(2018) Charbon de bois en Haïti : Une évaluation nationale des tendances de production et de consommation de charbon de bois

A National Assessment of Charcoal Production and Consumption Trends November 2018 9979_Charcoal_Haiti_CVR.indd 3 1/16/19 11:02 AM Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Andrew Tarter, Ph.D., anthropologist (principal investigator; lead author) Katie Kennedy Freeman, agricultural economist (World Bank Group) Christopher Ward, consultant (World Bank Group) Klas Sander, Ph.D., natural resources economist (World Bank Group) Kenson Theus, sociologist, Haiti Takes Root (research team leader, J/P HRO) Barbara Coello, economist (World Bank Group) Yarine Fawaz, economist (CEMFI) Melinda Miles, anthropologist, Haiti Takes Root (J/P HRO) Tarig Tagalasfia G. Ahmed, consultant, Humphrey Fellow (Cornell University) Financing for this study was provided by the Program on Forests (PROFOR) CHARCOAL in Haiti A National Assessment of Charcoal Production and Consumption Trends 9979_Charcoal_Haiti.indd 1 2/6/19 10:38 AM © 2017 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. 9979_Charcoal_Haiti.indd 2 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends iii ACKNO WLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v EXECUTIVE SUMMAR Y AND KEY FINDINGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii I. INTRODUCTION AND BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Intr oduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Backgr ound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Haitian Go vernment Legislative and Executive Efforts to Address Deforestation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 The Original For ests of Haiti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 The Defor estation of Haiti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Curren t Estimates of Arboreal Coverage in Haiti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 The Historical Production o f Charcoal in Haiti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 The Use o f Non-Charcoal Wood in Haiti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 The Historical Supply of Charcoal to Port-au-Prince . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 Contempor ary Charcoal Production in Port-au-Prince . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 II. RESEARCH QUESTIONS AND METHODOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Resear ch Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Principal Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Methodological Similarities and Diff erences from Previous Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Timeframe o f the Fieldwork Scoping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Spatial Sampling: Cor e and Periphery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Data Standar dization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 III. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Descripti ve Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Differ ences between the Three Sampling Periods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Comparison of All Three Sampling Periods Based on One Shared Day . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Comparison of All Sampling Periods Based on Total Shared Hours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Reconstruction o f the Peak Week . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 Robustness Check on Number of Trucks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 Variation within Days and across the Week . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Regional Diff erences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 The Eff ects of the Farthest Periphery Stations and Feeder Roads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 The Southern Peninsula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Centr al Plateau . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 North . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 East o f Port-au-Prince . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Contents 9979_Charcoal_Haiti.indd 3 2/6/19 10:38 AM iv Charcoal in Haiti Charcoal Entering Haiti from the Dominican Republic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Estimates of Charcoal Entering Haiti from the Dominican Republic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Alternativ e Hypotheses Concerning Charcoal Entering from the Dominican Republic . . . . . . . . . . . . . . . . . . . . . . . . . 30 IV. ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33 Estimated Annual Consumption in Port-au-Prince . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Changes in the R egional Supply of Charcoal to Port-au-Prince . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Estimated Annual Consump tion at the National Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Economic Value of the Charcoal Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Charcoal as Compar ed to GDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Charcoal R elated to Other Commodities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Employmen t in the Charcoal Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 V. IMPACT OF MATTHEW ON CHARCOAL PRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Hurricane Matthe w . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41 VI. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 R esearch Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Explaining Increases in Pr oduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 The Sustainability of Haitian Charcoal Woodlot Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Social and Ecological Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Impro ving Existing Woodlot Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 VII. POLICY IMPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 The Charcoal Industry in the National Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Implication of Charcoal Production Stigma on Policies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Policy Designs Inhibited by a L ack of Data on Charcoal Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 VIII. AREAS OF ADDITIONAL RESEARCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 U ndertake Additional Research on Key Facets of the Charcoal Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Undertak e Additional Work in the Policy Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 IX. REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 X. ANNEXES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59 Annex 1—Fieldwork Timeframe Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Annex 2 —Logistical Details about the Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Annex 3—P ossible Sources of Errors and Adopted Mitigation Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Annex 4—Complementary Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Annex 5—Post-Hurricane Matthew Arboreal Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Annex 6— The Gendered Aspects of Charcoal Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 9979_Charcoal_Haiti.indd 4 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends v Acknowledgments The report is a product of the Agriculture Global Practice of the World Bank. It was prepared by a team led by Katie Kennedy Freeman (Agriculture) and Klas Sander (Environment) and authored by anthropologist and expert consultant Andrew Tarter, PhD. Research was led by a team from nonprofit organization J/P HRO in conjunction with the Haiti Takes Root initiative. The final report considers feedback from multiple experts across various institutions. The authors would like to thank the peer reviewers, Dana Rysankova (Senior Energy Specialist), Joanne Gaskell (Senior Agriculture Economist), Andrew Mitchell (Senior Forestry Specialist), and Erika Felix (Bioenergy Specialist, Food and Agriculture Organization of the UN (FAO)). In addition, the team would like to thank Raju Singh (Program Leader), Pierre Xavier Bonneau (Program Leader), Preeti Ahuja (Practice Manager Agriculture), and Valerie Hickey (Practice Manager Environment) for their support and guidance. The team acknowledges and thanks the Program on Forests (PROFOR) for supporting this research and publication. 9979_Charcoal_Haiti.indd 5 2/6/19 10:38 AM 9979_Charcoal_Haiti.indd 6 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends vii passing trucks (August 2017, October 2017, and December 2017) that amounted to a total of 384 hours of observations, registering 10,404 unique enumerations of charcoal vehicles by 69 enumerators placed at 23 different roadside stations that controlled charcoal vehicles at multiple intersections of roads or maritime wharfs leading into Port-au-Prince from every direction of the country. 1 The estimations and subsequent extrapolations presented here are conservative, using midrange estimates on a number of variables, including charcoal bag carrying capacities for different-sized vehicles in the classificatory typology, an average weight assumption for charcoal bags, and the utilization of annual extrapolation methods (for Port-au-Prince and all of Haiti) based on extending data sampled during representative low and peak periods of charcoal production to corresponding low and peak seasons across the entire year. This research provides targeted answers to a narrow set of research questions, helping to fill an important information gap in Haiti. Most notably, the total volume of charcoal moving into Port-au-Prince has implications on the total required volume of primary production of biomass for charcoal and the total value of the charcoal value chains, demonstrating the magnitude of importance of charcoal production for Haiti. These two up-to- date figures may inform policy decisions for development and government programming related to landscape management, reforestation, tree planting, agroforestry, and agricultural projects in Haiti. 1 The research design also controlled for every known and relevant vehicular point of entry from the Dominican Republic at the time of the research. The authors acknowledge that charcoal may be entering Haiti from Route National Number 6, along the northern coast of Haiti, or through wharfs along the northern coast of Haiti (the only known major entry routes not controlled for). However, such charcoal would most certainly be headed to the nearby city of Cap Haitian, or other urban towns north of Haiti. A widely cited report from 1979 suggested that existing wood supplies in Haiti would be enough to meet increasing charcoal demand until around the year 2000, but that ongoing charcoal production could result in an environmental ‘apocalypse’ (Voltaire 1979, 21, 23) The prediction that wood supplies in Haiti would be exhausted by 2000 was also supported by a report on trends emerging from early remote sensing analyses of aerial photographs spanning from 1956 to 1978, for three different locations in Haiti (Cohen 1984, v–iv). And yet, some 40 years later, Haitians continue to produce large quantities of charcoal despite these dire predictions to the contrary. The research presented in this report directly addresses important and unresolved questions stemming from the unexpected fact that Haitians continue to meet approximately 80 percent of their national energy needs through firewood and charcoal production: 1. How much charcoal is consumed annually in the capital city of Port-au-Prince? 2. Which geographical regions produce the charcoal consumed in the capital? 3. How do these production areas variably supply charcoal to the capital? 4. In what ways have these trends changed over time? and 5. What percentage of charcoal is originating from the bordering Dominican Republic? This report draws on research spanning nearly half a century to answer these questions, presenting both longitudinal and cross-sectional data related to multiple aspects of charcoal production and consumption in Haiti. Data collection spanned two years, commencing in 2016 with literature reviews, key informant interviews, and regional scouting trips across Port- au-Prince and Haiti to identify the best locations to position research teams with the objective of enumerating passing charcoal trucks and boats. These preliminary stages were followed by three different periods of roadside sampling to count Executive Summary and Key Findings 9979_Charcoal_Haiti.indd 7 2/6/19 10:38 AM viii Charcoal in Haiti Key Findings and Conclusions The data and analyses permitted conclusions and estimates that help contextualize the enormity of charcoal production in Haiti in terms of scale, geographic scope, economic impact, and employment generation. All key findings and conclusions briefly summarized here are supported in more detail in the main body of the report. The main findings of the research are: • Approximately 438,000 metric tons of charcoal are consumed annually in Port-au-Prince. The annual charcoal consumption range for Port-au- Prince is based on 24/7 counts at six key enumeration stations controlling for the largest known charcoal entry points into the capital. Extrapolating from two weeks of sampling data (one week of sampling data from a peak charcoal production season and one week from a low production season), an estimated 352,014 to 524,394  metric tons, with a midrange estimate of 438,204 metric tons of charcoal are consumed annually in Port-au-Prince. • Approximately 946,500 metric tons of charcoal are consumed nationally in Haiti each year. The estimated annual charcoal consumption range for Haiti at the national level is based on a tons-to-population ratio created from the annual estimate for Port-au- Prince—by far the largest city and largest consumer of charcoal in the country. The tons-to-population ratio was applied to the entire urban population of Haiti 2 (inclusive of Port-au-Prince) as last reported by the Haitian government (IHSI 2015), suggesting a range between 759,470 and 1,133,537 metric tons, with an estimated midpoint of 946,506 metric tons of charcoal consumed nationally in Haiti each year. • Charcoal production is the second-largest agricultural value chain in Haiti. The charcoal sector’s outsized influence on Haiti’s economy is evidenced by its size relative to other agricultural commodities. It is the second largest agriculture value chain in the country, behind only mangoes, and dwarfing most other traditional pillars of the Haitian 2 Charcoal is primarily produced rurally and consumed in urban locations. In rural locations Haitian use firewood for cooking. rural economy, such as yams, bananas, beans, avocados, coffee, sugarcane, and corn. • Contrary to popular conception, the data show that a negligible amount of the charcoal consumed in Port-au-Prince, Haiti, originates in the Dominican Republic. Across five different enumeration stations controlling for charcoal entering Haiti from the Dominican Republic, the total amount of charcoal observed in the research presented in this report is equivalent to 2.48 percent of the amount consumed in Port-au-Prince during that same period. 