with profiled tread soles provide more ankle support 3. PROTECTIVE CLOTHING AND when negotiating uneven or rough terrain and help EQUIPMENT to minimize ankle injuries and slips and falls. Boots should be manufactured from leather or fire resistant Protective clothing is essential for firefighters. It is material. Rubberized “gumboots” are definitely very important to appreciate that forest firefighting is unsuitable for forest firefighting. Furthermore, safety quite different from structural or urban firefighting128 helmets with adjustable harness are recommended to and that the protective equipment that is used for the provide protection from falling objects, and leather latter is completely unsuited to forest firefighting. It is work gloves are recommended. Safety goggles are important to note that all personal safety equipment suggested for operating light machinery such as must be constructed of non-flammable materials, and blowers or chainsaws where flying debris can lodge in that construction from synthetic fabric or materials the operator’s eyes, and ear muffs are recommended that can melt or ignite when exposed to heat must for operators of machinery or powered tools. 128. Structural firefighting can involve short periods of very intense activity where firefighters may undertake search and rescue or very specialized suppression activity in extremely hostile environments, while kitted out in, and protected by, very heavy heat resistant clothing and self-contained breathing apparatus. Structural firefighting protective clothing and equipment is designed to protect personnel from extreme levels of heat, smoke and gasses, but it allows no dissipation of environmental heat (heat absorbed from the environment) or metabolic heat (heat created by personal exertion). Firefighters can only operate in these situations for very short periods, after which they must retreat so they can cool down by shedding their heavy protective clothing. 129. Structural firefighters often need to work ‘in close’ to internal building fires and they need serious protection from radiated heat and direct flame contact. They frequently work in very short shifts but then must retreat to cool down, else they will suffer heat exhaustion through not being able to dissipate metabolic and radiated heat. Forest firefighters likewise need protection from radiated and/or embers heat but they have more opportunity to regulate their distance from an active fire front and reduce their overall heat absorption. There is often a temptation to use structural firefighting equipment for forest firefighting but the practice is dangerous and more likely to induce heat stroke in firefighters. 193 Strengthening Forest Fire Management in India ANNEX 6 SUMMARY NOTES OF WORKSHOP PROCEEDINGS International Workshop on Forest Fire Prevention and Management organized by Ministry of Environment, Forest and Climate Change and The World Bank India Habitat Centre, New Delhi, India November 1-3, 2017 A series of blogs prepared by participants in connection with the workshop can be accessed through the following link: https://blogs.worldbank.org/endpovertyinsouthasia/category/tags/fightforestfire DAY 1, AFTERNOON – EARLY provinces and territories have been involved at each stage of development of the national FDRS, as are WARNING, FIRE DANGER RATING, other agencies such as the Department of Meteorology. AND FIRE DETECTION The FDRS is modular, with different pieces for fire weather, behavior, prediction, and possible impacts. The afternoon sessions of the first day highlighted Of these pieces, the assessment of fire weather is the the numerous entry points for data and technology most basic and essential. Numerous countries have in the FFPM process, including early warning, fire adapted Canada’s FDRS by calibrating it to local danger rating, active fire detection, and post-fire weather, fuels, and intended use. burnt area assessment. Participants highlighted that receiving timely notice of a fire, as is possible through Ross Smith (World Bank) discussed lessons learned satellite-based alerts, is critical for managing the fire. from the examples of Indonesia, Croatia, and Leveraging data and technology for FFPM requires Australia in developing their FDRS. The experiences additional support, however. Ground truthing and of these countries have reinforced the importance verification help to improve the accuracy and quality of developing a system rooted in local conditions of fire information. Moreover, care must be taken with local inputs, which is easily understood by fire to communicate this information with communities managers and land users and which can be reliably and and field staff to translate it into effective action. The effectively communicated to those people who use fire central government and the states can further the or are at risk from fire. Developing and popularizing development of data and technology for FFPM by a national FDRS requires a champion, and often the working together. Both levels of government have backing of a legislative requirement. Experience important roles to play. Furthermore, public outreach has shown that the forestry agencies are typically is critical to ensuring data and technologies are used the institutions that take on this role of champion. on the ground. Such awareness raising is just a piece of Mr. Smith applauded FSI for the work it has done a much broader and more fundamental engagement in developing a nascent FDRS for India. Mr. Smith that is needed with communities to prevent and suggested the system can be refined by continuing manage fires. Participants also discussed the need for to validate it against actual fires, weather, and fuel India to develop a National Forest Fire database. scenarios. Mr. Smith also encouraged FSI to continue to experiment, for example, with including different In the first session of the afternoon, Brian Simpson indices of drought and anthropogenic factors. (Canadian Forest Service) provided an overview of the development and functions of the Canadian Fire E. Vikram (Forest Survey of India, “FSI”) outlined Danger Rating System (FDRS), which formally started the recent advancements in FSI’s “pre-warning alert” in 1968. The FDRS is built upon scientific research, system for identifying areas of high fire danger. As Mr. which began in the 1920s and continues today. The Vikram explained, the purpose of the system is not Strengthening Forest Fire Management in India   194 to predict when and where fires will occur, but rather that use fires to understand their needs, how they to promote the efficient allocation and coordination meet those needs by using fire, what alternatives to of resources for fire prevention and control, and to fire exist for meeting those needs in areas where fire is identify areas of priority for risk mitigation. Discussion degrading forests, and how incentives might support centered on challenges with making effective use of shifting behaviors. Participants raised concern as to FSI’s alerts as a management tool and with translating whether global concerns – limiting fires to reduce pre-warning alerts into actions in the field. Participants carbon emissions – would have an impact on current agreed that states and local forest divisions should fire practices of communities. also consider local knowledge and conditions in identifying areas of high risk and vulnerability, and Numerous scientific studies on the effects of fires that community engagement is indispensable in have been performed in different locations in India, communicating fire danger and reducing risks. but the larger-scale impacts of fire and the extent to which fire is contributing to forest degradation across The second session of the day shifted to the detection India is still poorly understood. The role of invasive of fires and measurement of fire impacts. Charles species and other threats such as climate change on Ichoku (NASA, United States) surveyed the current forest fire regimes are also mostly unknown. These state of remote sensing technologies and scientific knowledge gaps can be filled by a focused research research for detecting forest fires and measuring their agenda on the impacts of fire across a range of forests, impacts. E. Vikram updated participants on recent climates, and topographies. India needs to develop a developments with FSI’s nationwide alert system robust methodology to evaluate the ecological effects for active fire detection. Participants discussed the and economic impacts of forest fires and to assess symbiotic relationship between the central government what fires imply for the country’s commitments for (MoEFCC and FSI) and the states in fire detection, climate change. To this end, there is also a need to the specific needs of states (e.g., Punjab’s concern incentivize forest department personnel to improve for monitoring agricultural fires in areas adjoining field reporting on the occurrence of fire (including forests), and the criticality of greater ground truthing burnt area), and to involve researchers from outside and verification of alerts. Current algorithms and the department. methods of detection cannot be improved without such on-the-ground information provided to FSI and other Tim McGuffog (Forestry Corporation of New South agencies/departments producing alerts by field staff. Wales, Australia) opened the morning’s presentations by describing fire prevention practices in Australia. Preventative burning is done regularly, and Australia DAY 2, MORNING – FIRE has issued national guidelines for prescribed burning PREVENTION which set forth required actions for the planning and implementation of burn operations. Fire managers The morning sessions of Day 2 focused on the have formed partnerships and closely involved prevention of forest fires. The consensus among aboriginal communities in FFPM planning and workshop participants—and within the international operations. As in India, these communities have long community of fire managers and scientists more used fire as a land management tool. Mr. McGuffog broadly—is that the total elimination of fires from also stressed that the health and safety of firefighters forests is unwise and unachievable. Rather, the goal and the public is a priority for fire prevention. of fire prevention should be to minimize the negative impacts of fire and to maximize the benefits, recognizing Pieter van Lierop (UN Food and Agricultural the responsible use of fire as a land management tool Organization, Rome) provided a global view of forest and fire’s place in traditional culture, practices, and fire trends and an outline of FAO’s Fire Management livelihoods in India. Out-of-control and intense fires Voluntary Guidelines. Fire prevention is framed are damaging and should be stopped, but periodic within the Guidelines as a part of integrated fire low-intensity fires may not be bad for forest quality management, along with early warning, preparedness, or health, and occasional fires can help maintain the response, restoration, and monitoring, which aims structure and species composition of some forests. to minimize impacts and maximize benefits to forest To prevent damaging and unwanted fires, forest ecology and society from fire. A community-based managers should engage with the rural communities approach is advocated by the Guidelines. 