Field Practicum Report Comparative Tree Planting Strategies: Impact and Application in Haiti

Material Information

Field Practicum Report Comparative Tree Planting Strategies: Impact and Application in Haiti
Goertz, Hans
Place of Publication:
Gainesville, FL
University of Florida
Publication Date:
Physical Description:
Project in lieu of thesis


Subjects / Keywords:
Agricultural land ( jstor )
Agroforestry ( jstor )
Canals ( jstor )
Communes ( jstor )
Control groups ( jstor )
Farmers ( jstor )
Personnel evaluation ( jstor )
Planting ( jstor )
Species ( jstor )
Trees ( jstor )
community mobilization
environmental behavior
watershed management


The 2010 earthquake brought the extreme vulnerability of Haiti into international focus. Seemingly caught in an endless cycle of social unrest and natural disaster, the country has effectively become a ward of the international community. At the heart of Haiti’s protracted struggle have been poor governance, insecurity and environmental degradation. Environmental degradation has been targeted by interventions ranging from local grassroots efforts to top-down-implemented national initiatives. Two of the most established approaches to tree planting in Haiti are reforestation and agroforestry. As part of a field practicum for the Masters of Sustainable Development Practice (MDP) program, two examples of these approaches, implemented by the International Organization for Migration (IOM) and by the Mennonite Central Committee (MCC), were studied. This paper reviews the impact and application of these different programs, and discusses their contributions to a comprehensive solution to Haiti’s environmental vulnerability. In closing, the paper highlights key lessons learned from the evaluations and their relevancy for environmental programming: touching on issues of project design, paid and voluntary labor as well as sustainability of activities.
General Note:
Sustainable Development Practice (MDP) Program final field practicum report
General Note:
The MDP Program is administered jointly by the Center for Latin American Studies and the Center for African Studies.

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
Copyright Hans Goertz. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
1039729348 ( OCLC )


This item is only available as the following downloads:

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E1U77AGSD_YELNFR INGEST_TIME 2016-04-19T22:30:09Z PACKAGE AA00023803_00001

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd


Field Practicum Report Comparative Tree Planting Strategies: Impact and Application in Haiti Hans Goertz Masters of Sustainable Development Practice Spring 2014 Committee Chair: Dr. Marianne Schmink Committee Member: Dr. Gerald Murray Comm ittee Member: Dr. Glenn Galloway


2 Acronyms and Abbreviations AECID Spanish Agency for International Development Cooperation AOP Agroforestry Outreach Program CIA Central Intelligence Agency FAO Foo d and Agriculture Organization GoH Government of Haiti IOM International Organization for Migration IEA International Energy Agency MCC Mennonite Central Committee MDG Millennium Development Goals MDP Masters of Sustainable Development Practice PREPEP Programme de Rvitalisation et de Promotion de l'Entente et de la Paix OXFAM Oxford Committee for Famine Relief PADF Pan American Development Foundation USAID United States Agency for Development WFP World Food Programme


3 Acknowledgements Kay koule twonpe soly, men li pa ka twonpe lapli. A leaky house fools the sun, but it cannot fool the rain. Haitian proverb I believe this proverb aptly captures the field practicum experience for many of us. As students, we are constructing our very first house. Since we have never built a house, we may likely assume that construction always goes to plan. We design big ambi tious houses based on many assumptions. These are submitted for review to our committee. Fortunately, the members of this committee are not new to construction. They have built houses before and can see the flaws in our plans. We work together to build a sturdier house. The host organization supports us through this process. They know the context best and can see that a house designed in Florida may not fare as well in Haiti. So more changes are made to the plan. Despite the best preparations, stude nts still encounter unexpected difficulties in constructing a house ; so we continue to draw on the support of our committee host organization and program to make repairs and change course as necessary. The final result of this partnership is a house that withstands the elements, that holds up under scrutiny and is of service to the host organization and communities. There are many parties that supported me through this practicum experience. First and foremost, I would like to express my sincere gratitu de to the members of my committee: Dr. Marianne Schmink, committee chair; Dr. Gerald M urray and Dr. Glenn Galloway. I could not have asked for a more complete committee; e ach member contributed in different and indispensable ways to a meaningful practicum Program Coordinator Cindy Tarter was also instrumental in getting my practicum off the ground. Through my practicum, I had the good fortune of working with two inspiring host organizations: the International Organization for Migration (IOM) and the Menn onite Central Committee (MCC). There are countless colleagues in each organization that assisted in this process, but several in particular that deserve recognition: Lisa Bedolla, IOM Head of Program Support; Francois Fournier, IOM Jacmel Head of Sub Offi ce; Kurt Hildebrand, MCC Country Representative; and Jean Remy Azor, MCC Desarmes Program Director. I am eternally grateful to have benefitted from the keen oversight and dedication of the IOM and MCC teams. I hope that this humble practicum contributes to their continued efforts to plant trees in partnership with Haitian communities.