3 Other key findings include: • Researchers are approaching consensus that current arboreal land coverage in Haiti is significantly higher than previously believed. Since the early 1980s, the Haitian government has stated that in ideal conditions (given realities of topography, geology, and meteorology) some 35–55  percent of the land surface of Haiti should be covered forests. Five recent land-cover studies have concluded that Haiti has a much higher than conventionally reported level of tree cover and/or forest cover, 4 with many reports estimating present tree coverage in Haiti at approximately 30 percent of the land surface. • There are clear annual low and peak seasons of charcoal production in Haiti. Evidence presented here demonstrates that charcoal production fluctuates significantly during the course of a year; there are clear peak and low seasons of charcoal production, typically lasting approximately six months each. Unexpected events such as disasters, droughts, agricultural pests, tropical storms, or political unrest—which disrupt the traditional agricultural calendar—may shift, shorten, or extend low and peak charcoal production seasons 3 These data are supported by other research, suggesting that the percentage of official charcoal exports from the Dominican Republic to Haiti has declined significantly over the last two decades, while Dominican exports to overseas markets have surged. In 2001, Haiti received over 50 percent of official Dominican charcoal exports, which were valued at only US$4k. By 2012, the value of Dominican charcoal exports had grown to between US$500,000 to US$1,200,000, and exports were exclusively to the U.S., Europe, and the Middle East, indicating new charcoal markets of higher value were found elsewhere. 4  Definitions of forest and tree cover vary, but controlling for differences in definitions, many studies converge at or near 30 percent. 9979_Charcoal_Haiti.indd 8 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends ix in Haiti. It is likely that mar ket fluctuations—either of price or of supply and demand—have a similar effect. • Charcoal production in Haiti is now decentralized throughout the country. The addition of approximately 1,200 new feeder roads (remote, smaller roads that later join national highways) since the late 1960s and improvements to existing roads have opened up most remaining areas in the country and permitted the decentralization of charcoal production in Haiti. • The national decentralization of charcoal production in Haiti has resulted in less pressure on some traditional production areas, permitting arboreal recovery and subsequent return to charcoal production. Although charcoal production has reached most corners of the country, decentralization has relieved pressure on traditional production areas, permitting natural arboreal regeneration and a return to increased charcoal activities. • The differential supply of charcoal to Port-au- Prince by region has shifted over time as a function of the influences of decentralization, changes to transportation, and arboreal recovery in historical production zones. Presently, the following geographical regions differentially supply charcoal to Port-au-Prince. The relative contributions from the following regions represent all the charcoal consumed in Port-au-Prince during our combined sampling periods: northwest (1%); island of La Gonâve (3.4%); Artibonite (9.7%); Central Plateau (20.3%); east of Port-au-Prince (18%); the southern (Tiburon) peninsula (41%); southeast (4.2%); due south of Port- au-Prince/Kenscoff/Furcy (0.1%); and the Dominican Republic (2.3%). • Regionally, the largest charcoal supplier to Port-au-Prince is the southern peninsula, and the three top charcoal production locations together produce approximately 80 percent of the charcoal consumed in Port-au-Prince. The Tiburon peninsula registered 41 percent of the total charcoal supplied to the capital. This position as the largest regional supplier of charcoal to Port-au-Prince has remained above 30 percent since 1978. The second and third largest suppliers of charcoal to the capital are the Central Plateau (20.3%) and the area east of Port- au-Prince (18%). Together with the southern peninsula (41.0%), these top three locations supply nearly 80 percent of the charcoal consumed in the capital. • The volume of charcoal counted at the farthest, most remote enumeration stations registered an amount equivalent to one-half of the total amount entering the capital. The four enumeration stations farthest from Port-au-Prince registered an amount of charcoal equivalent to approximately half the charcoal consumed in the capital during the same time period, demonstrating that high volumes of charcoal production occur at the far reaches of the country. • The amount of charcoal emerging from areas penetrated by feeder roads established since the late 1960s is equivalent to approximately one- third of the quantity consumed in the capital during that same period of time. These ‘feeder roads’ demonstrate the high level of decentralization of charcoal production in Haiti and the production capacity of these remote locations. • Maritime charcoal transport in Haiti has decreased in significance. While historically boats carried large percentages of charcoal into Port-au-Prince, data presented here show the total volume of maritime charcoal counted at the wharfs sampled at an amount equivalent to 6.7 percent of the charcoal consumed in the capital during the same period. This is likely a result of new and improved roads and road transport. • The overall production of charcoal has increased in all geographical areas of Haiti, even as relative supply of some areas has decreased. The annual estimate for charcoal consumed in Port-au- Prince is at least five times the amount from a similar study in 1985, suggesting that overall charcoal production has increased in virtually every area sampled, including those that show a decline in the relative percent of charcoal supplied to Port-au-Prince. • Haitians are not only still meeting their woodfuel needs, they are also producing charcoal at higher volumes, not only from new locations, but also from many of the same historical production regions. This suggests that at least part of the charcoal being produced in Haiti is made with biomass resources that are renewable. • Total charcoal sales in Port-au-Prince, Haiti, are approximately US$182 million per year. 9979_Charcoal_Haiti.indd 9 2/6/19 10:38 AM x Charcoal in Haiti The average cost of a large sack of charcoal was approximately 800 Haitian Gourdes (US$12.42) from July–August 2018. Using these figures, the total estimated value of the charcoal market in Port-au- Prince is US$182 million per year. • At the national level, total charcoal sales across Haiti are an estimated US$392 million per year. Using the same values for the price of charcoal as used in the Port-au-Prince calculation, the total value of the national charcoal market is an estimated US$392 million per year. • Based on the calculations of total sales, charcoal represents approximately 5 percent of Haiti’s GDP. The economic significance of the charcoal industry in Haiti can also be placed in context by comparing it to national GDP. Based on 2017 figures, charcoal represents 4.7 percent of GDP (US$8.408 billion). • Charcoal is over six times more valuable than Haiti’s total agriculture-related export market. When charcoal’s estimated national annual value (US$392,026,140) is compared to 2016 exports of crop and livestock products, charcoal is six times more valuable than all of these exports combined (US$62,479,200). Comparing to individual export commodities, charcoal is over 15 times more valuable than the highest-valued export in 2016 (essential oils; US$25.5  million in exports), 30 times more valuable than cacao exports (US$13.2 million), over 40 times more valuable than mango exports (US$9.2 million), and a startling 650  times more valuable than coffee exports (US$611,000). The results of this study vividly underscore that in Haiti, charcoal is big business. 5 Indeed, based on initial calculations described above, charcoal is the second largest agriculture- related value chain in the country. With a total market size in Port-au-Prince of approximately US$182 million per year and a national market value of approximately US$392 million per year, charcoal is one of Haiti’s most important crops. It contributes nearly 5 percent to GDP and has large impacts for employment in rural areas. 5 A recent study in the Haitian newspaper Le Nouvelliste references so-called ‘charcoal millionaires’. The charcoal sector’s large—and likely growing—scale stands in stark contrast to the decades of apocalyptic predictions of the rapidly approaching day when a charcoal maker would fell Haiti’s last tree. This dynamic sector has continued to defy these forecasts through an intriguing combination of increased geographic reach and the evolution of more sustainable production techniques, relying on sources of renewable biomass. This study and its innovative charcoal rapid-assessment methodology highlight important new steps in growing efforts to understand the charcoal sector in Haiti. The data presented here shed additional light on charcoal—a poorly understood commodity with multifaceted and far-reaching impacts on Haiti’s economy and environment. The persistent stigmatization of charcoal as a dirty, destructive, and illegal fuel source tends to lead to calls for controls on production, or for the replacement of charcoal with other, often more expensive, sources of imported cooking fuels. With these aspects more fully considered, policy makers have the opportunity to capitalize on the economic, environmental, and energy policy opportunities offered by charcoal production in Haiti. Despite these promising aspects, any such efforts to leverage these economic, environmental and energy opportunities need to equally recognize and mitigate the negative outcomes associated with charcoal production. In particular, the health risks related to the use of charcoal for cooking. This study’s narrow focus examining volumes of charcoal transported to Port au Prince, charcoal origin, and historical trends, does not consider the health impacts of the use of charcoal. However, large bodies of ongoing research by the World Health Organization (WHO), Global Alliance for Clean Cookstoves, academic researchers, and other organizations are carefully examining the relationship between health outcomes and cooking fuels/cooking practices. Although much work is still under way to define precisely where biomass stoves and fuels become harmful, it is widely accepted that the use of charcoal for cooking is more hazardous than modern alternatives like Liquid Propane Gas (LPG), natural gas, and electricity. When developing policy related to charcoal, health aspects need to be prioritized and further researched. Addressing, mitigating, and improving all aspects of the charcoal industry requires additional research and the collation of existing research on: the characteristics of the 9979_Charcoal_Haiti.indd 10 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends xi char coal value chain and its many actors; price behaviors and trends; agronomic analyses of current wood energy production techniques by farmers; measures of sustainability and the renewability of charcoal production under different conditions and in different settings; aspects of environmental degradation and/or improvement; cleaner kilns on the production side; safe labor and working conditions on the transportation side; cleaner burning stoves and ventilation systems on the consumption side; and charcoal consumption habits and preferences. Such knowledge, combined with an increasing openness toward engaging in and improving the sector, could provide significant improvements to one of Haiti’s largest value chains. 