195 Strengthening Forest Fire Management in India H.S. Suresh (Indian Institute of Science, Bangalore) DAY 2, AFTERNOON – FIRE discussed forest fire ecology from both the local and global perspectives. Globally, fires are a major SUPPRESSION influence in shaping vegetation structures and biomass In the afternoon, the workshop turned from fire and continue to be used by indigenous peoples to prevention to suppression. Discussion covered a manage their natural resources. In India, there is a wide range of issues. Participants discussed how the long tradition of people applying controlled burning current interpretations of Supreme Court rulings as a land management tool. The suppression of fire have created some uncertainty about the extent to under colonial silvicultural systems has resulted in which fire lines may be cleared and widened, where major changes in vegetation. Research in Tamil Nadu needed, and whether fuels such as fallen trees may has found that low-intensity fires do not impact overall be removed from within protected areas to prevent species composition, biomass, or regeneration and damaging fires. Participants also discussed equipment that areas which have experienced longer periods of and methods for fire suppression. The shared view fire exclusion tend to have higher fuel loads and are among participants is that forest departments do not more prone to canopy fires, which cause large-scale have adequate equipment but that equipment needs tree mortality. depend on geography, fuel types, and what is locally acceptable. Local solutions are often best but are Dmitry Krasovsky (Ministry of Forestry, Belarus) easily overlooked or lost. There is a common need presented the system for forest fire prevention and for training among field staff, with different levels of control in Belarus. As in India, remote sensing training tailored for different levels of responsibility technologies have played a pivotal role in fire detection for those in charge of crews’ safety on the fire line. and prevention in Belarus. Mr. Krasovsky emphasized Participants also agreed that there is a need for a more the importance of knowledge exchange for improving formal mechanism of knowledge exchange between fire prevention and control in Belarus and learning the state forest departments to share experiences and from other countries, including India. innovations on training, equipment, technologies, and policies. Finally, participants discussed the role Amitabh Agnihotri (Forest Department, Government of local communities in responding to forest fires. of Madhya Pradesh, India) described the initiatives Out-migration, labor shortages, and the erosion of that Madhya Pradesh has taken to strengthen FFPM. traditional communal institutions present challenges to As Mr. Agnihotri noted, involving communities community involvement in some areas, and the forest has been key to Madhya Pradesh’s success, as the department must be mindful of these constraints. forest department has worked with the Joint Forest Management Committees to ask local people for Alfredo Nolasco Morales (National Forestry solutions and to promote alternatives to burning for Commission, Mexico) described how FFPM has collecting non-timber forest products such as tendu undergone a paradigmatic shift in Mexico, from a leaves and mahua flowers. Another ingredient in policy of total suppression to a more integrated policy Madhya Pradesh’s success has been the effective use of of fire management, recognizing that some fires are technology for satellite detection of fires and near-real- beneficial - ecologically, socially, and economically. time alerts. At the same time, fire is seen as damaging Community-based fire management has accompanied to forest regeneration, which is already under stress this shift, involving local actors, agencies, and from livestock, and therefore controlled burning is NGOs; community volunteers to fight fires, conduct not used as consistently as a preventive strategy. prescribed burns, and implement fuel management plans; and incorporation of traditional and indigenous P.S. Nongbri (Forest Department, Government knowledge into fire management planning at the of Meghalaya, India) discussed the role of the local level. Achieving this shift took time and strong community in FFPM in Meghalaya, where 95 percent leadership and is reflected in the country’s new of the forest estate is under community and private 25-year strategic plan for FFPM. Mexico has also ownership outside the direct management of the forest implemented an Incident Command System (ICS) for department. In Meghalaya, the department is striving fire suppression, though this took significant time and to have close cooperation with the communities to training to institute. Mexico has been able to improve ensure more forest conservation and protection of the its approach to FFPM without increasing the budget. forest area. Strengthening Forest Fire Management in India   196 Instead, leaders focused on allocating resources more involvement of rural communities in the VSS, and the effectively and efficiently. creation of mobile forest firefighting squads. Mohan Raj Kafle (Department of Forests, Nepal) Ross Smith shared insights on firefighting training, presented FFPM in Nepal, where fires have emerged safety, and equipment. Seasonal and permanent as one of the major challenges threatening biodiversity firefighters require training to understand what they and forest ecosystems. Under the country’s Fire are required to do and how to best do it, and most Management Strategy of 2010, coordinated action importantly, when to retreat. Training should address is required at the national, district, and community basic fire behaviour principles so that firefighters level. The country has made significant efforts understand how and why fires behave to adopt the to strengthen and mobilize local communities as most effective suppression techniques at their disposal. implementers of FFPM, creating community-based There are also critical safety connotations that must forest fire management groups and conducting be introduced to firefighters, and crews should be FFPM operational planning with these groups. The trained in proper equipment use (e.g., when using Department of Forests also holds regular trainings chainsaws or other potentially dangerous tools). Mr. with volunteer firefighting groups as well as with Smith suggested that for India the focus in equipment public safety personnel, the army, paid forest staff, development should be on hand tools and small and student groups on fire safety and response. motorized equipment (leaf blowers, chainsaws, quad bikes, motor bikes, etc.) and protective clothing. He C. Jayaram (Forest Department, Government of emphasized there is no “silver bullet” – what will work Karnataka, India) presented an overview of FFPM in best under forest fire conditions in India is what the Karnataka and challenges for fire management in that local people develop or elect to use. state. Challenges include managing fires in protected areas, where Mr. Jayaram noted the blanket ban on fuel removal has hindered FFPM and conservation DAY 3, MORNING – WORKING objectives. Mr. Jayaram argued that the removal of WITH OTHER AGENCIES AND dead and fallen hardwood trees, which create the COMMUNITIES potential for intense and long-lasting fires, should be allowed on a limited basis as part of the management The third day opened with a panel discussion of protected areas. Mr. Jayaram also advocated for on institutional coordination and community regular training of forest personnel in firefighting engagement with Kamal Kishore (National Disaster equipment and methods, noting the recent death of Management Authority, India), Alfredo Nolasco a forest guard and serious injuries to a range forest Morales, and Ramesh Pandey (Fire Cell, Ministry of officer (RFO) and two forest watchers in Karnataka and Home Affairs). One issue for coordination discussed the impact this incident has had on the department. is how the state forest departments should coordinate He also highlighted the need for tools to assess the with disaster management authorities. District disaster damage and losses stemming from forest fires. management plans do not currently address forest fire risks, and forest fires have not been a priority at the Ombir Singh (Forest Research Institute, “FRI,” India) level of national policy. Institutionally, the respective discussed the research and development of forest roles and responsibilities of the forest department, firefighting tools by FRI. FRI has developed and disaster authorities, and local communities for promoted an equipment kit with lightweight tools managing and responding to fires is not entirely clear. for manual beating. Mr. Singh also discussed the Forest fires will continue to be primarily