4 Table of Contents Abstract 5 Introduction 6 Context of Haiti 7 Target Problem 11 Program Host Organizations 11 Program Approaches 12 Practicum Overview 18 Placement and In Country Supervisors 18 Field Sites 18 Practicum Design 23 Part 1: MCC Practicum 22 Part 2: IOM Practicum 23 Existing Program Measures 23 Practicum Measures 25 Knowledge, Attitude and Behavior Impact 25 Environmental Impact 26 Methodology 27 Part 1: IOM Evaluation Process 27 Part 2: IOM KAP Impact 28 Part 3: IOM Environmental Impact 32 Practicum Results 33 Part 1: IOM and MCC Evaluation Results Overview 34 Part 2: MCC Program Manual 35 Part 3: IOM and MCC Program Comparison 36 Part 4: IOM KAP Impact 39 Part 5: IOM Environmental Impact 46 Program Analysis 53 Lessons Learned 56 Conclusion 56 Appendices 63 References 64 Attachments: 1 3 see attached documents 66


5 Abstract: The 2010 earthquake brought the extreme vulnerability of Haiti into international focus. Seemingly caught in an endless cycle of social unrest and natural disaster, the country has effectively become a ward of the international community. At the heart of insecurity and environmental degradation. Environmental degradation has been targeted by interventions ranging from local grassroots efforts to top down implemented national initiatives. Two of the mo st established approaches to tree planting in Haiti are reforestation and agroforestry. As part of a field practicum for the Masters of Sustainable Development Practice (MDP) program, two examples of these approaches, implemented by the International Orga nization for Migration (IOM) and by the Mennonite Central Committee (MCC), were studied. This paper reviews the impact and application of these different programs, and vuln erability. In closing, the paper highlights key lessons learned from the evaluations and their relevancy for environmental programming: touching on issues of project design, paid and voluntary labor as well as sustainability of activities. Key Words: com munity mobilization, watershed management, reforestation, agroforestry, agriculture, livelihoods, environmental behavior


6 Introduction Support for environmental restoration has been building in Haiti for the last 50 years. Interventions have ranged from local grassroots efforts to top down implemented national initiatives. Throughout the history of environmental programming in Haiti, no consensus has been reached on how to best plant trees and reverse the devastating consequences of deforestation in the country. A groundswell of support for tree planting in Haiti has emerged from the aftermath of the 2010 earthquake and the increased at ambitious reforestation strategy and declared 2013 the National Year of the Environment. At the same time, external donors have mobilized funding to support tree planting act ivities. In order to capitalize on this unprecedented support, it is critical at this juncture to evaluate the efficacy of existing approaches to tree planting in shaping a comprehensive environmental strategy for Haiti. My field practicum was implemented in support of this objective. I studied two examples of the most established approaches to tree planting in Haiti: Cash for Work and a gr o forestry, as implemented by the International Organization for Migration (IOM) and by the Mennonite Ce ntr al Committee (MCC) The practicum was implemented in the watersheds of Petit Gove (IOM), Jac mel (IOM) and Desarmes (MCC). management program has planted trees and constructed erosion barriers thr ough Cash for Work labor in these watersheds since 2006. MCC has promoted agroforestry livelihoods among smallholder farmers sinc e 1992. To understand their best practices, I evaluate d activities and documented Common indicators were collected for both programs as part of a comparative study of their impact and applic ation across different contexts What worked? And where did it work best? This study was conducted to determine the contribution of these different approaches to an overall environme ntal strategy. in terms of its cost efficien cy, environmental impact and sustainability An important component of sustainability of