9979_Charcoal_Haiti.indd 11 2/6/19 10:38 AM 9979_Charcoal_Haiti.indd 12 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 1 Introduction Deforestation and charcoal production in Haiti are widely misunderstood phenomena. This is perhaps nowhere more evident than the misplaced belief that charcoal production is the principal driver of deforestation. It is a well-documented historical fact that Haiti’s primordial forests of valuable hardwoods largely vanished as a result of a series of other historical events, highlighted subsequently. In the present era, it is widely believed that some 2 percent of primordial forests remain in Haiti, despite that approximately one-third of the surface of Haiti remains covered in trees that supply the necessary woody biomass for ongoing charcoal production at the national level. Misinformation about deforestation and charcoal production in Haiti spread as part of a larger, global phenomenon: the so-called ‘woodfuel crisis’ of the 1970s and 1980s. 6 This belief drove development planning during those decades and continues to drive popular but misinformed myths around woodfuel use in Haiti and beyond. 7 To illustrate, a widely cited report from 1979 suggested that existing wood supplies in Haiti were enough to meet increasing charcoal demand until around the year 2000, but would ultimately result in an environmental ‘apocalypse’. 8 The prediction that wood supplies would be exhausted by 2000 was also supported by a report on trends emerging from the early remote sensing analyses of aerial photographs spanning from 1956 to 1978, in three different locations in Haiti. 9 Forty years later, Haiti is still covered with trees, 10 and Haitians continue to produce large quantities of charcoal to supply their domestic energy needs. This report directly addresses important and unresolved questions about charcoal production. Since Haitians meet 6 Eckholm 1975, 1984. 7 Leach and Mearns 1988; Mwampamba et al. 2013; Arnold and Dewees 1997; Hansfort and Mertz 2011; Bailis et al. 2017. 8  Voltaire 1979, 21, 23. 9 Cohen 1984, v–iv. 10 Tarter 2016; Tarter et al. 2016. approximately 80 percent of their national energy needs for cooking through firewood and charcoal production, 11 a deeper understanding of current charcoal production in Haiti in the context of a historical perspective is crucial to inform policy decisions for development and government programming related to landscape management, reforestation, tree planting, agroforestry, and agricultural projects in Haiti. The methodology for data collection was drawn from research spanning nearly half a decade to allow for comparison with previous results. The following section presents a brief review of historical and contemporary knowledge to address widespread misinformation and misconceptions about tree cover, forest cover, and deforestation in Haiti, setting the context for the research questions and methodologies subsequently presented. Background Haitian Government Legislative and Executive Efforts to Address Deforestation The first Haitian government efforts to slow the cutting of trees in Haiti were enacted in 1804, the year of Haitian independence, although they were driven more by agricultural production considerations rather than strictly environmental concerns (Bellande 2010, 3). Throughout the 19th century, various Rural Codes provided strictures against the cutting of trees in and around mountain ridges, natural springs, and the banks of rivers (see Bellande 2010 and Bellande 2015 for extensive details)—a strategy reflective of an understanding of linked human and environmental influences. 11 Charcoal in Haiti is used principally for cooking in urban areas, while wood is used for cooking in rural areas, although wood is used to a much lesser extent in urban bakeries and urban drycleaners. I . Introduction and Background 9979_Charcoal_Haiti.indd 1 2/6/19 10:38 AM 2 Charcoal in Haiti Such 19th century Haitian government legislation has addressed the ownership, utilization, protection, control, restoration, marketing, and establishment of reserves for natural resources for: 12 Forests: the Law of February 3, 1926; Law of August 20, 1955; the Rural Code of May 24, 1962, Law No. VIII; Decree of March 18, 1968; Decree of November 21, 1972; Decree of November 20, 1974. Soil: the Rural Code of May 24, 1962, Law No. V; the Constitution, Article 22; and the Decree of June 16, 1977. Land use and agriculture (in terms of both state and private land): the Law of July 26, 1927; the Constitution, Article 22; the Rural Code of May 24, 1962, Laws No. IV and V; and the Law of August 11, 1975. 13 The historical consensus is that while the Haitian government recognized early on the related challenges around proper land and natural resource management—and passed corresponding legislation to address these challenges—they were ultimately not equipped to effectively and equitably enforce these laws at the national level. While the exact percentage of tree or forest cover in Haiti during many of these historical periods is not well established, early historical records provide evidence, and several new studies have provided current arboreal coverage estimates. The Original Forests of Haiti Haiti was never fully cloaked in forests, largely due to the combined influence of geographical, topographical, and meteorological deterrents—the nation is located on the leeward (i.e., dry), western side of the island of Hispaniola, in the rain shadow of several large mountain ranges that block much of the precipitation carried on the northeastern and Caribbean trade winds. As a forester completing a survey of the Haiti’s timber reserves in 1945 confirmed: The appearance of many of the inland smaller mountains and plateaus does not indicate that they ever supported 12 This list is illustrative, not necessarily comprehensive. 13 USAID 1979, 19–27. much forest growth, and many rocky hillsides probably never supported heavy timber stands, even though the valleys and ravines are known to have yielded some high- quality timber. A general survey of the country indicates that most of the stories of former vast timber resources of Haiti were probably greatly exaggerated. Even allowing for the difference in rainfall and topography between the North, West, and South coasts, it is still obvious that many of the mountainsides in the central zone and on the West coast were never covered with the heavy mixed vegetation of the Northeast and Southwest, nor with the pine forests of the higher mountain ranges of the Southeast. 14 Estimates of initial forest cover and land capacity that have considered these natural determinants range from 35 percent (Haitian Ministry of Environment) 15 to 55 percent forest cover (Haitian government’s forestry plan of 1975). 16 Stated succinctly, Haiti was probably never more than roughly halfway covered with forests. The Deforestation of Haiti Many of the primordial forests of Haiti were felled during the colonial period to establish and support the plantation model of agricultural production that would become the precursor to modern industrialized agriculture. Subsequent tree felling was authorized through contracts established between the fledgling Haitian government and foreign timber concessionaries to pay off a post-independence war indemnity and the new republic’s early leaders. 17 Cultural and religious beliefs spared some trees from felling, but both the Catholic and Protestant churches in Haiti later targeted these same species in concerted efforts to destroy them. 18 While the exact percentage of tree or forest cover in Haiti during previous periods is unknown, several new studies have provided current arboreal coverage estimates, which are briefly presented below. 14 Klein 1945, 5. 15 World Bank 1982 (pg. 17), citing a 1980 report by the Haitian Government’s Département de l’Agriculture des Ressources Naturelles et du Développement Rural DARNDR—now ‘Ministère de l’Agriculture des Ressources Naturelles et du D éveloppement Rural (MARNDR). 16  USAID 1979: 33 17 See Tarter et al. 2016 for a lengthy history of deforestation in Haiti. 18 Tarter 2015b. 9979_Charcoal_Haiti.indd 2 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 3 Current Estimates of Arboreal Coverage in Haiti A frequently cited figure that repeatedly finds its way into the development literature posits that Haiti is only 2 percent forested. One of the major issues of contention related to this estimate results from the paucity of available data concerning when the figure first surfaced; it was almost assuredly based on extrapolating trends from incomplete data into the future. It is also not clear if the 2 percent figure refers to an estimated percentage of remaining forests, and if so, whether or not it is based on the false assumption that 100 percent of Haiti was covered with forests at some point in the past—an assumption that discounts the known environmental determinants of forest cover (geographical, topographical, and meteorological) discussed above. Perhaps the 2 percent figure simply represents an estimate of overall forest land cover at the time of its formulation, although there is no solid empirical evidence that primordial (or primary) forests or forest patches in fact covered 2 percent of the land. Complicating the issue, how a ‘forest’ is defined and measured varies significantly by individual, discipline, and institution, 19 casting increasing doubt on the validity of the 2 percent figure and what it actually represents. 20 While the 2 percent estimate probably originated in reference to original, primordial forests, it discounts secondary forests, woodlands, managed woodlots, tree plantations, agroforestry systems, and the many trees found on farms. Until the recent availability and close examination of high- ­ resolution, remotely sensed (satellite) imagery, researchers estimating arboreal coverage in Haiti relied exclusively on field-site visits or qualitative assessments from aerial flyovers. Six recent research studies summarized here present data based on new satellite data, and reach conclusions contrary to conventionally accepted estimates and the belief that Haiti’s landscape is devoid of tree cover: 19 Lund 2014 lists approximately 1,600 different established definitions for ‘forest’ and ‘forest land’. 20  Tarter 2016. The Large Island of La Gônave, Offshore from Port-au-Prince 21 In 2012, geographers conducted a land-use/land- change analysis of the large island of La Gônave, offshore to the northwest of Port-au-Prince, using two high-resolution Landsat satellite images twenty years apart, from 1990 and 2010. Both photos were selected from late January—the middle of Haiti’s dry season—to control for seasonal variability in vegetation and cloud cover. For their classification, researchers considered five different land-cover types that were mutually exclusive and exhaustive of all land types in the area, including separate land category classifications for forests and shrubs. Several complementary and higher resolution satellite photographs were consulted for the accuracy assessment of the image classification, in combination with the analysts’ knowledge of the area and a two- month field visit. 22 While the overall percentage of the forest land cover decreased over 20 years by 22.7 percent, the majority of 1990 agricultural lands were converted to shrub (45.01%), forest lands (34.23%), and 56.2 percent of the eroded land area in 1990 had been revegetated in 2010. Overall, the shrub coverage in La Gônave increased by 87.4 percent from 1990 to 2010. The entire land surface of La Gônave in 2010 (excluding water and masking the < 2% cloud cover) was 40.4 percent covered with woody shrubs and 46.0 percent covered with forest. A Nationwide Estimate of Forest and Tree Cover in Haiti 23 In an analogous study, a geographer, a geologist, and a natural resource management specialist estimated forest cover for the entire country of Haiti. The authors used 2010–2011 Landsat national satellite images of Haiti at 21 White et al. 2013. 22 A stratified random sample of validation points to ground-truth their classifications was used. The team averaged 61 verification points for each of the five land-use categories (n = 301 total), locating random geospatial coordinates with GPS units. The random sample strata were 15, 52-meter elevation increments, to look for elevation-based influences (White et al. 2013, 498). The overall accuracy of their 2010 classification was 87%, with a Kappa coefficient of 0.84. The results show that the percent of land area change on La Gônave from 1990 to 2010 for agricultural land, forest/DV, shrub, and barren/eroded land classes, were −39.73%, −22.69%, +87.37%, and −7.04%, respectively (White et al. 2013, 499). 23  Churches et al. 2014. 9979_Charcoal_Haiti.indd 3 2/6/19 10:38 AM 4 Charcoal in Haiti the dry season. After a series of standard renderings and corrections, the authors reclassified their satellite images using FAO’s forest class definition, thereby creating a low and high range of percent tree cover. 24 The results of the nationwide analyses showed that in 2010–2011, trees covered 29.4 percent to 32.3 percent of Haiti’s land surface, 25 and that shrub areas accounted for 45.7 to 48.6 percent of the land surface. 26 Rather than a thinly dispersed arboreal covering, trees are aggregated in fragments and patches (Churches et al. 2014). 