managed by proposed establishment of a Centre for Forest Fire the state forest departments but will also need to be a Management at FRI. part of disaster planning and the forest department should be involved in this process. Because other T.A.K. Sinha (Forest Department, Government of agencies, such as local police, fire departments, and Odisha, India) shared recent developments with disaster management agencies, may be called in case FFPM in Odisha, including the formulation of a of large fires, it is important that they are trained in Standard Operating Procedure for FFPM, regular fire forest fire suppression methods. And moreover, joint risk zoning to inform FFPM planning, popularizing trainings should be organized to coordinate between backpack leaf blowers for clearing fire lines, successful 197 Strengthening Forest Fire Management in India these departments. More generally, there is a need with earlier sessions, was the need for better data to professionalize the forest fire management service. and information on fires to inform management. At the same time, it must be recognized that other Panelists also pointed to the need to involve scientists agencies and institutions have their own constraints and experts from other fields, including atmospheric (e.g., fire departments are understaffed or under- sciences and IT, to develop new tools and methods for resourced in 95 percent of urban areas), which may FFPM. limit their possible involvement in FFPM. Beyond the disaster management authorities, other stakeholders P. Raghuveer (Forest Department, Government of are also involved in FFPM, including the media and Telangana, India) shared the recent experience of the public. Participants agreed that more needs to be Telangana with FFPM. Among the state’s successes done to educate the media and public about FFPM have been a campaign to curb burning for tendu and the role of fires in forest ecology. leaf collection and the creation of fire risk maps. Telangana reduced tendu burning by working with Discussants also emphasized the importance of field-level staff to identify the villages where fire informed fire management. As Mr. Pandey argued, use was highest, providing funds to those villages the low priority accorded to FFPM in national policy contingent on reducing fires, and encouraging the may be a product of missing data and knowledge on creation of village sub-committees for fire protection. fires. Without good data on fires and their impacts, The state has also created fire risk maps down to the it is difficult to weigh the appropriate level of public beat and compartment level, working with field staff investment and to convince policymakers of the need in the most fire-prone areas to assess why those areas for greater resources. Good data can also improve experience more fires than others and to identify the level of public accountability and credibility. Mr. appropriate solutions for the management of those Nolasco reinforced the need to capture, maintain, and areas, from providing extension services to fire-reliant use data for FFPM, particularly within the context of communities to increasing enforcement. climate change as the nature of fire risk is changing in many areas. Other participants agreed that better C. Sudhakar Reddy (National Remote Sensing Centre, science and data are urgently needed and there should “NRSC,” India) offered a survey of the current scientific be a public funding mechanism in place to support research on the impacts of forest fires in India. Mr. fire research. Moreover, research institutions should Reddy cited evidence that low-intensity fires may be be working in partnership with forest department beneficial, though frequent and repeat burning may to improve forest fire prevention and management affect seedlings and regeneration. Several studies practices. have been performed of carbon emissions for biomass burning in forest fires, though currently there is a lack A third key message of the morning’s discussion of field-based data on burning efficiency or available was on the importance of good leadership. As Mr. biomass, leading to large uncertainty in existing Nolasco reflected on his experience in Mexico, while estimates, particularly at the local level. Mr. Reddy no single agency can take on the burden of FFPM by pointed to a need for sampling of different vegetation itself, there is a need for an agency to take the lead. types on a regular basis and more accurate ground Good leadership means a commitment to delivering data. outcomes and care for constituents—including firefighters, scientists, rural communities, the media/ Kasturi Chakraborty (North Eastern Space public, and society at large. The lead agency should Applications Centre, “NESAC,” Meghalaya, India) form part of a network of leaders, with MoEFCC, discussed research by NESAC on forest fires in the legislators, the state forest departments, communities, Northeast, the most fire-prone region in India. Using NGOs, and other agencies—each of which has a remote sensing technologies and historic data on fires, respective role and responsibilities. NESAC has identified forest areas for management priority. NESAC has also conducted analyses of burnt area in the Northeast and is creating a forest DAY 3, MORNING – FIRE IMPACTS fire dashboard for the states in the region. Ms. Chakraborty pointed to a need for a stronger database, The final session of the workshop focused on the not just of forest cover and fire locations, but also of impacts of fire. A major theme of this session, as roads, assets, settlements, and other infrastructure, to Strengthening Forest Fire Management in India   198 assist with FFPM. Ms. Chakraborty also emphasized • Strong need the need for fire managers and scientists to involve • Collaboration with communities experts from other fields, including the atmospheric • Passion for fire management sciences and IT, rather than working in isolation. • Well structured • Quantify data • Stepping stone for more fruitful discussions PARTICIPANTS’ DESCRIPTIONS OF THE • Long way to go WORKSHOP • More budget needed for forest fires • Integration and sharing of information Participants were asked to describe the workshop in • Lot of work in FSI – long way to go in forest fire one word. The list of phrases they offered is below: • Systematic effort • Highly relevant and educative • Informative / theme on national park and • Sharing experience sanctuary missing • SOP for coordinated response • Good • Good exposure • Enlightening • Since Belarus has an interpreter, we need more • Assessment of loss to ecosystem than one word • Very interesting • Excellent seminar / more information on modern • Eye opener technology needed • Informative more funding required for forest fire • Excellent science without modern technology is • Must lead to better forest fire management required • Stock taking of useful information • Sharing of information and knowledge • Good practice around the world • Enjoyed the logistics • Information sharing and community participation • Fascinating and educative • Good beginning at the end • Extremely important • Forest fire volunteer • Nature protects, let’s protect her Workshop Agenda Day 1 (November 1) Time Session Title and Description Speakers 14:00 – 15:00 Welcome and Opening Remarks Introduction by Pyush Dogra, Senior Environmental Specialist (World Bank) Welcome address by Saibal Dasgupta, Additional Director General of Forests (MoEFCC) Keynote address by Siddhanta Das, Director General of Forests and Special Secretary (MoEFCC) Address by Alexander Antonovich Kulik, First Deputy Minister of Forestry (Ministry of Forestry, Belarus) Word of thanks by Pyush Dogra, Senior Environmental Specialist (World Bank) 199 Strengthening Forest Fire Management in India Time Session Title and Description Speakers 15:00 – 15:15 Coffee and tea break 15:15 – 16:30 Fire Danger Rating, Early Warning and Chair: Andrew Michael Mitchell, Senior Forecasting of Forest Fires, and Detection of Forestry Specialist (World Bank) Forest Fires (Part I) Panelists: The first part will present international experience • Brian Simpson, Fire Analyst and Modeler in developing early-warning systems to assess and (Canadian Forest Service)  forecast weather conditions of high fire danger, • Ross Smith, Consultant (World Bank) identify areas of high fire risk, and model the behavior of potential fires. 16:30 – 18:15 Fire Danger Rating, Early Warning and Chair: Saibal Dasgupta, Additional Director Forecasting of Forest Fires, and Detection of General of Forests (MoEFCC) Forest Fires (Part II) Panelists: The second part will focus on the detection of • Charles Ichoku, Research Physical Scientist active forest fires, including systems to notify (NASA, United States) fire managers and the public when fires occur. • E. Vikram, Deputy Director (Forest Survey of Presenters will discuss the use of remote sensing as India) well as on-the-ground systems for fire monitoring. 