7 activities was the change in attitude, knowledge and behavior in program participants. I developed a set of indicators in collaboration with the host organization to measure progress against these objectives. M easurements of the intervention group s were compared to matching constructed control groups to determine the extent to which progress can be attributed to the program was also evaluated on the basis of its environmental impact by measuring its tree survival rate However, the primary objective of the practicum with MCC was to document their methodology for tree planting in order to disseminate these lessons to other environmental partners. Therefore I did not evaluate the MCC agroforestry program in terms of its sustainability and other environmental measures, as w as done for the IOM watershed management program. In line with the different objective s in studying each program, the evaluation tools varied for the two organizations: documentary review s semi structured interviews and direct obser vation were conducted for both IOM and MCC, while oral surveys were conducted exclusively for IOM with the assistance of national counterpart s The results of the formative evaluation of these programs were compil ed in a report and presented to each host organization In dis cussion with the host organization s, I also develop ed a set of lessons learned to disseminate to other organizations engaged in environmental programming in Haiti these are included in the conclusions section of this report The practicum is intended to contribute to ongoing and future efforts by critically reviewing the de sign and implementation of these particular models This role provide d me with experience in data collection, analysis and presentation, which complement ed my previous experience in other stages of the project cycle. Context : Haiti is the poorest country in the Western Hemisphere. Ranking 158 th on the 2011 Human Development Index, Haiti lags behind in every significant measure of human and environmental welfare. 1 Seemingly caught in an endless cycle of social unrest and 1 UNDP 2012


8 natural disaster, the country has effectively become a ward of the international community. At the heart of been poor governance, dual destruction of the ecological basis for human 2 export plantation economy. At its colonial peak, Haiti was the largest exporter of sugar and coffee t o Europe. Vast sections of the island were cleared of trees to support production on plantations. After independence, a burgeoning timber industry cleared hardship introduce d by the eradication of the Creole pig in the 1980s and the U.S. trade embargo in 1991 further contributed to tree cutting, primarily for charcoal production. Deforestation continues today as agricultural plots became increasingly fragmented through inhe ritance and remaining trees compete with cash crops. 3 4 Topography has compounded the challenges of deforestation in Haiti. Approximately two thirds of the country is mountainous. 5 P opulation growth and declining productivity have forced farmers to clear steeper plots to cultivate crops. Most farmland is on slopes of over 20%, and the majority of farmers has several small plots totaling less than two ha. As a result, most hillsides a re visibly eroded, and according to USAID, a third of all land is severely degraded. 6 Haiti receives an average of 1,461 mm of rain annually, with the majority of rain falling during two concentrated seasons from April to May and during the hurricane season from July to November. 7 this annual 2 Schulz 1997 1998 3 Smucker, et al. 2002 4 Swartley and Toussaint 2006: 22 5 FAO 2012 6 White and Jickling 1995 7 FAO 2012


9 rainfall, as rain washes down denuded ravines. With much of the forest cover and topsoil gone in these upper catchments, many of Haiti's rivers are now highly uns table, changing rapidly from destructive flooding to inadequate flow during the course of the year. These drastic fluctuations pose a major threat to downstream settlements during the rainy season and to irrigated agriculture during the dry season. Haiti has a strong agricultural tradition dating back to the colonial era. This tradition has continued in rural areas where 75% of households is engaged in some form of agriculture. Despite this history, Haiti is the most food insecure country in the Western Hemisphere today. The World Food Programme (WFP) estimated in 2008 that 25% of households were food insecure. Most farmers do not produce enough to feed themselves, let alone the whole country: 68% of food items consumed by rural households are bought o n the market. 8 National agricultural production covers only 47% of the national food needs. Consequently, access to food is determined more by purchasing power than production. Food expenditure currently represents approximately 59% of household spendi ng in Haiti. 9 population has been poorly represented in government and decision making. This has historically led to the omission of agriculture and by extension, the environment from the national agenda. 10 The exclusion of the rural population has been reinforced by the reliance on the French language by the government, particularly in the judiciary branch. Land disputes settled by the legal system have generally f avored the French speaking Haitian elite. As such, the insecurity of land tenure and poor access to credit remain significant barriers to private investment for farmers. 11 Despite this, trees have found a 8 This could also be understood as a result of the market orientation of Haitian farmers, as opposed to a failure of subsistence agriculture. 9 WFP 2008 10 Cohen 2010: iv 11 Smucker, et al. 2002: 9 12