24 These image classifications were verified through the application of a stratified sample of 1,525 random reference points to higher resolution satellite imagery. Strata were based on the land use distributions from their initial classification (Churches et al. 2014, 209). Their ‘tree cover’ class had a users’ accuracy of 86% and a .81 Kappa statistic, and the overall classification accuracy ranged from 78% reference point counts to 83% class proportions. 25  Includes ‘water’, ‘wetlands’, ‘bare/non-vegetated’, and ‘cloud’ categories. 26 It should be noted that the Churches et al. 2014 study and the White et al. 2013 study used slightly different land-use classifications. Nevertheless, both studies use a single category that operationalizes trees in the same way, and similarly restricts smaller shrubs, and all other land uses. Said succinctly, the tree cover in both studies is similarly classified, though the non-tree cover category varies between studies. USAID Global Development Lab’s GeoCenter Land Cover Analysis of Haiti In 2016, the GeoCenter of USAID’s Global Development Lab undertook a land-cover analysis of Haiti in anticipation of a Notice of Funding Opportunity related to the U.S. Congressional earmark for reforestation efforts in Haiti. The analysis was based on different parameters applied to two different global datasets: (i)  Hansen/ University of Maryland; 27 and (ii) the World Forests 28 global dataset. Using the second dataset and associated definition of ‘forest’, USAID estimated forests span 9 percent of Haiti’s land area. When applying FAO’s standard definition of tree cover 29 for the first dataset, USAID found that 40 percent of Haiti’s land area fell under this forest cover definition. Finally, USAID applied custom parameters to the Hansen dataset, 30 suggesting 27 Hansen et al., 2013. 28 The BaseVue 2013 World Forests dataset parameters define ‘forest cover’ as trees higher than three meters in height with a closed canopy of >35%. 29  Designating 30 m 3 30 m units as tree covered if containing vegetation taller than 5 m and greater than 10% canopy cover. 30  USAID defined 9 intervals of tree cover canopy based on the Hansen 2000 tree cover dataset. Tree cover map produced by Churches et al . 2014 Bare/non-vegetated Shrub cover/herbaceous Tree cover Water Wetland No data 9979_Charcoal_Haiti.indd 4 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 5 that using the most stringent f orest cover definition (>90% canopy cover), forests cover an estimated 11  percent of Haiti’s land area, while an estimated 36 percent of Haiti’s land area would be considered forest cover when defined by a >50% canopy cover. Stated differently, when USAID applied three different forest definitions to two different global datasets, their analyses yielded present forest cover estimates that ranged from 9 to 36 percent of Haiti’s land surface. A Regional Study of the Greater Antilles 31 In a broader study of land use changes within the Greater Antilles 2001–2010, researchers observed that 26 Haitian municipalities 32 underwent significant changes to woody vegetation 33 (8 decreased and 18 increased); 36 municipalities underwent significant changes in agriculture/ herbaceous cover 34 (25 decreased and 11 increased); and 48 municipalities experienced significant changes in mixed- woody/plantations 35 (9 decreased and 39 increased). Across municipalities of significant change, there was an 8 percent loss of woody vegetation, a 114 percent loss of agriculture/ herbaceous, and a 133 percent increase in mixed-woody/ plantation. Conversely, in overall land percentages for the entire country, woody vegetation increased from 1 percent, agriculture decreased from 4 percent, and mixed-woody/ ­ plantations increased from 4 percent. Municipalities that experienced losses or gains were widely geographically distributed. 36 An Assessment of Aboveground Biomass in Haiti 37 A dynamic landscape model 38 was used to simulate changes in land cover if woodfuel demand in Haiti continues unabated. The modeling found that current demand might contribute to moderate levels of ecological degradation, but that “the situation is not as 31 Álvarez-Berríos et al. 2013. 32 Sections communales. 33 Woody vegetation was defined as trees and shrubs with >80% cover. 34 ‘Agriculture/herbaceous vegetation was annual crops, grasslands, and pastures with >80% cover’ (Álvarez-Berríos et al. 2013, 86). 35  “Mixed-woody/plantations was woody vegetation with a 20% to 80% cover, including agriculture/herbaceous vegetation or bare soil as background, as well as all forms of plantations and perennial agriculture” (Álvarez-Berríos et al. 2013, 86). 36  This points to the wide decentralization of charcoal production in Haiti. 37 Ghilardi et al. 2018. 38 Modeling Fuelwood Sustainability Scenarios (MoFuSS). severe as is typically portrayed.” Under a ‘business-as- usual’ scenario, “the simulated regenerative capacity of woody biomass is insufficient to meet Haiti’s increasing demand for wood energy and, as a result, between 2017 and 2027 stocks of aboveground (woody) biomass could decline by 4 percent (± 1%), equivalent to an annual loss of 302 kilos/ton (± 29%) of wood.” The authors acknowledge that “the input parameters utilized in this preliminary exploration carry large uncertainties,” given limits to primary data and model input assumptions. An Assessment of Biodiversity Linked to Remaining ‘Primary Forest’ in Haiti 39 In late 2018, a report was released indicating that ‘primary forest’ in Haiti has declined from an amount equal to 4.40 percent of the total land area in 1988, to an amount representing 0.32 percent of the land area in 2016. 40 The research uses a definition of ‘primary forest’ that assumes a “stringent 70 percent threshold [for tree canopy] and then eliminates cases of major regrowth (secondary growth) by following 30-m pixels back in time to make sure to always represent forested areas.” 41 Here the definition of ‘primary forest’ is restricted to those areas with a tree canopy cover equal to or greater than 70 percent, unchanged over the last 33 years (excluding areas of significant arboreal regrowth), and only on plots of land larger than 0.5 hectares. Although certain aspects of the methodology of this study need to be examined further, 42 the study identifies important primary forest areas that represent some of the hotspots of remaining and highly sensitive animal (and plant) biodiversity. 39 Hedges, Blair, S., Warren B. Cohen, Joel Timyan, and Zhiqiang Yang. 2018. Haiti’s biodiversity threatened by nearly complete loss of primary forest. Proceedings of the National Academies of the Sciences. Online/pre-print version accessed Nov. 1. https://doi.org/10.1073/pnas.1809753115. 40  Hedges et al. 2018, 1. 41 Ibid., brackets authors, parentheses original. 42 Two methodological questions should be raised: (1) The study’s generalization of results from the few areas of remaining ‘primary forests’ to the national level is based on the assumption that at one point in the past the entire land surface of Haiti was covered with such forests, whereas, as cited, the combined influences of geography, topography, and meteorology suggests that the natural capacity for forest cover in Haiti ranges from 35–55%; and (2) do all nine categories from Holdridge’s life zones—an ecological classificatory system developed in, and for, Haiti, but now applied as a globally recognized ­standard (Holdridge 1947, 1967)—meet this new definition of ‘primary forest’? For example, ‘Subtropical Dry Forest’ (19% of Haiti’s land) and ‘Subtropical Thorn Woodland’ (less than 1%) may not meet the 70% canopy threshold. Other serious concerns about the methodology remain but are beyond the scope of this report. 9979_Charcoal_Haiti.indd 5 2/6/19 10:38 AM 6 Charcoal in Haiti These findings, highlighting differences in how forests are conceptualized and measured, and collectively presenting overwhelming data contrary to popular depictions of charcoal and deforestation narratives about Haiti, induce the research questions posed in this report. The Historical Production of Charcoal in Haiti Charcoal production in Haiti commenced around the 1920s. 43, 44 Prior to that time, given the large rural population distribution, most Haitians met their domestic energy needs through firewood, which could be procured in rural areas, a task mostly relegated to children and women. 45 The rise of charcoal production in Haiti in the 1920s correlates with increasing urbanization that occurred not only in Haiti, and particularly in Port-au-Prince, but also as regional and global phenomena. Increased charcoal production was also 43 Tarter 2015a; Tarter et al. 2016. 44 Haitian agronomist, researcher, and author Alex Bellande has recently uncovered the earliest known reference to charcoal production in Haiti: a report of charcoal entering the capital by railway in 1909 from the Haitian newspaper Le Nouvelliste, listing a total of approximately 375 metric tons for six months during 1909 (Bellande, personal communication, 10/16/2018). 45  See Annex 6 for a discussion on the gendered division of labor and the differential gender effects on health related to the product, transport, marketing, and consumption of charcoal in Haiti. facilitated by the emphasis in the 1920s 46 to rehabilitate existing and build new transportation infrastructure, including roads, railways, and maritime wharfs. Increased urbanization and population densities called for more agricultural clearing and charcoal production. Research from other analogous locations throughout the world has debunked many myths surrounding woodfuels, especially the conclusion that charcoal production is the primary driver of deforestation. 47 Although there has been lengthy debate over whether charcoal production or agricultural clearing is the principal driver of deforestation in Haiti, 48 the widely cited historical figure suggesting that only 2 percent of Haiti’s primordial forests remain is disassociated from either argument. Charcoal in Haiti is largely produced from trees, not forests, trees found in woodlands, woodlots, agroforestry systems, and on farms. The Use of Non-Charcoal Wood in Haiti The present study examines charcoal use precisely because it is the principle use of wood in Haiti, primarily for cooking. 46 1919–1934. 47 Leach and Mearns 1988; Mwampamba et al. 2013; Arnold and Dewees 1997; Hansfort and Mertz 2011; Bailis et al. 2017. 48  Stevenson 1989. Charcoal sold by the bucket in a local market Traditional charcoal stoves 9979_Charcoal_Haiti.indd 6 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 7 Although rural dw ellers use wood or a mix of charcoal and wood for cooking, urban and peri-urban dwellers use almost exclusively charcoal. The use of charcoal in urban areas relates to the economics of moving wood versus charcoal (charcoal is lighter, more compact) and it burns less smoky than wood, making it more comfortable for densely populated areas. In Haiti, food cooked with charcoal is also a cultural preference, related to custom and stated preferences for the flavor of food cooked with charcoal, and for this reason, charcoal is the preferred fuel for most Haitians. The choice to use fuelwood for cooking in rural areas is related to cost—it is generally free to cut wood in rural areas, whereas charcoal carries a cost. Beyond uses for cooking, the literature occasionally references the ongoing use of wood in urban businesses such as dry cleaners and bakeries; however, these are few (around an estimated 100 across the country) and result in a negligible amount of wood compared to the demands from charcoal production. Other uses of wood in Haiti include furniture, doors, beds, coffins, and construction scaffolding. However, the demand for these products is limited (scaffolding polls are reused between jobs, and furniture and other carpentry works are imported). The use of and demand for charcoal in Haiti dwarfs that of wood, and for this reason this research focuses exclusively on the movement of charcoal toward the principle urban market of Port-au-Prince. The Historical Supply of Charcoal to Port-au-Prince The first historical areas of large-scale charcoal production in Haiti occurred east of the capital city of Port-au-Prince, and later shifted offshore to the nearby island of La Gonâve, as areas in closer proximity to the capital started producing charcoal. Then charcoal production moved into the northwest peninsula, and later swung to the more remote southern peninsula, and to a lesser extent into Haiti’s Central Plateau area. 49 Charcoal production designated for the capital, the largest consumer to date, commenced in areas close and accessible to Port-au- Prince before shifting toward geographically remote areas. By around 1980, estimates suggested that only 5 percent of charcoal consumed in Port-au-Prince came from the area east 49 Smucker 1981; Conway 1979; Voltaire 1979. of the capital, where large-scale production had historically commenced. 