18:30 – 19:30 Dinner and reception Day 2 (November 2) Time Session Title and Description Speakers 9:00 – 10:30 Prevention of Forest Fires (Part I) Chair: Abi Tamim Vanak, Fellow (Ashoka Trust for Research in Ecology and the Environment, The morning of the second day will be devoted India) entirely to forest fire prevention. Practitioners from India and other countries will discuss policies, Panelists: management strategies, and practices that have • Tim McGuffog, State Fire Manager (Forestry proven effective in reducing unwanted fires in a Corporation of NSW, Australia) variety of forest types, climates, topographies, and • Pieter Van Lierop, Forestry Officer (Food and social and economic contexts. Agriculture Organization of the United Nations, Italy) • H. S. Suresh, Researcher (Indian Institute of Science, Bangalore) 10:30 – 11:00 Coffee and tea break 11:00 – 12:30 Prevention of Forest Fires (Part II) Chair: Rupak De, Principal Chief Conservator of Forests (Government of The second half of the morning session will focus Uttar Pradesh) specifically on the role of local communities in preventing forest fires and how to strengthen the Panelists: effectiveness of community engagement. • Dmitry Krasovsky, Deputy Head of Department, Department of Forestry (Ministry of Forestry, Belarus) • Amitabh Agnihotri, Additional Principal Chief Conservator of Forests (Government of Madhya Pradesh) • P. S. Nongbri, Conservator of Forests, Government of Meghalaya Strengthening Forest Fire Management in India   200 Time Session Title and Description Speakers 12:30 – 14:00 Lunch break 14:00 – 15:30 Forest Fire Response and Suppression (Part I) Chair: Thomas Chandy, Principal Chief Conservator of Forests (Government of Sikkim) The afternoon of Day 2 will turn to the management of unwanted fire. International participants Panelists: will present on the systems their countries have • Alfredo Nolasco Morales, Manager of developed for fire response and the principal Fire Management (National Forestry Commission, methods of suppression that have been deployed Mexico) successfully in different forest types and terrain. • Mohan Raj Kafle, Under Secretary and Focal Point of Forest Fire Management (Department of Forests, Nepal) • C. Jayaram, Chief Wildlife Warden (Government of Karnataka) 15:30 – 16:00 Coffee and tea break 16:00 – 17:30 Forest Fire Response and Suppression (Part II) Chair: S. S. Negi, Former Director General of Forests (MoEFCC) The second half of the afternoon will include a discussion on firefighter safety and equipment. Panelists: • Ombir Singh, Scientist “E” (Forest Research Institute, India) • T. A. K. Sinha, Additional Principal Chief Conservator of Forests (Government of Odisha) • Ross Smith, Consultant (World Bank) 18:00 – 19:30 Dinner and reception Day 3 (November 3) Time Session Title and Description Speakers 9:00 – 10:30 Institutional Coordination Chair: Kamal Kishore, Member (National Disaster Management Authority, India) The third day will open with a panel discussion on strategies for enhancing coordination between Panelists: agencies responsible for forest management and • Alfredo Nolasco Morales, Manager of disaster response at the national, state, and local Fire Management (National Forestry Commission, level. Mexico) • Ramesh Pandey, Chief Conservator of Forests (Government of Uttar Pradesh) 10:30 – 11:00 Coffee and tea break 11:00 – 12:30 Forest Fire Impacts Chair: A. M. Singh, Principal Chief Conservator of Forests (Government of Assam) The morning session will then turn to assessing the environmental and economic impacts of forest Panelists: fires. Impacts to be discussed include the effects • P. Raghuveer, Additional Principal Chief of fire on biodiversity, ecological services, climate Conservator of Forests (Government of Telangana) change, and air quality. • C. Sudhakar Reddy, Scientist (National Remote Sensing Centre) • Kasturi Chakraborty, Scientist (North Eastern Space Applications Centre) 201 Strengthening Forest Fire Management in India Time Session Title and Description Speakers 12:30 – 12:45 Wrap-up and Thanks • Closing Remarks by Saibal Dasgupta, Additional Director General of Forests (MoEFCC) • Word of thanks by Urvashi Narain, Lead Economist (World Bank) 12:45 – 14:00 Lunch Strengthening Forest Fire Management in India   202 ANNEX 7 DATA SHEETS SENT TO STATE FOREST DEPARTMENTS Instructions Please complete this data sheet as fully as possible. You can either fill in responses directly in this document or create a new document. Where applicable, you can also attach other reports or documents that provide the required information for a question. Please note the name of the attachment with the information for that particular question. If the information asked by a question is not available or not collected, please indicate “No Info”. Please send the completed data sheet and any attachments to [...]. Basic Information 1. State or Union Territory: 2. Officer completing the data sheet: Name Designation Contact information General Forest Data 3. Please indicate the area of forest in hectares: Forest area by administrative category Total Area Area under valid (hectares) Working Plan, Working Scheme, or another plan (hectares) Total forest area in the state     Forest area under forest department (including categories below)     Reserved forest     Plantation forest     Assisted natural regeneration forest     Forest in protected areas (wildlife sanctuaries, national parks, etc.)     Non-forested lands under forest department     Other forest area under forest department (please specify)     Forest area under revenue department     Forest under van panchayats     Area of private forest (e.g., municipal or cantonment forest)     Area of communal forest     Area of forest under other departments or agencies     203 Strengthening Forest Fire Management in India 4 Have the boundaries of all forest area under the forest department been mapped and digitized in a GIS? If so, which office or person maintains this information? 5. Have the boundaries of other forest areas not under the forest department (e.g., revenue forest) been mapped and digitized in a GIS? If so, which office or person maintains this information? Fire Lines 6. Please indicate the length of fire lines in kilometers (by width of the fire line): Length of fire lines in forest area by administrative category Length of fire lines in km Forest area under the forest department 3 m wide or 5 m or wider less Total length of fire lines stipulated by Working Plans (WP)     Fire lines mapped and digitized on a GIS layer     Fire lines not functional or not maintained according to WP     Fire lines maintained annually#     Other forest areas not under the forest department Total length of fire lines stipulated by Working Schemes (WS) or     other required plans Actual length of functional, maintained fire lines     Fire lines not functional or not maintained according to WS or     other required plans Fire lines maintained annually #     Note: / # for lines cleaned two or three times per season, please include this only once here 7. Have the fire lines in areas under the forest department have been mapped and digitized in a GIS? If so, which office or person maintains this information? 8. Have the fire lines in other areas not under the forest department have been mapped and digitized in a GIS? If so, which office or person maintains this information? 9. If the actual length of clear, maintained fire lines under the forest department is less than stipulated than by the Working Plans, please comment on reasons for this. 10. If the actual length of clear, maintained fire lines in areas not under the forest department is less than stipulated than by the Working Schemes and other plans, please comment on reasons for this. Controlled Burning 11. Please state the area of forest subject to controlled burning (in hectares): Forest area under the forest department Area (ha) Annual area stipulated for controlled burning by WP Actual area of controlled burning done annually Other forest area not under the forest department Annual area stipulated for controlled burning by WS or other required plan Actual area of controlled burning done annually Strengthening Forest Fire Management in India   204 Fire Causes, Reporting of Fire Incidents, and Burnt Area 12. Please state the total number of fire incidents and area burnt by forest fires over the past 10 years: Number of fires 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 State forest Non-state forest Area burnt 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 (hectares) State forest Non-state forest 13. Are all fire incidents investigated and reported on? 14. Please state what information is required to be collected for each fire incident reported by field staff (or provide a copy of the report form). 15. Please describe how the causes of fire are classified in fire incident reports for your state (e.g. agricultural escape, Tendu leaf, Mahua flower, accident, unknown, etc.). Are causes of fire classified differently for areas under forest department control versus in other forest? 16. To the best of your knowledge, please indicate the number of fire incidents in your state by cause per year for the last 10 years, including where the cause is listed as “unknown.” Cause (please 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 specify) Unknown 17. To the best of your knowledge, please provide the area of forest burned by cause of fire per year for the last 10 years in your state, including where the cause is listed as “unknown” (in hectares). Cause (please 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 specify) Unknown 205 Strengthening Forest Fire Management in India 18. If monetary damages are reported for forest fires, please state the damages by cause of fire per year for the last 10 years, including where the cause is listed as “unknown” (in Rupees). Cause (please 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 specify) Unknown 19. Please describe the methodology used to estimate damages. 20. Please report how burnt area is assessed (i.e., actual total area burnt, area treated, or area of counter fires, area at detection, etc.). Satellite-Based Monitoring of Forest Fires 21. Are field staff required to report back on fire alerts provided by FSI or state monitoring systems? What is the reporting rate (number of field reports received vs number of alerts)? 22. What percent of the fire alerts provided by FSI/state system prove to be false? 23. What percent of total fire incidents reported by field staff are not detected by the FSI/state system? Strengthening Forest Fire Management in India   206 ANNEX 8 QUESTIONNAIRE SENT TO STATE DISASTER MANAGEMENT AGENCIES Welcome! 8. Do you have any role in recovery after a fire event? 9. Are forest fires included in the State and/or District This survey will collect information about the Disaster Management Plan? If so, how? prevention and management of forest fires in your 10. Once a forest fire takes place, does the relevant area. SDMA and/or District Disaster Management Authority (DDMA) team have representation The questions are part of a study by the World Bank (either standing or ad-hoc) from the forest for the Ministry of Environment, Forest and Climate department? Change. 