10 foothold on informally owned land in Haiti as sub stantiated by studies of several tree planting villages. 12 A 1995 study by Smucker and Timyan identified informal land ownership as the key determinant to tree planting. Haiti, land is instead purchased and subdivided informally with community members bearing witness to the transfer of ownership. These informa l land tenure arrangements provide secure control over trees, and by extension, the motivation to plant trees. 13 A statistical analysis by Bannister and Nair of land tenure and tree planting behavior confirmed this same tendency for more trees to be planted on plots under more secure tenure. 14 Charcoal remai ns the dominant source of energy in most urban and town households. It is estimated that over 80% of charcoal is consumed in the capital of Port au Prince. Mos t rural households rely on fuel wood and only produce charcoal for market. Charcoal production has nearly tripled over the last three decades, from 12 million metric tons in 1980 to 32 million metric tons in 2010. 15 Declining soil fertility and falling agricultural commodity prices have led many rural Haitians to intensify charcoal production as a means of securing cash income. Estimates of the national energy need met by charcoal or firewood in Haiti range from 66 85%. The reasons for this reliance are clear: charcoal is cheap, accessible and more reliable than the supply of propane or electricity Charcoal consumption is primarily urban, while rural households rely primarily on fuel wood. The decision of rural households to use fuel wood is both economical and practical: fuel wood is more available in rural areas, cheaper, and less labor intensi ve to produce than charcoal. 16 12 Murray and Bannister 2004: 7 13 Smucker and Timyan 2005 14 Murray and Bannister 2004: 7 15 Van der Plas 2007:21, 30 16 Howard 1998:18; Van der Plas 2007:3


11 Target Problem : Deforestation and soil erosion are turning parts of the country into a sea. 17 Over the last 50 years, arable land per capita decreased from 0.38 hectares to about 0.16 hectares as a result of population growth, over cultivation, deforestation and erosion. 18 Trees are vital to the so il and water cycle that farmers depend on. Their leaves contribute organic matter and nutrients to the soil and release moisture that forms rain clouds, while their roots stabilize the soil and channel water into the ground. As soil fertility declines, Haiti can no longer feed itself and now imports mor e than half of its food. 19 Food insecurity, frequent flooding and widespread unemployment have made tree planting a government and donor priority in Haiti. Program Host Organizations The International Organization for Migration (IOM): Established in 1951, IOM is one of the leading intergovernmental organizations working on migration issues. membership consists of 127 member States, and 21 observers including States, international governmental and non governmental organizations. S ince 1992, IOM has held observer status at the United Nations and has cooperation agreements with many UN agencies. community development and stabilization to communal go vernance, border management and counter trafficking. In 2004, IOM launched PREPEP ( Programme de Rvitalisation et de Promotion de l'Entente et de la Paix ), a conflict mitigation and community stabilization program through funding from the United States A gency for International Development (USAID). PREPEP engages communities in conflict prone areas in their own stabilization and improvement through the implementation of small scale, high impact projects. IOM has pursued watershed management intervention s in six regional offices a s part of a broader disaster risk reduction (DRR) initiative under the PREPEP. Watershed 17 Foxx 2012:105 108 18 Lundahl 2002 19 WFP 2008


12 management is an integrated response to a variety of identified problems: severe flooding, declining soil fertility, volatile river flows and widespread unemployment. 20 In Millennium Development Goal (MDG) fund and the Spanish Agency for International Development Cooperation (AECID). The Mennonite Central Committee (MCC): Founded in 1920, MCC addresses basic human needs such as water food and shelter and works alongside churches and communities worldwide in a variety of efforts to build peace MCC has worked in Haiti since 1958. MCC workers serve with two human rights networks, which give workshops on human rights and document human rights violations. MCC suppo rts other partner organizations with programs in education, job training, literacy, conflict resolution and microfinance. In 1992, MCC launched a tree planting program in the Artibonite Valley. The program has since expanded to include community tree nur series and environmental education. Since the earthquake, MCC has also been involved in relief and reconstruction activities in the directly impacted area. 21 Program Approach es : Support for environmental restoration has been building in Haiti for the last 50 years. Interventions have ranged from local grassroots efforts to top down implemented national initiatives. Throughout the history of environmental programming in Haiti, n o single overarching strategy for reversing deforestation has emerged. Two of the most established approaches to tree planting are reforestation and agroforestry. T wo examples of these approaches are the current programs implemented by IOM and MCC IOM P rogram Approach : IOM implements watershed management in the upper catchments of several mountain ranges prioritized for intervention by the Haitian 20 IOM 2012 21 MCC 2011