50 Likewise, only 5 percent of charcoal consumed in Port-au-Prince came from central Haiti; 10 percent came from the large offshore island of La Gônave; 50 percent originated from the northwestern peninsula; and 30 percent from the southern peninsula. Voltaire predicted charcoal production would eventually shift from the rapidly depleting areas of La Gônave and the Northwest to the more wooded areas of the Central Plateau and Grand Anse. 51 Table 1 displays these and subsequent estimations of varied regional contributions to the charcoal consumed in the capital city of Port-au-Prince. The first comprehensive, robust survey on the production and consumption of charcoal in Haiti 52 was administered some six years after Voltaire’s report and predictions. According to the study, by 1985 the northwestern area of Haiti was supplying 34.2 percent of charcoal consumed in the capital; the large off-shore island of La Gônave supplied 7 percent; the Central Plateau area supplied 12.7 percent; the southeast supplied 10.3 percent; an area in the center of the southern peninsula provided 28.1 percent; and an area toward the end of the southern peninsula supplied 7.7 percent (see Table 1). Contemporary Charcoal Production in Port-au-Prince According to the Haitian government’s National Institute for Statistics and Information (IHSI), Port-au-Prince has an estimated population of 2,618,894, 53 which is over seven times the population of the second largest city of Gonaïves (356,324) and nearly ten times the population of the third largest city of Cap-Haïtien (274,404). As such, the capital remains the largest consumer of charcoal in the country, and the logical location from which to sample and extrapolate to produce an estimate of annual charcoal consumption for the capital and the urban population of the country. 50 Voltaire 1979. 51 Ibid. 52 Grosenick and McGowan 1986. 53 IHSI 2015. 9979_Charcoal_Haiti.indd 7 2/6/19 10:38 AM 8 Charcoal in Haiti TABLE 1: Regional Contributions to Annual Charcoal Consumption in Port-au-Prince Year Main production areas Est. % contribution Source 1930s East of Port-au-Prince; Island of La Gonave; Northwestern peninsula ~100% Smucker 1981; Conway 1979 1979 East of Port-au-Prince ~5% Voltaire 1979 Island of La Gonave ~10% Northwest ~50% Central Plateau ~5% Southern peninsula ~30% 1985 East of Port-au-Prince ~0% Grosenick and McGowan 1986  Island of La Gonave ~7% Northwest ~34.2% Cen ter (Croix-des-Bouquets, Hinche, St . Marc) ~12.7% Southern peninsula  (Center of southern peninsula ~28.1%; tip of southern peninsula/ Grand Anse ~7.7%) ~35.8% Southeas t (Jacmel  and Thiotte) ~10.3% 1990 Gr and Anse ~13 ESMAP 1991 Island of La Gonâve ~5 Northwest ~21 Central (Central Plateau and east of Port-au-Prince) ~13 South ~2 Southeast ~3 West ~26 Artibonite ~13 North ~4 2005–2007 The southern peninsula (Grand Anse, Belle-Anse,  Aquin, the south coas t); the Northwest; and the Central Plateau (Maïssade,  Thomonde, Thomassique, Pignon, Cerca-Cavajal, Hinche, Mirebalais, Boucan Carr., Saut d’Eau, and Lascahobas) ~100% ESMAP 2007 9979_Charcoal_Haiti.indd 8 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 9 passing charcoal vehicle or incoming charcoal vessel, they also recorded the date and time of each observation, permitting hourly and daily analyses. Methodological Similarities and Differences from Previous Studies The methodology presented below is based largely on an earlier study of charcoal consumption in Port-au-Prince, produced by researchers from the University of Maine, working in conjunction with the decade-long Agroforestry Outreach Project (AOP). 55 This methodological approach was loosely replicated in various forms in several subsequent studies (see Table 2). By modeling the methodology on aspects of these earlier studies, it is possible to provide both cross-sectional (2017) and longitudinal (diachronic changes over almost four decades) estimates of nationwide charcoal production and consumption trends in Haiti. Other aspects of the methodology are independent of these earlier studies, in order to address new research questions pertinent to current production and consumption patterns in Haiti. Timeframe of the Fieldwork Scoping 56 During a preliminary scoping phase (2016–2017) the research team conducted visits and undertook contextual interviews to identify the important charcoal depot (wholesaler) locations in Port-au-Prince. The team examined storage points accessible by roads and sea, and in warehouses and wharfs. The interviews and observations allowed the team to develop a typology of charcoal transportation vehicles (trucks and boats). In addition, the research team visited locations throughout the country to identify the places along roads and at key intersections where enumeration stations would be placed during the survey phase. 55 Grosenick and McGowan 1986. 56 See Annex 1. Research Questions As a general, historical, and widely observed trend, charcoal in Haiti is produced rurally and transported for consumption to urban areas, in contrast to wood, which is largely consumed in rural areas. 54 Since the majority of charcoal produced in Haiti is geared toward consumption in the capital city of Port-au-Prince, the following, overarching questions guided this research design: (R1) How much charcoal is consumed annually in the capital city of Port-au-Prince? (R2) Which geographical regions produce the charcoal consumed in the capital? (R3) How do these production areas variably supply charcoal to the capital? (R4) In what ways have these trends changed over the last 40 years? (R5) What percentage of charcoal consumed in the capital is originating from the bordering Dominican Republic? Methodology Principal Method The principal method of this study mirrored previous estimation methods—to count charcoal transport vehicles (trucks) and vessels (boats) entering the capital city of Port-au-Prince and/or passing through important roads or crossroads en route to the capital. Teams of three enumerators were positioned at roadside stations and maritime wharfs alongside 23 carefully identified locations throughout the country. The enumerators took shifts manning the stations for 24 hours a day, over three different sampling periods. Enumerators not only tallied every single 54 Exceptions include wood used in bakeries and drycleaners, although these are negligible amounts in comparison to the amount of wood that is transformed into charcoal. II . Research Questions and Methodology 9979_Charcoal_Haiti.indd 9 2/6/19 10:38 AM 10 Charcoal in Haiti TABLE 2: Differences between Previous Charcoal Study Approaches Study Principal method Sampling period(s) Locations Time period Grosenick and McGowan 1986 Stopping vehicles at checkpoints and counting charcoal bags Three one-week periods, although due to political unrest, only two weeks were utilized: • July 1985; and • May 1986 Police checkpoints on two major highways that controlled all traffic into the capital from the North, Southeast, and Southwest . 24 hours High traffic periods on lesser roads and on important maritime wharfs: • Avant Poste de Police (police s tation) at Cazeau (north and southeast); • Avant Poste de Police in Brochette/ Carr efour (southwest); • Intersection of Rue Huc and Delmas ( major entry location for charcoal on donkeys); • Djoumbala on Rue Freres (major entry location for charcoal on donkeys); • Cité Soleil wharf (maritime arri vals); • Martissant/La Rochelle wharf ( maritime arrivals); and • Jérémie wharf (maritime arrivals). High traffic periods E SMAP 1991 Counts of the number of charcoal shipments entering Port-au- Prince One week in February 1990; Averaged with counts from the first trimester of 1989 • Gr • Wharf • Wharf Cite Soleil; • Wharf • Wharf • M • Croix-des-Bouquets; and • Cr. Not clear ESMAP 2007 Assessing arrivals: number of bags, transportation means used, source of charcoal One week in the capital; Buttressed with supplementary surveys conducted over an undetermined amount of time • Carrefour Shada (northern en trance); • Rond Point de la Croix des Bouque ts Police Station (Central Plateau and border zone); • Gressier (southern entrance); • Cité Soleil wharf; • J • La Rochelle wharf; • Mariani; and • Croix-des-Bouquets market. Partial daytime co verage (areas listed to the left) Partial nighttime coverage (Carrefour Shada, Croix des Bouquets, and Gressier, only) While the studies outlined in Table 2 vary in sampling strategy, duration, and robustness, they represent the best and only data available. This study takes Earl’s (1976) and Voltaire’s estimates (1979) as baselines, although they do not use methodologies similar to those presented in Table 2. All comparisons made in the analyses and concluding chapter of this report recognize these limits and control a series of variables to increase comparability between and across these early reports. Sampling Spatial Sampling: Core and Periphery Figure 1 identifies the location of 23 different charcoal ­vehicle/vessel enumeration stations located on major roads, at important intersections, and at maritime wharfs across most of Haiti. Enumeration stations farthest from the capital or at 9979_Charcoal_Haiti.indd 10 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 11 the intersection of two major routes provide ‘periphery’ counts that further disaggregate the location of charcoal production, while those located around the capital are referred to as ‘core’ stations. More precisely: • The core accounts for the six final enumeration stations that all trucks and boats must pass through to enter the metropolitan area of Port-au-Prince. 58 The core data do not suffer from double-counting that may 57 See the entire list of the enumeration stations in Table 5. 58 The metropolitan population of Port-au-Prince encompasses all areas included in our ‘core’ (the communes of Port-au-Prince, Delmas, Cite Soleil, Tabarre, Carrefour, and Pétion-Ville). occur in the periphery, whereby different enumeration stations count the same truck on the same day en route into the capital. Instead, core counts represent the final amounts of charcoal entering the metropolitan area of Port-au-Prince and are used to make estimations of the consumption level of charcoal in the capital, and later to make extrapolations to the entire country based on a tons-to-population ratio and associated economic national estimates. • The periphery stations are all other enumeration stations outside of the Port-au-Prince area. Many of the periphery stations are located at intersections, and thus control for charcoal flows headed in the direction toward the capital from more than one location. FIGURE 1: Geographical Dispersion of 23 Enumeration Sites Note: The red points represent the six core stations (Croix-des-Bouquets, Kafou Thomazeau/Mòn Kabrit, Kenscoff, Mariani, Titanyen, Wharf Jérémie). The green points represent the periphery points. 57 Hinche 1 and 2 overlap due to their close proximity. 9979_Charcoal_Haiti.indd 11 2/6/19 10:38 AM 12 Charcoal in Haiti All subsequent results are noted as core or periphery results. Final counts presented in the subsequent analysis are based on the six ‘core’ stations in order to avoid double-counting of trucks. The periphery stations allow for the determination of approximate geographical origins of production for charcoal supplied to Port-au-Prince. Temporal Sampling: The Annual Calendar for Charcoal Production 59 Previous research demonstrated seasonal fluctuations in annual charcoal production in Haiti. The scoping phase of this research found similar results—interview respondents confirmed that there are low and high seasons of charcoal production. High seasons of charcoal production correspond to periods when households anticipate higher than normal financial expenditures. According to the interviews conducted during this research and more generally on Haitian households’ expenditures, school-associated costs are one of the most frequently reported costs for which the rural population must make provision. The two periods during which primary school enrollment fees must be paid include the first semester of the school year (August/September), and the second semester (December/January). 60, 61 Trees are also harvested for charcoal during increased periods of drought, when crop failure is widespread. In the case of crop failure, charcoal production intensifies when it becomes apparent that seasonal rains are insufficient to produce a marketable crop, although droughts are often experienced differentially by region. Charcoal production lulls during the rainy seasons and during periods when farmers and their families are occupied with the preparation of fields, planting, weeding, harvesting, processing, and marketing of agricultural food crops. Despite these difficulties in determining exact seasonal fluctuations in charcoal production on an annual basis, three intervals of time are generally considered as peak charcoal 59 ‘High’ and ‘peak’ are used interchangeably as synonyms across the report. 60 This also corresponds to the pre-Christmas expenditure period. 61 The 2017 school year began on September 4, as decreed by the Ministry of Education’s (MENFP) academic calendar. However, it is common in rural Haiti that children do not return to school on the start date. Many families instead delay until they are able to gather the money to pay for school fees, books, and uniforms. (MENFP Ministère de l’Education Nationale et de la Formation Professionnelle.) production periods across rural Haiti: (1) August/September; (2) December/January; and (3) May/June. It is also generally believed that the first range is the highest production period, but regional differences should not be discounted. The remaining intervals of time throughout the year may be considered as lower production periods. These seasonal differences provide for six months of peak charcoal production, and six months of low production in a typical year. Timeline of Data Collection 62 The research protocol was initially designed for data collection during two sampling periods of one week each: (i) One week in a high-production season, August 2017; and (ii) One week in a low-production season, October 2017. The aim of this design was to collect data from weekly periods that could be applied to associated low and high charcoal production seasons throughout the year, producing an annual estimate that considers seasonal production fluctuations. However, due to logistical issues 63 during the first sampling period, a third sampling period of 72 hours was added in December 2017—another peak period—to make up for hours missed during the peak period of August 2017 (see Table 3). 