11. How do SDMAs and/or DDMAs engage with communities regarding forest fires (for example, Your input is very important to the findings and discussing forest fires with communities during recommendations of the study. We appreciate your awareness-building exercises)? time and thought in responding. 12. Is there a mechanism in place for the SDMA/ SDRF to co-ordinate with the forest department Any answers you provide will be anonymous. Your regarding forest fires? If so, please provide details. identifying information will not be shared with 13. Please provide any further comments or anyone outside the team of World Bank researchers suggestions that you may have for improving co- completing this study. ordination with the forest department regarding forest fires. For any other suggestions, questions, or comments, 14. Do you have any other comments, questions, or please email […] suggestions that have not been covered? 15. May we contact you if we have any more questions Thank you for your participation! about forest fires in your area? If so, please provide 1. Please enter the name of your State or Union your contact information below. Territory. 2. Please enter the name of your organization and Name: your designation. Salutation (Dr., Mr., Ms., etc.): 3. What actions are taken to handle the occurrence Email: of forest fires in the state? Phone: 4. Based on past fire incidences, what kind of response mechanism has been developed? !Note: Your name and contact information will not be 5. What role does SDMA play in controlling forest used for any other purpose or shared with anyone outside fires compared to other agencies in terms of pre- the team of World Bank researchers completing the fire, real time or post-fire occurrence? assessment without your consent. Should you have any 6. What are the plans ahead to tackle forest fires? questions or concerns, please contact the study team at 7. Can you throw some light on some of the key actions […]. taken to manage forest fires which happened in the recent past? Thank you for completing the survey! 207 Strengthening Forest Fire Management in India REFERENCES • Abt, K., R.J. Huggett, Jr., T.P. Holmes. 2008. Monsoon Dynamics and Climate Change.” Annu. “Designing Economic Impact Assessments for Rev. Earth Planet. Sci. 43: 2.1-2.49. USFS Wildfire Programs” in The Economics of Forest • Anderson, Hal E. 1982. “Aids to Determining Disturbances: Wildfire, Storms, and Invasive Species, edited Fuel Models for Estimating Fire Behavior.” by T.P. Holmes, J. Prestemon, K.L. Abt, 151-166. U.S. Department of Agriculture, Forest Service, Springer Science + Business Media B.V. General Technical Report INT-22, Intermountain • AFAC (Australian Fire and Emergency Services Forest and Range Experiment Station, Ogden, Authorities Council). 2016. “National Position on Utah. Prescribed Burning.” AFAC Publication No. 2036, • Andreae, M.O. and P. Merlet. 2001. “Emission and East Melbourne, Vic., Australia, October 2016. Trace Gases and Aerosols from Biomass Burning.” • Agnihotri, Amitabh. 2017. “Madhya Pradesh Global Biochemical Cycles 15, no. 4: 955-966. Forest Department – an overview”, Forest • Angelsen, Arild, Pamela Jagger, Ronnie Department, Government of Madhya Pradesh, Babigumira, Brian Belcher, Nicholas J. Hogarth, India, presentation and discussion at the workshop Simone Bauch, Jane Börner, Carsten Smith-Hall, on Forest Fire Prevention and Management Sven Wunder. 2014. “Environmental Income organized by MoEFCC and the World Bank in and Rural Livelihoods: A Global Comparative New Delhi, November 2017. Analysis.” World Development 64, Supplement 1: • Alexander, M.E. 2008. “A Primer on the Canadian S12-S28. Forest Fire Danger Rating System (CFFDRS).” • Assam Forest Policy. 2004. Presentation to the Alaska Division of Forestry • Atkinson, Giles and Haripriya Gundimeda. 2006. Spring Operations Meeting, Fairbanks, Alaska, 4 “Accounting for India’s Forest Wealth.” Working April 2008, https://www.frames.gov/catalog/16722. Paper 5/2006, Madras School of Economics, • Alkhatib, Ahmad A.A. 2014. “A Review on Forest Chennai, India, February 2006. Fire Detection Techniques.” International Journal • Badarinath, K.V.S., and Krishna Prasad Vadrevu. of Distributed Sensor Networks 2014, http://dx.doi. 2011. “Carbon Dioxide Emissions from Forest org/10.1155/2014/597368. Biomass Burning in India.” Global Environmental • An, Zhisheng, Wu Guoxiong, Li JIanping, Sun Research 15: 45-52. Youbin, Liu Yimin, Zhou Weijian, Cai Yanjun, • Bahuguna, V.K. 1999. “Forest Fire Prevention Duan Anmin, Li, Mao Jiangyu, Cheng Hai, Shi and Control Strategies in India.” http://www.fire. Zhengguo, Tan Liangcheng, Yan Hong, Ao Hong, uni-freiburg.de/iffn/country/in/in_1.htm. Chang Hong, and Feng Juan. 2015. “Global Strengthening Forest Fire Management in India   208 • Banerjee, S.P. and Sumer Chand. 1981. “Physico- for Reliable Forest Fires Detection.” Procedia Chemical Properties and Moisture Characteristics Computer Science 19: 794-801. of Soils as Influenced by Forest Fire.” The Indian • Camia, Andrea, Tracy Durrant, and Jesús San- Forester 107, no. 3: 178-182. Miguel-Ayanz, “Harmonized Classification • BAPPENAS-ADB (Indonesian Ministry of National Scheme of Fire Causes in the EU Adopted for the Development Planning and Asian Development European Fire Database of EFFIS,” European Bank). 1999. “Causes, Extent, Impact and Costs Commission Joint Research Centre, EUR 25923 of 1997/1998 Fires and Drought.” Final Report, EN (2013), http://effis.jrc.ec.europa.eu/media/ Annex 1 and 2. Planning for Fire Prevention and cms_page_media/42/LB-NA-25-923-EN-N.pdf Drought Management Project. ADB TA 2999- • Chakraborty, Kasturi, Partho Protim Mondal, INO, Jakarta. Mayuri Chabukdhara, and S. Sudhakar. 2014. • Barbero, Renaud. John T. Abatzoglou, Timothy J. “Forest Fire Scenario and Challenges of Mitigation Brown. 2015. “Seasonal Reversal of the Influence of During Fire Season in North East India.” The El Niño-Southern Oscillation on Very Large Wildfire International Archives of the Photogrammetry, Occurrence in the Interior Northwestern United Remote Sensing and Spatial Information Sciences States.” Geophysical Research Letters 42: 3538-3545. XL-8: 27-33. Accessed June 29, 2017. http://www. • Belcher, Brian, Ramadhani Achdiawan, Sonya int-arch-photogramm-remote-sens-spatial-inf-sci. Dewi. 2015. “Forest-Based Livelihoods Strategies net/XL-8/27/2014/isprsarchives-XL-8-27-2014. Conditioned by Market Remoteness and Forest pdf. Proximity in Jharkhand, India.” World Development • Chandra, K.K. and Atul Kumar Bhardwaj. 2015. 66: 269-79. “Incidence of Forest Fire in India and Its Effect • Bhagwat, Shonil A., Elinor Breman, Tarsh on Terrestrial Ecosystem Dynamics, Nutrient Thekaekara, Thomas F. Thorton, and Katherine and Microbial Status of Soil.” International Journal of J. Willis. 2012. “A Battle Lost? Report on Two Agriculture and Forestry 5, no. 2: 69-78. Centuries of Invasion and Management of Lantana • Chandran, Manoj. 2012. “Study on the camara L. in Australia, India, and South Africa.” Impact of Ban on Green Felling of Chir Pine PLoS ONE 7, no. 3: e32407. (Pinus roxburghii, Sarg.). Uttarakhand Forest • Bhandari, B.S. 1995. “Recovery of a Submontane Department, Dehradun, India. Grazing Land Following Summer Burning.” D. • Chhattisgarh Forest and Culture Department. Phil. Thesis, HNB, Garhwal University, Srinagar- 2001. Chhattisgarh State Forest Policy. Raipur. Garhwal. Retrieved from http://www.cgforest.com/wildlife/ • Bhandari, B.S., J.P. Mehta, and S.C. Tiwari. 1997. media/ForestProtection_wl.pdf. “Animal Behavior and Carrying Capacity of Burnt • DAFF (Department of Agriculture, Forestry and Unburnt Submontane Grazing Lands.” Trop. and Fisheries, South Africa). 2013. Publication Ecol. 38, no. 1: 149-152. of the Fire Danger Rating System for General • Bhandari, B.S., J.P. Mehta, and R.L. Semwal. Information in Terms of Section 9(1) on the 2011. “Forest Fire in Uttarakhand Himalaya: An National Veld and Forest Fire Act, 1998 (Act No. Overview” in Glimpses of Forestry Research in the Indian 101 of 1998). Notice 1099 of 2013. Government Himalayan Region, edited by G.C.S. Negi and P.P. Gazette No. 37014, 15 November 2013. Dhyani, 167-175. Almora, India: ENVIS Centre • Darlong, V.T. 2002. “Traditional Community- on Himalayan Ecology. Based Fire Management Among the Mizo Shifting • Bhattacharya, Prodyut, Lolita Pradhan, and Cultivators of Mizoram in Northeast India” Ganesh Yadav. 2010. “Joint Forest Management in Communities in Flames: Proceedings of an International in India: Experiences of Two Decades.” Resources, Conference on Community Involvement in Fire Management, Conservation and Recycling 54: 469-480. edited by Peter Moore, David Ganz, Lay Cheng Tan, Thomas Enters and Patrick B. Durst, http:// • Bouabdellah, Kechar, Houache Noureddine, www.fao.org/docrep/005/AC798E/ac798e0j.htm. Sekhri Larbi. 2013. “Using Wireless Networks 209 Strengthening Forest Fire Management in India • de Groot, William J., Robert Field, Michael Brady, Fires.” FSI, Dehradun, Uttarakhand, India. Orbita Roswintiarti, Maznorizan Mohamad. 2007. • FSI. 2015. India State of Forest Report 2015. Ministry of “Development of the Indonesian and Malaysian Environment, Forest and Climate Change, Fire Danger Rating Systems.” Mitig Adapt Strat Glob • FSI. 2016. “Forest Fire Burnt Area Assessment.” Change 12: 165-80. Presentation to World Bank study team at FSI, • Deccan Chronicle. 2017. “312 forest fires daily Dehradun, Uttarakhand, 14 December 2016. in state.” March 10. http://www.deccanchronicle. • FSI. 2017a. “Forest Fire Pre-Warning by FSI.” com/nation/current-affairs/100317/312-forest- Presentation to World Bank study team, June fires-daily-in-telangana.html. 2017. • Department of Environment and Forest • FSI. 2017b. “Forest Fire Related Interventions Government of Assam, India. 