13 Ministry of the Environment. 22 The rationale of this approach is that if you catch rain where it falls then you can prevent rain runoff from reaching the velocity and volume that causes erosion and flooding downstream. Agricultural technicians lead paid community members in the construction of stonewalls and contour canal s in vulnerable ravines on public a nd private lands. Trees and deep rooted grass species are then planted to reinforce these barriers. These hedgerows capture water where it falls, nourishing plants and stabilizing soil. Overtime, these hedgerows form productive terraces that can be cultiva ted. IOM establishes tree nurseries in intervention sites to produce trees. The nurseries are run by community members and subsidized by IOM. IOM pays nursery workers at a rate of between $0.40 and $0.60 per tree produced Site selection : The process of selecting a si te for intervention begins at the request of host communities. Representatives submit a written request to the regional IOM office to implement watershed m anagement activities, or other type of project Following this request, IOM dispatches agricultural technicians to the community to determine if the site is appropriate for intervention. Next, IOM staff meet with the community to determine their commitment to the project. This is a key step in the selection process since projec t requests do not always have broad community support which can threaten project implementation and sustainability The meeting also serves to inform communities of the responsibilities they assume with project implementation. Once the organization has secured a commitment from the community and approval from the donor, IOM technicians ide ntify intervention sites in the targeted watershed. Sites are identified on the basis of their environmental state, contribution to flooding and proximity to the commu nity. Selected sites are located in the upper catchments to maximize the downstream benefits of activities. Sites are characterized by steep slopes, sparse tree cover and marginal soil. As such, the land is not in competition with traditional agricultur e. While the lands are privately owned, they frequently have 22 Fournier 2013


14 approaches the l andowners at intervention sites to negotiate access to the land. Landowners commit to not pla nting root crops and harvesting trees planted on their land for ten years. In exchange, they receive the environmental benefits of the watershed management activities and the economic benefits of eventually harvesting the trees. Participant recruitment : There are t wo types of participants recruited by IOM : Landowners : are selected based on the location and environmental state of their land. They must fulfill a set of requirements before participating in the project: agree to host intervention activities, produce clear title to their land and sign a maintenance contract pledging to protect project investments for 10 years. These investments in erosion control, soil fertility and tree cove r improve the quality of their land. Farmers are free to cultivate the land and harvest the trees as they see fit after this 10 year period. Wage laborers: the project workforce is comprised of residents of the host community. IOM does not directly selec t participants, but rather engages recognized community leaders to compile worker lists. IOM draws participants from this list working their way down from the top as far as project activities and funding permit. Participants work two week rotations in o rder to extend the opportunity to work to as many residents as possible. Residents must be at least 18 years of age to participate in activities. At least 40% of the participants selected are female. In practice, men and women play different roles in project implementation. Women are charged with carrying trees from nurseries to intervention sites where they are planted Men are primarily engaged in the construction of erosion barriers. At the end of their two week r otation, participants are paid the national minimum wage of 250 HTG wage per day, approximately $6.25. These projects are an important source of income in rural communities with few formal employment opportunities. MCC Program Approach : MCC takes a livelihoods approach to tree planting that promotes the micro economic benefits of agroforestry over the macro ecological benefits


15 of reforestation MCC subsidizes the production of trees at community run nurseries at a rate of $0.1 5 to $0.35 per tree produced These trees are distributed freely to community members once per y ear, during the rainy season. Distribution p articipants receive between 60 100 trees to plant on their own private holdings. They have complete control over the management of these trees: how they are planted, raised and harvested. MCC conducts follow While primarily economic driven, this approach also provides indirect ecological b enefits Farmers te nd to harvest trees as needed, rather than all at once. In addition, many of the forestry species planted will coppice and provide multiple harvests over their lifetime. The net result of this livelihood strategy is increased tree cover on these land hol dings Site Selection : The MCC agroforestry program was first established in Desarmes and has since expanded to 22 comm unities in the Central Plateau. Originally MCC produced trees at professionally run tree nurseries in Desarmes to distribute to these s urrounding communities. Communities were selected based on their proximity to Desarmes. Since then, MCC has decentralized production to community run nurseries. MCC trains and provides inputs to residents to establish nurseries in t heir communities. H ost c ommunities are responsible for securing land for the nursery, but are advised by MCC to locate land near a water source and accessible to other community members. N urseries distribute trees to community residents. Distributions start early in the mo rning so that participants can plant trees during the cooler part of the day to give them the best chance of survival. Participants decide where to plant the trees they receive. In practice, participants plant trees in dedicated tree stands or interspers ed alongside traditional crops in silviculture arrangements. This usually translates to planting trees on small to medium private holdings, in valley lands or on gentle slopes with favorable soil conditions as opposed to the marginal conditions encounte red in the upper catchments targeted by IOM for watershed management. Participant recruitment : There are two types of participants recruited by MCC: Tree nursery committee members : Each community nursery is run by a