64 Reconstruction of a Peak and a Low Period The sample data for the peak charcoal production season were divided between two periods (1 and 3, see Table 3). In order to construct a complete peak week for comparison to the complete week from low production season, sampling periods weeks 1 and 3 were combined. When there were observations from both peak periods, the means of tons and the mean number of trucks were used. When there were data from only one period, values from the observed period were imputed into the new peak season week. This methodology is described in further detail in Section III Results. 62 See Annex 1 for more details. 63 An error was made by the firm contracted to collect the data. All enumerators were mistakenly sent home 17 hours early during the first sampling period. The firm mitigated the mistake by resampling a few days during the following peak season (sampling period 3 in Table 3). 64  See Annex 2 for further details about the logistical organization on the field. 9979_Charcoal_Haiti.indd 12 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 13 TABLE 3: Description of the Three Sampling Periods Day Su M T W Th F Sa Total hours Period 1 (8/26/17, 7 pm– 9/1/17, 7 pm) 24 hrs 24 hrs 24 hrs 24 hrs 24 hrs 19 hrs (12 am–7 pm) *5 hrs (7 pm–12 am) 144 Period 2 (10/23/17, 7 am–10/30/17, 7 am) 24 hrs *24 hrs 24 hrs 24 hrs 24 hrs 24 hrs 24 hrs 168 Period 3 (12/15/17, 7 am–12/18/17, 7 am) 24 hrs 7 hrs (12 am 67 –7 am) No data No data No data *17 hrs (7 am–12 am) 24 hrs 72 Total hours 72 55 48 48 48 60 53 384 * = Sampling period start time Data Standardization Standardizing Vehicle Classifications Roadside enumerations were based on a tripartite typology of transportation vehicles (small, medium, or large trucks); enumerators counted and classified different-sized charcoal- carrying vehicles. During Phase I, a series of ‘carrying capacity ranges’ were established for each of these three truck categories, based on trucks at full capacity with large bags. 65 The estimated average weight of a large charcoal bag was initially established by repeatedly weighing large bags (n = 72 bags) at four different locations (two wharfs and two charcoal depots) throughout Port-au-Prince, with an average of 41.93 kilograms per large charcoal bag. 66 Previous researchers also varied as to the weight they ascribed to a large bag of charcoal. Since the most relevant early studies used a 30 kg 65 A.M. denotes midnight; P.M denotes noon. 66 The range of the weight of the 72 large bags was wide, from 20 kg large bags to 65 kg large bags. Since weights varied so much, the mode of all the bags was also not a viable option. In Haiti charcoal bags vary in weight tremendously by the type of wood that produced the charcoal, the moisture content of the charcoal, the sizes of the charcoal pieces, the occasional addition of rocks, and slight variations in the sizes of bags themselves. average for a large charcoal bag, 67 and other important studies reported counts of different-sized bags but also reported weight totals in tons, 68 for the sake of comparability across time, this study uses a 30 kg estimate for a large sack of charcoal. Given that the average of the 72 bags of charcoal weighed was 41.93 kgs, the 30 kg number used in the present calculations is a conservative estimate. Table 4 outlines the carrying capacities and associated ranges of metric tons ascribed to enumerator counts of small, medium, and large trucks. The subsequent data are analyzed as: (1) counts of the different categories in the tripartite vehicle typology; and (2) total weight based on the average (midrange) of the metric ton ranges for each type of truck, as displayed in Table 4. Distinction among Vessels and Vehicles The three wharfs that were sampled also function as de facto charcoal depots, with some bags of charcoal stacked high and sitting for indeterminate periods of time. Those enumerators stationed at maritime wharfs were instructed, trained, and 67 Earl 1976; Voltaire 1979; Smucker 1981; Grosenick and McGowan 1986. 68 Stevenson 1989; ESMAP 1991; ESMAP 2007. TABLE 4: Conversion of Categorical Counts to Metric Ton Averages Truck Size Truck carrying capacity (full capacity with large sacks) Truck weight range (metric tons) Low-end range Midrange High-end range Small truck 25–50 0.75 1.125 1.5 M edium truck 100–200 3 4 .5 6 L arge truck 300–400 9 10 .5 12 9979_Charcoal_Haiti.indd 13 2/6/19 10:38 AM 14 Charcoal in Haiti equipped to count not only charcoal vessels entering the wharfs, but also to record trucks departing the wharfs, which had come to fill their beds with charcoal. In the final analysis, wharf counts were based on those vehicles departing wharfs with beds full of charcoal, rather than on incoming boats. This ensured that charcoal en route to the capital was counted during the respective sampling periods, rather than counting charcoal that may be stored at wharfs for undetermined durations of time, that could have extended the amounts headed for the capital during sampling periods. Enumerators also noted the origins of boats entering the wharfs, and the quantities of charcoal such boats carried, in order to estimate percentages that reflected different origins of charcoal. This effort was similar to those enumerators station at intersections, noting the specific direction each vehicle originated from. However, all respondents at Wharf Marigot noted that charcoal arrived from the southeast, all respondents at Wharf Jérémie (in Port-au-Prince) reported maritime arrivals came from Grand Anse, and all respondents at Wharf Arcahaie reported charcoal incoming from the island of La Gonâve. 69 69 This is only unusual in the case of Wharf Arcahaie, where the historical literature indicated that charcoal offloaded at the wharf is separated by that coming from the island of La Gonâve and that originating in the northwest. It appears that the advent of better transportation has changed this historical trend. Large truck Medium truck Small truck Wharf Archahaie from La Gonâve 9979_Charcoal_Haiti.indd 14 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 15 73 These figures represent raw data, before any standardization of the dataset occurred. Descriptive Statistics Enumerators made a total of 10,436 70 observations of small, medium, or large trucks during the 384 hours of data collection across three different sampling periods, in the 23 enumeration stations, (see Table 5). These observations consisted of 3,790 trucks entering Port-au-Prince (core) and 6646 trucks heading toward the capital (periphery). 71 The data collection and classification were potentially susceptible to different sources of errors. In order to mitigate any potential errors, several measures were adopted (see Annex 3 for more details). Two locations had no observations: Belladère and Anse-à-Pitres. The Belladère enumeration station was at an intersection controlling two routes of entry near the border with the Dominican Republic. Anse-à-Pitres was controlling for charcoal entering Haiti at the southern-most land point along the Haiti- Dominican Republic border. In subsequent figures, the two stations with no observations will not be displayed. 72 In addition, some other enumeration stations were outliers, failing to generate substantial observations. This was the case for the Cerca-la-Source enumeration station, where only 31 observations were made at an intersection controlling for two routes of entry along the central border with the Dominican Republic. There were also only six total observations from the Kenscoff enumeration station in Pétion-Ville, which controlled for charcoal entering the capital from the large and wooded mountain ranges due south that tower above Port-au-Prince. 70 These figures represent raw data, before any standardization of the dataset occurred. 71  In the periphery stations some observations are repeats (i.e., the same truck recorded at more than one enumeration station, en route to the capital). 72  Enumerators were still posted at these stations for the full 384 hours of observation across all three sampling periods. III . Results TABLE 5: Total Number of Observations by Location and Size 73 Enumeration station Truck size Grand totalL M S Anse-à-Pitres 0 0 0 0 Belladère 0 0 0 0 Carrefour Dufort 631 391 4131,435 Carrefour Mariani 530 338 315 1,183 Carrefour Moussignac 586 154 442 1,182 Les Cayes 347163374 884 Cerca-La-Source 14 13 4 31 Croix-des-Bouquets 293 161179 633 Hinche 1 87 28 69 184 Hinche 2 112 36 91 239 Jérémie 120 28 45 193 Mòn Kabrit/Thomazeau 319 120 669 1,108 Kenscoff 1 0 5 6 Malpassee 102 38 50 190 Miragoâne 515206 455 1,176 Mirbalais_1 235 45 131 411 Mirbalais_2 54 26 31 111 Pont-Sondé 112 60 193 365 Thiotte 87 8 5 100 Titanyen 163 144 246 553 Wharf Archahaie 44 28 8 80 Wharf Jérémie 5 22 280 307 Wharf Marigot 2 22 41 65 Grand total 4,359 2,031 4,046 10,436 9979_Charcoal_Haiti.indd 15 2/6/19 10:38 AM 16 Charcoal in Haiti Differences between the Three Sampling Periods Comparison of All Three Sampling Periods Based on One Shared Day Sunday is the only complete (24-hour) day when data were collected across all three sampling periods. Figure 2 shows a comparison of the total metric tons circulating in the periphery, and passing into the core on Sunday, for all three sampling periods combined. From Figure 2, sampling period 3 appears to have the highest flows of metric tons into the core; this trend is repeated in the periphery but does not account for double-counting along the same route, or for charcoal that was offloaded prior to entering Port-au-Prince. Establishing seasonal differences in charcoal production based on only one day can create certain caveats in the data interpretation, including statistical overgeneralization of results considering daily variations within a weekly period. To mitigate this challenge, data are compared in several ways to highlight the real variation between sampling periods. Figure 2 compares the means of tons per hour passing through enumeration stations during the 24 overlapping hours between FIGURE 2: Total Charcoal Production by Periphery and Core for All Periods (in metric tons, Sunday—24 hours) Periphery Core 221 0 1,000 2,000 Metric tons 3,000 133 the three sample periods 1, 2, and 3 in the periphery and in the core enumeration points. It appears that more charcoal is circulating in the periphery than in the core. Generally speaking, this is true, as urban areas outside of Port-au-Prince consume charcoal as well, and not all charcoal passing through periphery stations also passed into the core. However, the totals from the periphery were also subject to double-counting of the same trucks by different enumeration stations along the same road. Comparison of All Sampling Periods Based on Total Shared Hours This section compares the 47-hour period shared across all three sampling periods, as shown in Table 6. TABLE 6: Description of the 47 Hours Overlap across All Sample Periods Weekday Time period Number of hours Sunday All  24 Monday (12 am to 6 am) 7 Friday (7 am to 6 pm) 12 Saturday (7 pm to 11 pm) 4 Total 47 9979_Charcoal_Haiti.indd 16 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 17 When these temporal ov erlaps between all sampling periods are added together, we can compare metric ton flows of charcoal into Port-au-Prince across all three sampling periods, based on a combined 47 shared hours (see Figure 3). Figure 3 shows that on average sampling periods 1 and 3 both register more overall metric tons of charcoal entering the core of Port-au-Prince than sampling period 2. Given that daily hours of sunlight vary by season, Figure 4 shows only daylight hours in order to control for this seasonal difference. It displays the average tons per hour across the three sampling periods’ daylight overlapping hours (7 am–7 pm). Figure 4 shows that that quantity of charcoal transported during daylight hours (7 am–7 pm) is higher per hour in December than in August. In other words, the average total tons of charcoal transported per daylight hour is higher in the wintertime. This is logical given that, during the month of August there are more hours of daylight, permitting trucks transporting charcoal to travel over a greater number of daylight hours than in the winter. Reconstruction of the Peak Week Despite small observable variations, the difference between sampling period 1 and 2 is not statically significant. Figure 2, Figure 3, and Figure 4 demonstrate what the qualitative interviews suggested: Period samples 1 and 3 (peak season) have on average more metric tons entering Port-au-Prince than sample period 2 (low season). As described in Section II, due to 17 hours of missing data during sample period 1, sample periods 1 and 3 are merged to recreate a full week of peak. When there are observations from both periods, tons of charcoal and number of trucks are averaged across the two periods. Where data from only one period was observed, values from the observed period are imputed to the unique peak week construction. Figure 5 reproduces Figure 3, and Figure 6 reproduces Figure 4, but in each case with the newly reconstructed peak week. These statistical robustness checks confirm the hypothesis of low and high charcoal production season. Figures 5 and 6 demonstrate that the average tons per hour for the peak weeks are significantly higher than in the low weeks. Robustness Check on Number of Trucks Figure 7 compares the mean number of trucks of each size per hour for the low week and for the reconstructed peak FIGURE 3: Average Tons per Hour across the Three Sampling Periods on the 47-hour Overlap 30 40 50 Tons of charcoal per hour 60 2 Sampling period 1 3 9979_Charcoal_Haiti.indd 17 2/6/19 10:38 AM [... middle sections omitted for long document ...] 