2015. “Assam State by Forest Survey of India.” Background note for Action Plan on Climate Change (2015-2020).” World Bank study team. • Department of Environment, Science and • FSI. 2018. India State of Forest Report 2017. Forest Technology, Government of Himachal Pradesh. Survey of India, Ministry of Environment, Forest n.d. “State Strategy and Action Plan on Climate and Climate Change, Dehradun, Uttarakhand, Change Himachal Pradesh - 2012.” February 2018, http://fsi.nic.in/details. • DWAF (Department of Water Affairs and Forestry, php?pgID=sb_64. South Africa). 2005. The National Fire Danger • Ganguly D, Rasch PJ, Wang H, Yoon JH. 2012. Rating System in Terms of the National Veld and “Climate Response of the Sout Asian Monsoon Forest Fire Act, 1998 (Act No. 101 of 1998). Notice System to Anthropogenic Aerosols.” J. Geophys. Res. 1054 of 2005, Government Gazette No. 27735, 8 117: D13209. July 2005. • Giglio, Luis. 2015. “MODIS Collection 6 Active • EC JRC (European Commission Joint Research Fire Product User’s Guide,” Department of Centre). 2016. Global Human Settlement Layer, Geographical Sciences, University of Maryland, available through JRC Data Catalogue, http://data. United States, 18 March 2015, https://cdn. jrc.ec.europa.eu/. earthdata.nasa.gov/conduit/upload/3865/ • Ecosystem Marketplace. 2016. “Raising Ambition: MODIS_C6_Fire_User_Guide_A.pdf. State of the Voluntary Carbon Markets 2016.” • Giglio, L., Csiszar, I., Justice, C.O. 2006. “Global Forest Trends’ Ecosystem Marketplace report, distribution and seasonality of active fires as Washington, DC. observed with the Terra and Aqua MODIS • FAO (United Nations Forest and Agricultural sensors.” Journal of Geophysical Research - Biogeosciences Organization). 2014. “Global Forest Resources 111, G02016, doi:10.1029/2005JG000142. Assessment 2015, Country Report: India.” FAO, • Giglio, L., Descloitres, J., Justice, C.O., Kaufman, Rome, Italy, http://www.fao.org/3/a-az238e.pdf. Y. 2003. “An enhanced contextual fire detection • FFMG (Australia and New Zealand’s Forest Fire algorithm for MODIS.” Remote Sensing of Environment Management Group). 2013. National Research 87:273-282. Priorities to 2020 and Beyond: National Bushfire • Government of India. 2006. Report of the National Management Policy Statement for Forests and Forest Commission. Rangelands. • Government of Jharkhand. 2014. “Jharkhand- • FSI (Forest Survey of India, Ministry of Action Plan on Climate Change.” Accessed 2017. Environment, Forest and Climate Change). http://www.moef.nic.in/ccd-sapcc. 2011. India State of Forest Report 2011. Forest Survey of India, Ministry of Environment, Forest and • Guhuthakurta, P., M. Rajeevan, D.R. Sikka, and Climate Change, Dehradun, Uttarakhand, Ajit Tyagi. 2014. “Observed Changes in Southwest February 2018, http://fsi.nic.in/details. Monsoon Rainfall Over India During 1901-2011.” php?pgID=sb_16. International Journal of Climatology, DOI: 10.1002/ joc.4095. • FSI. 2012. “Vulnerability of India’s Forests to Strengthening Forest Fire Management in India   210 • Guswanto, Eko Heriyanto. 2009. “Operational • Isaacs, Zeena. 2004. “National Fire Danger- Weather Systems for National Fire Danger.” Jurnal Rating System by June.” Engineering News, 9 Meteorlogi da Geofisika 10, no. 2, http://puslitbang. April 2004, http://www.engineeringnews.co.za/ bmkg.go.id/jmg/index.php/jmg/article/view/37. article/national-fire-dangerrating-system-by- • Guttman, N. B. 1999. “Accepting the Standardized june-2004-04-09/rep_id:4136. Precipitation Index: a calculation algorithm.” J. • IPCC (Intergovernmental Panel on Climate Amer. Water Resource Assoc. 35, no. 2: 311-322. Change). 2014. Climate Change 2014: Mitigation of • Hansen, M. C., P. V. Potapov, R. Moore, M. Climate Change, Contribution of Working Group III to the Fifth Hancher, S. A. Turubanova, A. Tyukavina, D. Assessment Report of the Intergovernmental Panel on Climate Thau, S. V. Stehman, S. J. Goetz, T. R. Loveland, Change. Seyboth, A. Adler, I. Baum, S. Brunner, A. Kommareddy, A. Egorov, L. Chini, C. O. P. Eickemeier, B. Kriemann, J. Savolainen, S. Justice, and J. R. G. Townshend. 2013. “High- Schlömer, C. von Stechow, T. Zwickel and J.C. Resolution Global Maps of 21st-Century Forest Minx, eds. Cambridge, UK: Cambridge University Cover Change.” Science 342 (15 November): Press. 850–53. Data available on-line from: http:// • Jhariya, M.K. S.S. Bargali, S.L. Swamy, B. earthenginepartners.appspot.com/science-2013- Kittur, Kiran Bargali, and Geeta V. Pawar. 2014. global-forest. “Impact of Forest Fire on Biomass and Carbon • Himachal Pradesh Forest Department. 2005. H. Storage Pattern of Tropical Deciduous Forests in P. Forest Sector Policy & Strategy, http://hpforest. Bhoramdeo Wildlife Sanctuary, Chhattisgarh.” nic.in/pages/display/NjVzZDRhhXQ0ZjY1cw==- International Journal of Ecology and Environmental Sciences 40, policylaws. no. 1: 57-74. • Hintze, Jerry L., and Ray D. Nelson. 1998. “Violin • Joseph, Shijo, K. Anitha, M.S.R. Murthy. 2009. Plots: A Box Plot-Density Trace Synergism.” The “Forest Fire in India: A Review of the Knowledge American Statistician 52, no. 2: 181-84. Base.” Journal of Forest Research 14: 127-134. • Hiremath, Ankila J. and Bharath Sundaram. 2005. • Keetch, John, and George Byram. 1968. A Drought “The Fire-Lantana Cycle Hypothesis in Indian Index for Forest Fire Control. U.S. Department of Forests.” Conservation and Society 3, no. 1: 26-42. Agriculture – Forest Service, Southeastern Forest Experiment Station, Asheville, North Carolina. • Hiremath, Ankila J. and Bharath Sundaram. 2013. “Invasive Plant Species in Indian Protected Areas: • Kerala Forest Department. n.d. “Kerala Forest Conserving Biodiversity in Cultural Landscapes” Statistics - 2016.” http://www.forest.kerala.gov.in/ in Plant Invasions in Protected Areas: Patterns, Problems, and images/pdf/cov1.pdf. Challenges, edited by L.C. Foxcroft et al. Dordrecht: • Kittur, B.H., S.L. Swamy, S.S. Bargali, and Manoj Springer. Kumar Jhariya. 2014. “Wildland Fires and • Holden, Sandra R. and Kathleen K. Treseder. Moist Deciduous Forests of Chhattisgarh, India: 2013. “A Meta-Analysis of Soil Microbial Biomass Divergent Component Assessment.” Journal of Responses to Forest Disturbances.” Frontiers in Forestry Research 25, no. 4: 857-866. Microbiology 4, DOI: 10.3389/fmicb.2013.00163. • Kodandapani, Narendran, Mark Cochrane, • HP (Himachal Pradesh Forest Department). 2015. and R. Sukumar. 2004. “Conservation Threat of Forest Manual, Volume IV (Technical Matters). Increasing Fire Frequencies in the Western Ghats, India.” Conservation Biology 18, no. 6: 1553-1561. • Ichoku, C. and L. Ellison. 2014. “Global Top- Down Smoke-Aerosol Emissions Estimation Using • Kodandapani, Narendran, Mark Cochrane, and Fire Radiative Power Measurements.” Atmospheric R. Sukumar. 2009. “Forest Fire Regimes and Chemistry and Physics 14: 6643-6667. Their Ecological Effects in Seasonally Dry Tropical Ecosystems in the Western Ghats, India” in Tropical • Ichoku, C. and Y.J. Kaufman. 2005. “A Method to Fire Ecology: Climate Change, Land Use and Ecosystem Dynamics, Derive Smoke Emission Rates from MODIS Fire edited by M.A. Cochrane, 335-354. Springer-Praxis, Radiative Energy Measurements.” IEEE Transactions DOI: 10.1007/978-3-540-77381-8_12. on Geoscience and Remote Sensing 43, no. 11: 2636-2649. 211 Strengthening Forest Fire Management in India • Konsam, Boddymore, Shyam Singh Phartyal, in Development PN-ACB-404, Washington, DC, Munesh Kumar, and Nagendra Prasad Todaria. United States. 2017. “Life After Fire for Understorey Plant • Madhya Pradesh Forest Department. 2005. Community in Subtropical Chir Pine Forest of Madhya Pradesh State Forest Policy, http://mfp. Garhwal Himalaya.” The Indian Forester 143, no. 8: mpforest.org/eco/pdf/forestpolicy.pdf. 759-766. • Maithani, G.P., V.K. Bahuguna, and Pyare Lal. • Khaki, Bilal A., V.R.R. Singh, Akhlag Q. Wani 1986. “Effect of Forest Fires on the Ground and Raj K. Thakur. 2015. “Effect of Forest Fire on Vegetation of a Moist Deciduous Sal (Shorea Soil Nutrients in Blue Pine (Pinus Wallichiana A.B. Robusta) Forest.” The Indian Forester 112, no. 8: 646- Jackson) Ecosystems.” The Indian Forester 141, no. 4: 678. 355-360. • Jaganadha Matta. 2009. “Rebuilding rural India: • Krishna, P. Hari , and C. Sudhakar Reddy. 2012. potential for further investments in forestry and “Assessment of increasing threat of forest fires in green jobs.” Unasylva 233, vol. 60. http://www.fao. Rajasthan, India using multi-temporal remote org/docrep/012/i1025e/i1025e07.htm. sensing data (2005–2010).” Current Science, May 10. • Mbatha, K .R. , and D. Ward. 2010. “The Effects of • Kulik, Aleksander A. and Dmitry Krasovsky. Grazing, Fire, Nitrogen and Water Availability on 2017. “Forest Fire Protection in the Republic Nutritional Quality of Grass in Semi-Arid Savanna, of Belarus.” Presentation and discussion at the South Africa.” J. Arid Environ. 74: 1294-1301. International Workshop on Forest Fire Prevention • McGuffog, Tim. 2017. “Fire Prevention.” and Management organized by MoEFCC and the Presentation and discussion at the International World Bank, New Delhi, India, 2 November 2017. Workshop on Forest Fire Prevention and • Kumar, K. Krishna, B. Rajagopalan, M. Cane, 1999: Management organized by MoEFCC and the On the Weakening Relationship Between the Indian World Bank, New Delhi, India, 2 November 2017. Monsoon and ENSO. Science, 284, 2156 -2159. • Merrill, D.F. and M.E. Alexander, eds. 1987. • Kumar, K. Krishna, B. Rajagopalan, M. Hoerling, Glossary of Forest Fire Management Terms. National G. Bates, M. Cane, 2006: Unraveling the Research Council of Canada, Ottawa, Pub. NRCC Mystery of Indian Monsoon Failure During El No. 26516. Niño. Science, 314, 115-119. • MoEF (Ministry of Environment and Forest, • Lü, X.T., Lü, F.M., Zhou, L.S., Han, X., and Government of India). 1988. National Forest Han, X .G.. 2012. “Stoichiometric Response of Policy. MoEF Resolution No. 3-1/86-FP, issued 7 Dominant Grasses to Fire and Mowing in a Semi- December 1988. Arid Grassland.” J. Arid Environ. 