16 committee comprised of a nursery manager, assistant manager, secretary, treasurer and five general members. When the nursery is first established, the community elects all nine members of the committee. These members then participate in a random drawing to determine the length of their service at the nursery generally between 1 3 years. After this period has elapsed, new committee members are selected to replace outgoing members. This rotation system ensures that o ther members have the opportunity to participate in nursery management, while still maintaining continuity of members from year to year. MCC contracts the committee to produce a mutually agreed upon number of trees to be distri buted and sold to the commu nity. The production season runs from January through August. During this period the nursery manager is paid a pays all committee members a subsidy at the end of the season f or the amount of trees produced and distributed. The committee has the opportunity to earn additional money by selling seedlings to the community or to other organizations. The committee takes responsibility for engaging comm unity members to purchase tr ees, while MCC serves as an intermediary for tree sales to other organizations. Distribution recipients : The day of distribution is publicized over the radio and door to door in host communities The practice has been ritualized in these communities over the last 20 plus years as something to expect and plan for each year. Every member of the community is invited to participate in the distribution and even surrounding communities that do not have their own nurseries. Participants are instructed to brin g a container to receive trees in. Nurseries distribute a set amount of trees to each participant depending on the number of trees available and participants present. Multiple members of a household typically attend distributions to increase the number o f trees the family receives. Children are encouraged to participate as well, but receive fewer trees. There are no requirements for participation other than arriving at the appointed time. Participants have the option to purchase trees at a subsidized r ate if they would


17 like additional trees than their distributed allotment. Program Objectives : These approaches have succeeded in planting millions of trees over the last ten years, 23 but by pursuing different end objectives: On the one hand, MCC promotes trees as a cash crop for farmers to plant, manage and harvest at regular intervals alongside traditional crops. This agroforestry approach has cascading environmental benefits, but t long term livelihoods. Alternatively, IOM plants trees as a part of a larger watershed management strategy to mitigate erosion and flooding. This takes place on a macro environmental level, but also creates short term employment for Cash for Work participants and contributes to long term agricultural livelihoods for landowners by mitigating erosion. Over the long term the objectives of these two programs are clearly complementary. Program Background : Over the course program has constructed 278,453 m of stone check dams, excavated over 107,582 m of contour canal s and planted 2.7 million trees. These Cash for Work interventions have created over 128,000 short term jobs. In ad dition, a total of 54 community nurseries were established to provide project trees. 24 The majority of these nurseries have since been closed as project activities have come to a close in many areas raising concerns about the sustainability of this compo nen t of the intervention. Through project planning meetings, IOM has raised environmental awareness and engaged communities in the sustainable management of their land. MCC began agroforestry work in 1983 in the Artibonite Valley in central Haiti. The program has since assisted farmers to plant more than 6 million tree seedlings. Currently, MCC supports 23 community tree nurseries that raise and distribute more than 400,000 seedlings annually to smallholders in the valley The nurseries carry out annual distributions, providing 75 100 trees to each household. Distributions are 23 IOM has planted over 3.2 million trees to date in the watersheds of Petit Gove and Jacmel, which were studied for this evaluation. The MCC agroforestry program has distributed a total of 7.6 million trees in the Central Plateau. 24 IOM PRE PEP 2012


18 reinforced with extension training to instruct beneficiaries to plant, tend and manage tree stands. MCC also sponsors environmental education in 12 schools in the commune of Desarmes. At present, 280 students are participating in the year long classes, taking field trips to community tree nurseries and starting nurseries at their schools. Practicum Overview Placement and Host Organizations : From May until August 2012, I worked under the supervision of IOM and MCC at their respective sites in Petit Gove Jacmel and Desarmes. The IOM Head of Office host ed me at his house in Jacmel I reside d at a private guesthouse in Petit Gove as there is no longer an IOM field presence in the region MCC host ed me at their field office in Desarmes. At several stages of the practicum, I resided in Port au Prince meeting with central office staff and analyzing field data. During my time in the capital, my parents hosted me at their residence in Bourdon. I negotiate d for both organizations to provide an assistant to accompany me on data collection in the field. In Petit Gove, I paid team of three former IOM employees to assist me in admini stering an oral survey. I had access to a private vehicle during the practicum to travel to tree planting sites and community nurseries in the se regions. In Country Supervisors : Country Supervisors Lisa Bedolla, Director, Program Support Unit, IOM, Port au Prince Haiti Kurt Hildebrand, Country Representative, MCC, Port au Prince Haiti Field Supervisors Francois Fournier, Head of Office, IOM, Jacmel, Haiti Jean Remy Azor, Program Director, MCC, Desarmes, Haiti