60 Charcoal in Haiti A major component of the training was to describe the survey methodology. Enumerators learned about the different types of transportation mediums (including the categories of trucks and boats) that they would be counting. The training used photos to provide examples of the different kinds of trucks the enumerators should look for and how to classify them using the tripartite typology. Enumerators also learned about the different types of enumeration stations (e.g., intersections, principal roads, and wharfs) to be used in the study. At intersections, enumerators were instructed to record the type of vehicle, the direction it was traveling from, and the time of the observation. On the principal roads and intersections, the enumerators had only to record the category of the truck and time of passage. At wharfs, enumerators counted both the number of charcoal bags coming into the port off boats and recorded where the boats originated. Enumerators also recorded the type of truck and direction of travel for trucks leaving wharfs full of charcoal. Each enumeration station was assigned a team of three enumerators. All night shifts were required to have two enumerators present as a security measure. Wharf teams only did day shifts because all the wharfs close at night. The training emphasized the idea that this was a team activity rather than an individual activity. In order for stations to be successful, each person had to participate and follow the cooperatively developed schedules to meet the data collection requirements. Given the large geographic scope of the study, the research team wanted to have a system in place to ensure accountability (e.g., people are at their stations at the required times), reliability (enumerators are properly classifying trucks), and data backup (classifications matched with time stamps based on locations). Logic for, and Creation of, Enumerator ‘WhatsApp’ Groups Given the large geographic scope of the study, the research team wanted to have a system in place to ensure accountability (e.g., people are at their stations at the required times), reliability (enumerators are properly classifying trucks), and data backup (classifications matched with time stamps based on locations). Thus, each group was also asked to select a group name for pre-created ‘WhatsApp’ 159 groups. All enumerators were added to two different WhatsApp groups, which served different functions: Group 1 The first group was used to track work shifts. Whenever someone started their shift, the team had to send a text that confirmed who was coming on-shift and who was going off-shift. The required message format was as follows: Kenson + Richard + (James)—Pa Gade Solèy/JER James + (Richard + Kenson)—Pa Gade Solèy/JER In the above example, the text provides the name of the three team members. The person(s) whose name(s) appears in parenthesis are off-duty, while the other person(s) are on-duty. The team name appears after their names (Pa Gade Soley) followed by the specific enumeration station (JER, or Jérémie). WhatsApp adds time stamps automatically. Group 2 Enumerators utilized the second WhatsApp group to post pictures of every truck that passed during daylight hours. No photographs were taken during nighttime hours due to security concerns (blinding drivers with flash, identifying enumerator locations, etc.). Sharing photographs of trucks with all enumerators alerted the next station down the line that a truck should be coming and facilitated tracking trucks from one station to the next. This group chat was also used to disseminate any pertinent information that should be shared with the larger team (e.g., weather conditions, security concerns, etc.). The format of each text would include the picture, location of origin, size, followed by the specific enumerator team name, and station (e.g., Large/RD  102/Pa Gade Solèy/JER). WhatsApp adds time stamps automatically. 159 WhatsApp is the most popular and widely used, free, chat application for phones in the world. It allows users to share messages and images, and time stamps both. 9979_Charcoal_Haiti.indd 60 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 61 The Wha tsApp groups had the added benefit of promoting coordination and communication among enumerations teams, fostering a spirit of healthy competition, and boosting overall morale. Field Visits The Research Team Leader (fifth author) made a series of surprise enumeration site visits, taking a team photograph with people present at the station and posting it to the WhatsApp group. This provided motivation for teams to stay engaged and at their station in case they would receive a surprise visit. It also turned into a fun, team-building activity. People did not know when or where the Research Team Leader would show up and would try to track or anticipate his sporadic movements. Each enumeration station received at least one surprise site visit during the course of the study, although this was not achieved in one sampling period alone. None of the field visits were announced to the enumerator teams ahead of time to promote accountability/engagement. Given the geographic scope of the study, JP/HRO provided two interns who also conducted more frequent field visits to the core stations controlling the entrance of charcoal to Port-au-Prince. This helped to alleviate the workload/travel burden of the Research Team Leader. There were several sites where no or very few trucks passed, which caused some frustration for these enumeration teams. These sites were along the border with the Dominican Republic (e.g., Cerca la Source, Belladère, Anse-à-Pitres, Kenscoff). The Research Team Leader visited all of these sites to talk to the team members and to explain to them that noting the absence of passing trucks is a form of data, and not to feel discouraged. Security Issues There were no major security incidents during the study period, although some enumerators raised concerns about safety during night shifts. Every team was provided a letter of introduction to take to the local police station upon their initial arrival, to ensure that authorities were aware of the study and knew about the work the teams would be conducting, particularly during the night. Teams also exchanged phone numbers with police officers to have a local emergency contact. In addition, each team engaged two local motorcycle taxi drivers (e.g., prepaying services for the week) to have on call for transportation in case of emergency or to facilitate travel to/ from the enumeration station and nearby hotels. Other precautions included requiring two people on night shifts and utilizing WhatsApp group chats at night. When possible, enumeration stations were in proximity to police stations and some teams were actually able to conduct night shifts within the boundaries of roadside police stations. In places where police stations were not conveniently located, the enumerators would arrange to work from the front porch or rooftop of a private home during the nighttime hours. One of the wharfs (Wharf Jérémie, located in Port-au-Prince) was flagged as a potential security risk due to its proximity to Cité Soleil, an area notorious for previous gang and kidnapping activity. 160 The Research Team Leader visited the team there and made contact with local gang leaders to introduce the study and explain who would be working there. None of the team members had any issues at the wharf after this introduction and communication with the local gang leaders. Adaptations in the Field The Morne Cabrit station had an unforeseen security concern. Multiple crimes had previously been reported at that intersection, and when the team presented their letter of introduction to the local police, the police strongly recommended that the team not conduct night shifts at the site. Once this concern was raised, the Research Team Leader traveled to the station to assess the situation. As a compromise, the night shift location was moved to the police station some 500 yards from the intersection. The Research Team Leader confirmed that the alternate night location still permitted the team to record the same data from the intersection (truck categories and origin). 160 Reports differ on the current state of security in and around Cité Soleil. 9979_Charcoal_Haiti.indd 61 2/6/19 10:38 AM 62 Charcoal in Haiti Annex 3—Possible Sources of Errors and Adopted Mitigation Measures The following list includes known potential sources of error and explains the steps taken to mitigate these possibilities: • Enumeration stations were placed on the wrong roads, intersections and/or wharfs: Phase I of the research (see above) was explicitly designed to visit all areas of the country that would be surveyed, and through observations and contextual interviews with a range of different individuals, to properly identify through observations and contextual interviews with a range of different individuals, which roads, intersections, and wharfs were the most-utilized by charcoal transporters within Haiti and across the border from the Dominican Republic. 161 The possibility exists that routes and ports shifted in the period of time between Phase I and data collection, but this is unlikely, as new roads and new port are detectable by satellite imagery. • Enumerator falls asleep; • Enumerator misses a truck; • Enumerator fabricates counts; and/or • Enumerator abandons the station. These possible sources of error were mitigated in multiple ways that address each concern. First, all stations were overseen by three different enumerators, 24-hours per day, 162 reducing the likelihood of any individual enumerator errors occurring. Second, all enumerators were instructed to take pictures of passing vehicles with their phones, and to share the photographs (using the WhatsApp Groups described above) with all the other enumerators, which served as a cross-check for the above concerns. Third, the Research Team Leader (fourth author of this report) had daily, periodic phone check-ins, and executed ongoing surprise visits to the enumeration stations, to ensure they were being properly managed as instructed during the training period discussed above. Not one enumerator abandoned station during any of the three sampling periods. 161 While charcoal entrance into Port-au-Prince used to occur by donkey transport as well, those days have ended with the introduction of motorcycles. Donkeys are still used in rural areas with poor access to major transportation routes, but they debark charcoal at locations where motorcycles or trucks arrive, rather than make the trip into Port-au-Prince. Likewise, motorcycles transport charcoal to/from truck loading locations, but not directly into Port- au-Prince as a primary mode of transportation 162  Except wharfs, which close at night since boats do not dock or unload in the dark. • Data entry error: Data were cross-checked with original enumeration sheets as well as through the WhatsApp Group photos—which record the date and time of the event. In some instances enumerators improperly noted ‘pm’ after midnight had passed, but this is an easily recognizable and correctable error. In the few instances where data were missing, we returned to the sheets to see if the issue was one of transcription. If not, the mode value (small, medium, or large) for a given area during the same sampling period, day and time was applied. In the few instances at intersections where enumerators failed to record the direction of a truck’s origin, the ratio for the given intersection was applied randomly to any missing values. Such instances occurred only a handful of times across all data sets from the three different sampling periods. • Political unrest or extreme weather events disrupt charcoal transportation flows: The first firm contracted in 2016 pulled out due to a lack of capacity to execute the study. When we tried in October to sample, we were delayed due to the passage of Hurricane Matthew. However, there were no political unrest or extreme weather events during any of the three sampling periods of 2017. • Skewing of numbers by the passage of Hurricane Matthew: This represents the largest potential source of error in the data set, but it is mitigated in four ways. First, the figures coming out of the southern peninsula can be cross-checked in reference to figures from the United Nations Environment Program’s charcoal surveys conducted in the same areas affected by the hurricane, just before the passage of the storm (UNEP 2016). Second, the percentage of charcoal coming out of the southern peninsula can be cross-checked with percentages reported since 1985, controlling for gradual increases over time. Third, ten months had passed after the hurricane before and before the first sampling period, which allowed the charcoal market to stabilize. Finally, arboreal changes to the area where Hurricane Matthew passed were detailed extensively through separate World Bank research (See Box 1 in main body of report and Annex 5). These efforts all represent opportunities to catch and control for any spike or decline in charcoal production from the South since the passage of Hurricane Matthew. 