78: 154-160. • MoEF. 1999. National Forestry Action Programme-India, • Lundgren, S. 1999. “The national fire Volume I: Status of Forestry in India. New Delhi, India. management analysis system (NFMAS) past 2000: • MoEF. 2006. Report of the National Forest Commission. A new horizon” in Proceedings of the Symposium on Fire New Delhi, India. Economics, Planning and Policy: Bottom Lines, edited by A. González-Cabán and P.N. Omi. Gen. Tech. Rep. • MoEF. 2009. “India Forestry Sector Outlook.” PSW-GTR-173, USDA Forest Service, Pacific Asia Pacific Forestry Sector Outlook Study II Southwest Research Station, Albany, CA. Working Paper Series, Working Paper No. APFSOS II/WP/2009/06, United Nations Food and • Lutakamale, Albert, Shubi Kaijage. 2017. “Wildfire Agricultural Organization Regional Office for Asia Monitoring and Detection System Using Wireless and the Pacific, Bangkok, Thailand, 2009, http:// Sensor Network: A Case Study of Tanzania.” www.fao.org/docrep/014/am251e/am251e00.pdf. Wireless Sensor Network 9: 274-89. • MoEF. 2009b. The Guidelines on State • Lynch, Owen. 1992. “Securing Community- Compensatory Fund Management and Planning Based Tenurial Rights in the Tropical Forests of Authority (State CAMPA), Dew Delhi, India, 2 Asia: An Overview of Current and Prospective July 2009, http://envfor.nic.in/sites/default/files/ Strategies.” World Resources Institute, Issues CAMPA-guidelines_0.pdf. Strengthening Forest Fire Management in India   212 • MoEF. Undated. “Joint Forest Management: A by Food and Agriculture Organization of the Handbook.” Report from JICA-MoEF project on United Nations, Asia-Pacific Forestry Commission, “Capacity Development for Forest Management and Forest Department of Sri Lanka, Colombo, and Training of Personnel,” New Delhi, India, Sri Lanka, 15-25 May 2017, http://www.fao.org/ http://ifs.nic.in/Dynamic/pdf/JFM%20handbook. fileadmin/templates/rap/files/meetings/2017/49_ pdf. Day9.pdf. • MoEFCC (Ministry of Environment, Forest • Narayanaraj, G., and M.C. Wimberly. 2011. and Climate Change, Government of India). “Influences of forest roads on the spatial pattern of 2014. National Working Plan Code – 2014 wildfire boundaries.” International Journal of Wildland (For Sustainable Management of Forests and Fire 20: 792-803, doi:10.1071/WF10032. Biodiversity in India). • Natural Resources Canada. 2016. “The Canadian • MoEFCC. 2017. Guidelines for Investment Policy Forest Fire Danger Rating System,” http://www. and Procedure for Investment of Funds Lying nrcan.gc.ca/forests/fire-insects-disturbances/ with Ad-Hoc CAMPA, revised with effect 16 fire/14470. November 2017, http://envfor.nic.in/sites/default/ • Nayak, Bhibhu Prasad, Priyanka Kohli, J.V. files/Guidelines_for_Investment_Policy_and_ Sharma. 2013. “Livelihood of Local Communities Procedure_0_0.pdf. and Forest Degradation in India: Issues for • MoEFCC-GIZ (Indian Ministry of Environment, REDD+.” Policy brief for “Developing Country Forest and Climate Change, Government of India, Participation in Addressing Climate Change: and Deutsche Gesellshaft für Internationale Analyzing Issues and Options for Implementing Zusammenarbeit). 2014. “The Economics of NAMAs and REDD” project, Norwegian Ministry Ecosystems and Biodiversity India Initiative: of Foreign Affairs (MFA) and The Energy and Interim Report – Working Document.” MoEFCC Resources Institute (TERI), New Delhi, India. and Indo-German Biodiversity Programme, GIZ, • NCA (National Commission on Agriculture). 1976. New Delhi, India. Report of the National Commission on Agriculture, • Mohanty, Sthitapragyan , Prashanta K. Patra, Part IX, Forestry. Ministry of Agriculture and Sudhansu S. Sahoo, and Asit Mohanty. 2017. Irrigation, Government of India, New Delhi. “Forecasting of solar energy with application • NFC (National Forestry Commission). 2006. for a growing economy like India: Survey and Report of the National Forest Commission. implication.” Renewable and Sustainable Energy Reviews Ministry of Environment and Forests, Government 78: 539-553. of India, New Delhi. • Mondal, N. and R. Sukumar. 2015. “Regeneration • Ninan, K.N., A. Kontoleon. 2016. “Valuing Forest of Juvenile Woody Plants After Fire in a Seasonally Ecosystem Services and Disservices – Case Study Dry Tropical Forest of Southern India.” of a Protected Area in India.” Ecosystem Services 20: BIOTROPICA 47, no. 3: 330-338. 1-14. • Mondal, N. and R. Sukumar. 2016. “Fires in • Nolasco Morales, Alfredo. 2017. Presentation and Seasonally Dry Tropical Forest: Testing the discussion at the International Workshop on Forest Varying Constraints Hypothesis across a Regional Fire Prevention and Management organized by Rainfall Gradient.” PLOS ONE 11(7): e0159691. MoEFCC and the World Bank, New Delhi, India, doi:10.1371/journal.pone.0159691 November 2017. • Nagbhushanam, D. 2013. Working Plan for • Nolasco Morales, Alfredo. 2017. “Mexico’s National Hyderabad Forest Department, 2013-14 to 22-23. Forest Fire Management Program,” http://blogs. Hyderabad, Telangana, India. worldbank.org/endpovertyinsouthasia/mexico-s- • Nair, CTS. 2017. “Joint Forest Management in national-forest-fire-management-program. India,” Presentation at 10th Executive Forest • OneWorld (OneWorld Foundation India). Policy Course, Revisiting the Poverty Reduction 2010. “Case Study: Fire Alert Messaging System Agenda in the Context of the SDGs: Opportunities in Madhya Pradesh.” Report for Governance and Challenges for Asia-Pacific Forestry, organized 213 Strengthening Forest Fire Management in India Knowledge Centre, Department of Administrative Science and Technology, Environment and Forests. Reforms and Public Grievances, Ministry of 2016. “Forest Fires and its effect on Environment, Personnel, Public Grievances and Pensions, Forests, Bio-diversity and Wildlife and Remedial/ Government of India. Preventive Measures.” • Oxford, Adam. 2017. “Wildfires Versus Big Data: • Ramakrishnan, P.S. 2007. “Traditional Forest How South Africa’s 49-Year-Old Fire Warning Knowledge and Sustainable Forestry: A North- System Has Stood the Test of Time.” htxt. africa, 2 East India Perspective.” Forest Ecology and Management February 2017, https://www.htxt.co.za/2017/02/02/ 249: 91-99. wildfires-versus-big-data-how-south-africas-49- • Ramakrishnan, P.S. and P.M. Vitousek. 1989. year-old-fire-warning-system-has-stood-the-test- “Ecosystem Level Processes and the Consequences of-time-but-is-slowly-being-replaced/. of Biological Invasion” in Biological Invasions: A Global • Pai, Rekha, Ankila J. HIremath and Umakant. Perspective, edited by J.A. Drake, H.A. Mooney, F. 2007. “Rethinking Forest Fires: Proceedings Di Castri, R.H. Grove, K.J. Kruger, M. Rejmanek, of the National Workshop on Forest Fires.” and M. Williamson, 281-300. Chichester, United Workshop organized by Government of India and Kingdom: Wiley and Sons. the Ashoka Trust for Research in Ecology and the • Reddy, C. Sudhakar, C.S. Jha, P.G. Diwakar, and Environment, New Delhi, India, 13-14 November V.K. Dadhwal. 2015. “Nationwide Classification 2007. of Forest Types of India Using Remote Sensing • Parashar, Amit and Sas Biswas. 2003. “The Impact and GIS.” Environmental Monitoring and Assessment, DOI: of Forest Fire on Forest Biodiversity in the Indian 10.1007/s10661-015-4990-8. Himalayas (Uttaranchal),” XII World Forestry • Reddy, C. Sudhakar, V.V.L. Padma Alekhya, K.R.L. Congress, Quebec City, Canada, Report 0358-B1, Saranya, K. Athira, C.S. Jha, P.G. Diwakar, and V.K. available at http://www.fao.org/docrep/ARTICLE/ Dadhwal. 2017a. “Monitoring of Fire Incidences WFC/XII/0358­B1.HTM. in Vegetation Types and Protected Areas of India: • Poffenberger, Mark, S.K. Barik, Dhrupad Implications on Carbon Emissions.” J. Earth Syst. Sci. Choudhury, Vincent Darlong, Vishal Gupta, S. 126: 11. Palit, Ivan Roy, Ibobi Singh, B.K. Tiwari, Sanjay • Reddy, C. Sudhakar, C.S. Jha, G. Manaswini, V.V.L. Upadhyay. 2006. “Communities and Forest Padma Alekhya, S. Vazeed Pasha, K.V. Satish, Management in Northeast India.” Background P.G. Diwakar, V.K. Dadhwal. 2017b. “Nationwide Paper No. 12 for study on Development and Assessment of Forest Burnt Area in India using Growth in Northeast India: The Natural Resourcesat-2 AWiFS Data.” Current Science 112, no. Resources, Water, and Environment Nexus, World 7: 1521-1532. Bank, March 2006. • Riley, Shawn J., Stephen D. DeGloria, and Robert • , Stephen. 1994. “India’s Cycle of Fire.” Elliot. 1999. “A terrain ruggedness index that Environmental History Review 18, no. 3: 1-20. quantifies topographic heterogeneity.” Intermountain • Raghuveer, P. 2017. “Forest Fire Prevention and Journal of Sciences 5, no. 1-4: 23-27. Management – Experiences from Telangana.” • Rodgers, W.A. 1986. “The Role of Fire in The Presentation and discussion at the International Management of Wildlife Habitats: A Review.” The Workshop on Forest Fire Prevention and Indian Forester 112, no. 10: 845-855. Management organized by MoEFCC and the • Roxy, Mathew Koll, Kapoor Ritika, Pascal Terray, World Bank, New Delhi, India, 3 November Raghu Murtugudde, Karumuri Ashok, and B.N. 2017. Goswami. 2015. “Drying of Indian Subcontinent • Rai, S.N. and Alok Saxena. 1997. “The Extent of by Rapid Indian Ocean Warming and a Weakening Forest Fire, Grazing, and Regeneration Status in Land-Sea Thermal Gradient.” Nature Communications Inventories Forest Areas in India.” The Indian Forester 6: 7423. 123, no. 8: 689-702. • Roy, P.S. 2004. “Forest Fore and Degradation • Rajya Sabha. Parliament of India. Department- Assessment Using Satellite Remote Sensing and related Parliamentary Standing Committee on Strengthening Forest Fire Management in India   214 Geographic Information System” in Satellite Remote Thesis, HNB Garhwal University, Srinagar- Sensing and GIS Applications in Agricultural Meteorology, edited Garhwal. by M.V.K. Sivakumar, P.S. Roy, K. Harmsen, and • Semwal, R.L. and J.P. Mehta. 1996. “Ecology of S.K. Saha, 361-400. Geneva, Switzerland: World Forest Fires in Chir Pine (Pinus roxburghii Sarg.) Forests Meteorological Organisation. of Garhwal Himalaya.” Curr. Sci. 70: 426-427. • Saha, Sonali and Ankila Hiremath. 2003. • Settele, J., R. Scholes, R. Betts, S. Bunn, P. Leadley, “Anthropogenic Fires in India: A Tale of Two D. Nepstad, J.T. Overpeck, and M.A. Taboada, Forests.” Aridlands Newsletter 54, November/ 2014: Terrestrial and inland water systems. 