19 Practicum Field Sites: Petit Gove IOM: located 72 kilometers east of Port au Prince, in the Ouest department. The population of the municipality was 117 504 in the 2006 census, with an estimated 12,000 living in the town. 25 Petit Gove is neighbored by the sister town of Grand Gove. Eu ropean traders settled both towns in the mid 17 th century. Their ports are still active today and have a reputation for importing contraband. 26 Located on the National 2 Road, the primary economic activity is commerce. The area supplies Port au Prince with fish, charcoal, fruit, and some crops. It has been the site of unrest dating back to the 2010 national elections. Protestors have barricad ed the National Road several times during this period to put pressure on the government to meet their demands. These demands range from recognition of voting fraud to repair of the municipal electrical grid. Petit Gove is in the area directly impacted b y the 2010 earthquake. The municipality has since been the scene of international relief and reconstruction efforts, including watershed management activities by IOM and other organizations. These activities have been implemented to mitigate flooding in downstream areas while also creating employment for populations displaced by the earthquake. IOM has operated a regional office in the city since 2006. The organization has undertaken its most extensive tree planting in the country in this commune of Pet it Gove. A random selection of the 16 total sites where IOM has intervened in the commune produced 6 field sites that were visited for the evaluation, plus 3 matching control communities. These sites are listed in the Evaluation Results section on page 25. They can be broadly characterized as rural, high elevation, agrarian communities. The majority of residents are engaged in the production of cash crops and livestock for sale in urban markets. Jacmel IOM: is set on the southern coast in the Sud Est department. The French founded Jacmel in 1698 on an older indigenous Ta no settlement. Starting in the 18 th century, the port city rose to prominence for its coffee production and trade. The city fell 25 UN 2006 26 Clammer 2012: 141


20 into decline in the 1930s during the American occu pation when a policy of centralization concentrated resources and trade in Port au Prince. 27 Despite its proximity to the capital, Jacmel is fairly isolated by the surrounding mountains which make passage to the city difficult. In recent years, Jacmel has regained some of its past prosperity as a major tourist destination, trading in handicrafts and hosting an annual Carnival celebration. The population of the municipality is 137,966, 28 with an estimated 40,000 living in the city itself. astal location has made the city ideal for shipping and tourism, it has also made it vulnerable to flooding. The city is at sea level, bordered by the confluence of the Riviere Gauche and and Grande Riviere and surrounded by the Chaine de La Selle mountai n range. In response to this flooding, IOM has been engaged in watershed management in the Jacmel watershed since 2012. The organization has intervened upstream in La Vall e to reduce rain and sediment runoff into the rivers that flow through Jacmel. Th ese efforts have complemented the construction of gabion walls along the river to mitigate flooding in the area. La Vall e is part of the Chaine de La Selle mountain range located north of the city of Jacmel on the western bank of the Grande Riveire La Vall e has historically been a coffee well suited to cultivating traditional shade species of coffee. However, large scale coffee production has disappeared in the last generation and only exists today on a subsistence level. Residents attribute this decline to the unreliable coffee export market, deforestation of the shade canopy and urban migration. The area is comprised of five communities. IOM has interven ed in three of these communities: Ternier, Laviale and Durivier. Field visits were conducted at these intervention communities as well as at the matching control community of Lakoupe. The community of Ridor e was excluded from the study because of its ur ban setting and wealthier populace. 27 Clammer, 2012: 143 144 28 UN 2006