9979_Charcoal_Haiti.indd 62 2/6/19 10:38 AM Annex 4—Complementary Analysis FIGURE 22: Hourly Flows of Charcoal Entering PaP, Day by Day 200 150 100 50 0 012345678 911121314151617181910 Sunday Tuesday Thursday Saturday Monday Wednesday Friday Peak 20212223 200 150 100 50 0 0123456789 11121314151617181910 20212223 200 150 100 50 0 012345678 911121314151617181910 20212223 200 150 100 50 0 0123456789 11121314151617181910 20212223 200 150 100 50 0 012345678 911121314151617181910 20212223 200 150 100 50 0 012345678 911121314151617181910 20212223 200 150 100 50 0 0123456789 11121314151617181910 20212223 Low 9979_Charcoal_Haiti.indd 63 2/6/19 10:38 AM 64 Charcoal in Haiti Annex 5—Post-Hurricane Matthew Arboreal Assessment In 2015 Hurricane Matthew passed over Haiti, creating widespread damage and thrusting an estimated 800,000 to 1.55  million Haitians in a state of food insecurity, with approximately 280,000 categorized as severely food-insecure. The storm adversely affected crops, trees, and physical infrastructure. An estimated two out of three farmers lost approximately 75 percent of their animal livestock (FAO 2017; UN World Food Program 2017). Agricultural damage assessments ranged from $573.5 million (the Haitian Ministry of Planning and External Cooperation) to $604 million (the Haitian Ministry of Agriculture, World Bank, and FAO) (FAO 2017). Total damages from the storm, from an estimation based on Haiti’s 2015 GDP, were reported from to 2.8 billion USD (approximately one-third of Haiti’s GNP) (World Bank et  al. 2017) to $8.88 billion 163 (suggesting that Matthew destroyed the equivalent of 11.4 percent of the country’s total production of goods and services). The Haitian government-led Damage and Loss Assessment (DALA) conducted immediately following the storm used satellite imagery, interviews, and key conversations to determine losses across the country. In the agriculture sector, calculations included estimations of the value of crops lost, as well as damaged and lost trees. It was clear from this analysis that Hurricane Matthew’s passage over the southern Tiburon Peninsula resulted in massive damage to arboreal systems, including: Haiti’s remaining forests; fragmented tree stands; and a multitude of individual trees found on farms, in courtyards, on steep slopes, in deep ravines, along riverbanks, delineating property boundaries, lining roadsides, and in other isolated locations. The storm’s damage to trees ranged from complete felling, snapping of trunks at various heights, snapped branches, and partial or total loss of foliage. However, it proved difficult to decipher satellite imagery for damaged, broken trees or standing dead trees versus standing live trees, and due to this, the full impact of the storm on tree resources was not fully known. As such, the World Bank team undertook a Post-Hurricane Matthew arboreal assessment of the Grand Anse and Sud Departments (composing the lion’s share of an area colloquially referred to as the ‘Grand Sud’) of Haiti—the areas hardest 163 http://www.businessinsider.com/haiti-hurricane-matthew-economic- impact-2016-10 hit by the storm. 164 The study was conducted approximately ten months and two agricultural seasons after the passage of Hurricane Matthew. The time passed since the phenomena of interest permitted trees to recover, ensured that answers to questions more accurately reflected final outcomes for trees, permitted tree-based markets to stabilize, and allowed for a better understanding of how farmers used newly opened lands in the subsequent agricultural season. The pertinent conclusions of the full report are summarized below, and are derived from the combined findings of land-based transects (n=298) and tree surveys (n=1,682) administered by a research team that visited six different locations along the broad path of Hurricane Matthew (see Figure 23). The transects were designed to directly observe the species and quantities of trees that fell or were damaged during the hurricane, and to discover what farmers did with such trees (green points, Figure 23). The surveys were targeted toward individuals in urban or peri-urban areas that directly observed the processing, marketing, and transportation of trees, charcoal, and planks in volumes far surpassing those of more isolated rural landowners (blue points, Figure 23). The main results of the study demonstrate the types of trees fallen across the sample region and the uses of these trees by farmers, with strong implications for the charcoal market. Fallen coconut, breadfruit, and mango trees together represented 52 percent of all of the fallen trees. Farmers who owned the plots of land where sample transects took place were asked about their primary use of trees knocked over during the storm. Across the samples from all regions, Haitian farmers overwhelmingly produced charcoal from trees felled during Hurricane Matthew. Some 14 percent of respondents from transects across all regions (see Figure 23) indicated they let fallen trees remain where they were, 6 percent transformed their fallen trees into planks, and only 1 percent used their fallen trees for construction purposes (Figure 24). The research showed that the number of bags of charcoal produced by landowners sampled in transects ranged widely, with a 15,724 bag total from 261 responding landowners, an 164 Other areas that were indirectly impacted include the Nippes department, the large offshore island of La Gonâve, the southwestern and northwestern portions of the Artibonite department, and the Nord-Ouest department. 9979_Charcoal_Haiti.indd 64 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 65 FIGURE 23: Approximate Locations of Transects and Surveys Relative to the Path of Matthew The red line represents the pathway of Hurricane Matthew. FIGURE 24: Primary U se of Downed Trees across All Regions Charcoal Nothing Planks Construction 78.57% 13.93% 1.07% 6.43% 9979_Charcoal_Haiti.indd 65 2/6/19 10:38 AM 66 Charcoal in Haiti average of approximately sixty bags per transect, with four different landowners reportedly producing approximately half of all the charcoal bags across all samples. With an influx of charcoal in the market, the widespread hypothesis was that prices of charcoal would fall. This research demonstrated otherwise. Across all samples, 74.28 percent responded that the price had increased since the passage of Hurricane Matthew (Figure 25). The majority of respondents in each region also responded, although there were very slight regional differences. Across all samples, whether considering the mode or the mean, the price for a large sack of charcoal increased by 80–100 Haitian Gourdes (~1.25–1.55 USD) after the passage of the storm (Figure 26). The mode price for charcoal in the most remote locations (Dame-Marie and Abricots) was the lowest before and after the storm; conversely, those locations closest to the national highway (Les Cayes and nearby Camp-Perrin) had the highest before and after prices. Jérémie, with a well-established albeit slower maritime route to Port-au-Prince, occupies the middle of this range. These findings suggest that transportation costs to Port-au-Prince dictates charcoal prices. When asked about a post-hurricane correction of market prices, the vast majority of respondents across the samples (92.82%) believed the price of charcoal would continue to rise, while a small minority believed prices would return to pre-hurricane prices (6.56%), and even fewer still believed that prices would remain at their current rates (0.62%). When survey respondents were asked why they thought the price of charcoal would continue to rise, many concluded that prices would be dictated by supply—trees were significantly diminished in the area as a result of the hurricane. Respondents reported that charcoal trucks came more frequently to the Grand Sud region after the passage of Hurricane Matthew, likely a prospective response to an increase in supply. Across all samples, 62.33 percent of respondents indicated charcoal trucks came more frequently; 32.23 percent reported that trucks came less frequently; and 5.45 percent responded that trucks came with the same frequency as before the storm (Figure 27). FIGURE 25: Changes to the Price of Charcoal 50 100 Count 250 200 150 0 Abricots 217 13.51% 194 12.08% 209 13.01% 209 13.01% 180 11.21% 184 11.46% 89 5.54% 6 0.37% 79 4.92% 57 3.55% 51 3.18% 66 4.11% 34 2.12% Camp- Perrin Dame- Marie Region Price Increase Decrease No change Les Cayes Portà Piment Jérémie 15 0.93%4 0.25% 1 0.06% 10 0.62% 1 0.06% 9979_Charcoal_Haiti.indd 66 2/6/19 10:38 AM A National Assessment of Charcoal Production and Consumption Trends 67 FIGURE 26: M ode Price Change for Large Sack of Charcoal 100 200 Mode 600 500 400 300 0 Abricots 250 150 400 300 150 250 200 300 400 600 200 300 Camp- Perrin Dame- Marie Region Before After Les Cayes Port à Piment Jérémie FIGURE 27: Frequency of Charcoal Truck Visits after Hurricane Matthew 50 100 Count 200 150 0 Abricots 18 1.20% 83 5.51% 9 0.60% 80 5.32% 172 11.43% 158 10.50% 8 0.53% 82 5.45% 18 1.20% 71 4.72% 155 10.30% 162 10.76% 174 11.56% 117 7.77% 75 4.98% 23 1.53% 6 0.40% 94 6.25% Camp- Perrin Dame- Marie Region Frquency of visits Same Less often More often Jérémie Port à Piment Les Cayes 9979_Charcoal_Haiti.indd 67 2/6/19 10:38 AM 68 Charcoal in Haiti These results can be interpreted by considering demand for charcoal, planks, and construction. Urban demand for charcoal in Haiti is high and constant, the demand for lumber is lower on both counts. While lumber fetches a higher price as a wood- derived product, it often does not sell quickly, thus even those farmers that may have been able to pay plank sawyers up front, nevertheless hedged their bets on charcoal, given their need for immediate cash after the storm. That many people reported a lack of capital to pay plank sawyers suggests that remuneration is requested immediately by sawyers, rather than sought out after planks eventually sell, providing further evidence of a lower overall market demand for planks. The opposite is sometimes true of charcoal—some participants in the charcoal value chain may not be remunerated until charcoal is finally purchased. One possible explanation emerges for the anomalous post- hurricane increase in charcoal prices commensurate with an increase in charcoal production. Charcoal trucks came more frequently to the Grand Sud region after the passage of Hurricane Matthew—perhaps an initial response by prospective purchasers to an increase in supply. Across all samples, 62.33 percent of respondents indicated charcoal trucks came more frequently after the storm. An initial increase in charcoal trucks may have precipitated a jump in the commodity price, or sparked a gold rush for charcoal. Annex 6—The Gendered Aspects of Charcoal Production Historically, women and children were responsible for the collection of firewood for rural, domestic consumption in Haiti (Murray and Alvarez 1973). While the conclusions from the most comprehensive review on the gendered aspects of charcoal production in Haiti is dated, it remains remarkably accurate some 40 years later: Charcoal production overall is not considered to be a woman’s occupation, but women are sometimes involved to varying degrees. In areas of greater concentration of charcoal production, women tend to be more extensively involved; poorer women are more active in charcoal making than those better off. Women rarely make charcoal independently of men. Where men and women work jointly at charcoal making, women’s activities are usually centered around raking out coals and sacking charcoal. Men predominate in the wholesaling of charcoal although some women are active wholesalers. Retail sale of charcoal is handled almost exclusively by women (J. Smucker 1981, 35). Increased women’s participation in charcoal production in the area east of Port-au-Prince was noted as a deviation from an historical norm (Conway 1979, 13–14), but it is not clear whether this deviation was widespread or continued as a trend. Most recent research presents a gendered division of labor around charcoal production and consumption in a manner similarly described above. Beyond the gendered division of labor related to charcoal production and consumption, there are also differential health concerns affecting women, men, and children. While charcoal produces less smoke than traditional firewood, women and children are still disproportionately and adversely affected by smoke during the preparation of meals with charcoal (Murray and Alvarez 1973). Men are typically more adversely affected by smoke and the inhalation of charcoal particulate matter during the kilning, coal raking, and bagging tasks of charcoal production. Men are also more subject to inhaling harmful charcoal particulate matter during the transport of charcoal from rural to urban areas, while women are subject to the same dangers at presumably higher levels during charcoal retailing in urban areas. 9979_Charcoal_Haiti.indd 68 2/6/19 10:38 AM 9979_Charcoal_Haiti_CVR.indd 2 1/16/19 11:02 AM