2014. December, https://cals.arizona.edu/OALS/ALN/ “Terrestrial and Inland Water Systems” in Climate aln54/saha.html. Change 2014: Impacts, Adaptation, and Vulnerability. Part A: • Saha, Sonali and Henry F. Howe. 2003. “Species Global and Sectoral Aspects. Contribution of Working Group II Composition and Fire in a Dry Deciduous Forest.” to the Fifth Assessment Report of the Intergovernmental Panel on Ecology 84, no. 12: 3118-3123. Climate Change, edited by Field, C.B., V.R. Barros, • Saranya, K.R.L., C. Sudhakar Reddy, and P.V.V. D.J. Dokken, K.J. Mach, M.D. Mastrandrea, Prasada Rao. 2016. “Estimating Carbon Emissions T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, from Forest Fires Over a Decade in Similipal R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. Biosphere Reserve, India.” Remote Sensing Applications: MacCracken, P.R. Mastrandrea, and L.L. White, Society and Environment 4: 61-67. 271-359. Cambridge University Press, Cambridge, United Kingdom and New York, NY. • Satendra, and A.D. Kaushik. 2014. Forest Fire Disaster Management. New Delhi: National Institute of • Shakesby, R.A. and S.H. Doerr. 2006. “Wildfire Disaster Management, Ministry of Home Affairs. as a Hydrological and Geomorphological Agent.” Earth-Science Reviews 74: 269-307. • Saxena, Alok. 2012. “Wildfire Risk Management in India and the Community” in Wildfire and Community: • Shilla, U., and B.K. Tiwari. 2015. “Impact of Facilitating Preparedness and Resilience, edited by Douglas fire and grazing on plant diversity of grassland Paton and Fantina Tedim, 129-49. Springfield, ecosystem of Cherrapunjee.” Keanean Journal of Illinois: Charles C Thomas. Science, 67-78. • Schmerbeck, J. and P. Fiener. 2015. “Wildfires, • Simpson, B. 2017. “The Canadian Forest Fire Ecosystem Services, and Biodiversity in Tropical Danger Rating System,” http://blogs.worldbank. Dry Forest in India.” Environmental Management 56: org/endpovertyinsouthasia/canadian-forest-fire- 355-372. danger-rating-system. • Schmerbeck, J., A. Kohli, K. Seeland. 2015. • Singh, J.S., Y.S. Rawat, and O.P. Chaturvedi. 1984. “Ecosystem Services and Forest Fires in India “Replacement of Oak Forest with Pine in the – Context and Policy Implications from a Case Himalaya Affects the Nitrogen Cycle.” Nature 311: Study in Andhra Pradesh.” Forest Policy and Economics 54-56. 50: 337-346. • Singh, Ombir. 2017. “FRI Initiatives of Fire Fighter • Schnepf, Chris, Andy Perleberg, Max Bennett, Safety & Equipment and Fire Suppression.” Bob Parker, Stephen Arthur Fitzgerald, Marty Presentation and discussion at the International Main, Ronald Mahoney. 2010. “Fire Behavior Workshop on Forest Fire Prevention and Basics” in Reducing Fire Risk on Your Forest Property. Management organized by MoEFCC and the Oregon State University Extension Service, World Bank, New Delhi, India, 2 November 2017. PNW 618, Corvallis, Oregon, October 2010. • Sinha, T.A.K. 2017. “Forest Fire Protection in https://ir.library.oregonstate.edu/concern/ Odisha.” Presentation and discussion at the administrative_report_or_publications/ International Workshop on Forest Fire Prevention ws859f89n?locale=en and Management organized by MoEFCC and the • Semwal, R.L. 1990. “Effect of Fire on Pine (Pinus World Bank, New Delhi, India, 2 November 2017. roxburghii Sar.) Forest and a Grazing Land (Charysopogon • Smith, Ross. 2017. “Fire Danger Rating Systems.” aciculatus Trinn.) of Garhwal Himalaya.” D. Phil. Background note to World Bank assessment of 215 Strengthening Forest Fire Management in India forest fire prevention and management in India, Rating: the Canadian Experience.” International New Delhi, April 2017. Journal of Wildland Fire 15: 121-35. • Sommers, W.T., R.A. Loehman, and C.C. Hardy. • Telangana (Telangana Forest Department). 2015. 2014. “Wildland Fire Emissions, Carbon, and Telangana State of Forest Report 2015. Government of Climate: Science Overview and Knowledge Telangana, Hyderabad. Needs.” Forest Ecology and Management 317: 1-8. • Thekaekara, Tarsh , Abi Tamim Vanak, Ankila • Springate-Baginski, Oliver and Piers Blaikie. J Hirema, Nitin D Rai, Jayashree Ratnam, and 2007. “Annexation, Struggle and Response: Raman Sukumar. 2017. “Notes from the Other Forest, People and Power in India and Nepal” in Side of a Forest Fire.” Economic & Political Weekly Vol. Forests, People and Power: The Political Ecology of Reform in LII Nos. 25 & 26. South Asia, edited by Oliver Springate-Baginski and • Tiwari, B.K., H. Tynsong, M.M. Lynrah, E. Piers Blaikie, 27-50. New York: Earthscan. Lapasam, S. Deb, D. Sharma. 2013. “Institutional • Standing Committee (Department-related Arrangement and Typology of Community Forests Parliamentary Standing Committee on Science of Meghalaya, Mizoram, and Nagaland of North- & Technology, Environment & Forests). 2016. East India.” Journal of Forestry Research 24, no. 1: 179- “Forest Fires and its Effect on Environment, 186. Forests, Biodiversity and Wildlife and Remedial/ • Tiwari, B.K., H. Tynsong, M.B. Lynser. 2010. Preventive Measures.” Report No. 293, Parliament “Forest Management Practices of the Tribal People of India, Rajya Sabha, 16 December 2016. of Meghalaya, North-East India.” Journal of Tropical • Sundaram, Bharath, Siddhartha Krishnan, Ankila Forest Science, July 2010: 329-342. J. HIremath, and Gladwin Joseph. 2012. “Ecology • Tuomi, M., T. Thum, H. Järvinen, S. Fronzek, and Impacts of the Invasive Species, Lantana camara, B. Berg, M. Harmon, J.A. Trofymow, S. Sevanto, in a Social-Ecological System in South India: and J. Liski. 2009. “Leaf Litter Decomposition – Perspectives from Local Knowledge.” Human Ecology Estimates of Global Variability Based on Yasso07 40, no. 6: 931-42. Model.” Ecological Modelling 220: 3362-3371. • Suresh, H.S. 2017. “Forest Fires: Need for • UN ECLAC (United Nations Economic Rethinking Management Strategies.” End Commission for Latin American and the Poverty in South Asia, World Bank blog, 8 Caribbean). 2014. Handbook for Disaster Assessment. December 2017, http://blogs.worldbank.org/ Santiago, Chile: United Nations. endpovertyinsouthasia/forest-fires-need- • US DOI-DOA (Departments of Interior and rethinking-management-strategies. Agriculture, United States). 2014. The National • Srivastava, P. and A. Garg. “Forest Fires in India: Strategy: The Final Phase in the Development of the Regional and Temporal Analyses.” Journal of Tropical National Cohesive Wildland Fire Management Forest Science 25, no. 2: 228-239. Strategy, Washington, DC, April 2014, • Stambler, Kimberly and Joseph Barbara. 2011. h t t p s : / / w w w. f o r e s t s a n d r a n g e l a n d s . “Engineering the Incident Command and gov/strategy/documents/strategy/ Multiagency Coordination Systems.” Journal of CSPhaseIIINationalStrategyApr2014.pdf. Homeland Security and Emergency Management 8, no. 1, • Uttarakhand Biodiversity Board. 2017. “Study DOI: 10.2202/1547-7355.1838. on ‘Science-Policy Interface’ Focusing on Access • Swetnam, T.W. and J.L. Betancourt. 1990. “Fire- & Benefit Sharing (ABS) and Transboundary Southern Oscillation Relations in the Southwestern Issues.” Kailash Sacred Landscape Conservation United States.” Science 249: 1017-1020. and Development Initiative, Uttarakhand • Tamil Nadu Forest Department. 2017. “Spatial Biodiversity Board, Dehradun, India, http://www. & Temporal Analysis of Decadal Forest Fire Data sbb.uk.gov.in/files/books/KSL_Study_Report- (2006-2015).” Burfal-15May2017-Small-Merge.pdf. • Taylor, S.W. and M.E. Alexander. 2006. “Science, • Valappil, Narayanan I., and K. Swarupanandan. Technology, and Human Factors in Fire Danger 1996. “Regeneration Dynamics and Sylvigenesis in Strengthening Forest Fire Management in India   216 the Moist Deciduous Forests of Southwest India.” Africa.” Report prepared for Department of Water New Forests 11, no. 3: 185-205. Affairs and Forestry, South Africa, ENV-P-C 2000- • van Leeuwen, T.T., G.R. van der Werf, A.A. 073. Hoffmann, R.G. Detmers, G. Rücker, N.H.F. • World Bank. 2005. “India: Unlocking French, S. Archibald, J.A. Carvalho Jr., G.D. Cook, Opportunities for Forest-Dependent People in W.J. de Groot, C. Hély, E.S. Kasischke, S. Kloster, India, Volume II: Appendixes.” World Bank J.L. McCarty, M.L. Pettinari, P. Savadogo, E.C. Report No. 34481-IN, Washington, DC, 28 Alvarado, L. Boschetti, S. Manuri, C.P. Meyer, F. December 2005. Siegert, L.A. Trollopse, W.S.W. Trollope. 2014. • World Bank. 2006. “India: Unlocking “Biomass Burning Fuel Consumption Rates: A Opportunities for Forest-Dependent People in Field Measurement Database.” Biogeosciences 11: India, Volume I: Main Report.” World Bank 7305-7329. Report No. 34481-IN, Washington, DC, 6 • Venkataraman, C., G. Habib, D. Kadamba, February 2006. M. Shrivastava, J.F. Leon, B. Crouzille, O. • World Bank. 2015. Kazakhstan Forest Protection Boucher, D.G. Streets. 2006. “Emissions from and Reforestation Project, http://documents. Open Biomass Burning in India: Integrating worldbank.org/curated/en/298651468195014908/ the Inventory Approach with High-Resolution pdf/ICR3561-ICR-KZ-P078301-P087485-Box- Moderate Resolution Imaging Spectroradiometer 394380B-PUBLIC-12-30-2015.pdf. (MODIS) Active-Fire and Land Cover Data.” Global • World Bank. 2016. “The cost of fire: an economic Biogeochemical Cycles 20, DOI:10.1029/2005GB002547. analysis of Indonesia’s 2015 fire crisis.” Indonesia • Verma, Satyam and S. Jayakumar. 2012. “Impact sustainable landscapes knowledge, note no. of Forest Fire on Physical, Chemical and Biological 1, http://documents.worldbank.org/curated/ Properties of Soil: A Review.” Proceedings of the en/776101467990969768/The-cost-of-fire-an- International Academy of Ecology and Environmental Sciences 2, economic-analysis-of-Indonesia-s-2015-fire-crisis. no. 3: 168-176. • Wunder, Sven, Arild Angelsen, Brian Belcher. • Willis, Carla, Brian van Wilgen, Kevin Tolhurst, 2014. “Forests, Livelihoods, and Conservation: Colin Everson, Peter D’Abreton, Lionel Pero, Broadening the Empirical Base.” World Development Gavin Fleming. 2001. “The Development of a 64, Supplement 1: S1-S11. National Fire Danger Rating System for South 217 Strengthening Forest Fire Management in India