21 Desarmes, MCC: The MCC agroforestry program services 22 communities from their regional office in Desarmes in the Central Plateau region of the Artibonite department. The town of Desarmes has approximately 16,000 resid ents. 29 Commerce is the main economic activity in the area, but is generally supplemented with subsistence or cash crop agriculture. People travel far to buy and sell harvests in the markets. Commerce between towns and larger cities has been spurred by t he rehabilitation of regional roads. The road between Pon t S onde to Mirebalais was recently completed, opening up trade with Gonaives, Saint Marc and Port au Prince Construction poles and charcoal are two of the main products coming from Desarmes to the cities as they bring the highest profits. This is putting increas ing pressure on local trees 30 The lower plains in the valley surrounding Desarmes are covered by vegetation and trees, while the mountaintops are largely deforested. The Artibonite department receives approximately 1,800mm of annual rainfall. 31 Much of this is channeled through the Artibonite River which supports agriculture in the valley. Farmers cultivate rice in the surrounding plains. In the drier hillsides farmers plant pige on peas, corn, beans and sorghum. Plantain, bananas and legumes are planted in ravines. Fruit trees are becoming more numerous after a long decline under the pressures of livestock predation and charcoal production. Goats, cattle and chickens are widely raised in the valley. Grazing is managed during the agricultural season, but animals are often left to roam during the dry season posing a threat to new trees. 32 The government provides minimal services to the area. There are some national schools and a court system with two judges, but these systems suffer from poor management and underfunding. There are several private schools but th e tuition costs are prohibitive Rural students who graduate from the education system have limited career opportunit ies. 33 There are two health dispensaries in the area plus the Albert Schweitzer 29 UN 2006 30 Goertz 2013: 7 9 31 FAO 2012 32 MCC 2010 33 MCC 2012


22 Hospital in Deschapelles and the Health Friends Hospital in Vert, but increasingly more people are going to other places such as Cange Ti Riviere Artibonite and Port au Princ e to get better health care and services. DINEPA, the municipal water authority, has worked with several international partners including MCC to install public water fountains. These water points are managed by committees that collect user fees for maintenance and repairs. A senator for the department has led efforts to connect Desarmes to the power grid in Vert, but the system suffers from disruptions and shortages. Several local businesses and households have solar panels or generators. Gas pow ered stoves are becoming more common among urban residents, but the majority of the population still depends on charcoal for its energy needs. Table 1. Practicum Timetable Dates Location Activity May 16 USA Haiti Depart for Haiti May 16 2 8 Port au Prince, HT Me t with MCC and IOM central office staff May 30 Jun 12 Jacmel, HT Monitor ed ongoing IOM reforestation activities Jun 12 Jun 21 Petit Gove HT Evaluate d completed IOM reforestation activities Jun 21 Jun 23 Port au Prince, HT Analyzed collected field data at central office Jun 23 Aug 10 Desarmes, HT Monitor ed ongoing MCC reforestation activities Aug 10 Aug 1 3 Port au Prince, HT P resent ed preliminary findings to host organizations Aug 1 3 Haiti USA Return ed to Gainesville


23 Practicum Design Part 1: MCC Practicum Upon the req uest of the host organization, I conducted a study of the MCC Agroforestry program as part of my practicum. MCC decided that an evaluation would not advance the since after 20 years the organization was confident in the success of the program and had a secure funding stream to continue activities (whereas IOM intended to use the results of the survey as a potential fundraising tool). was to disseminate the best practices of the Agroforestry program. In service of this goal, I documented the MCC approach in a program manual. The manual draws on a review of program documents, interviews with key informants and field site visits. I worked with MCC to circu late this document to other partners enga ged in tree planting in Haiti. MCC supports 22 community nurseries in the central plateau as part of its Agroforestry program; 20 of these nurseries were visited during the practicum see figure 1. No surveys were administered, a s these visits were primarily to assist in tree distributions and document program practices. The remainder of the Practicum Design section is devoted to the IOM evaluation. Discussion of the MCC study resumes in the Practicum Results section. Practicum Design Part 2: IOM Practicum Following the plan of the initial practicum proposal, I conducted an evaluation of the IOM watershed management program. The o utcomes for the program can be divided into two categories: environmental and behavioral. I assessed the achievement of these outcomes through documentary review surveys interviews and site visits I developed a scoring scale f or each of these instruments to measure the magnitude of change. Existing Program Measures : While no external evaluation has been conducted to date, IOM staff have tracked outputs for program activities. The program has institutionalized this monitoring process through the development of a management information system.


24 This project database organizes program activities into individual micro grants, which are assi gned a set of indicators and targets corresponding to their objectives These indicators are tracked through weekly progress reports and final reports submitted by the field offices. The grants are either implemented as a single project or as phases in a larger project. The IOM project database was designed as a central reporting system to compile outputs of actual results submitted by field staff. During implementation, the database was utilized to systematically organize information to inform administr ative decisions, track progress towards program goals and meet donor reporting requirements. While this monitoring system had the potential to serve as a dynamic, real time evaluation tool, the execution of the system suffered from several shortcomings : 34 1. The database did not track information at project sites and communities providing no point of comparison for program progress. 2. The data submitted by the field offices did not always reflect results on the ground. 3. The donor selected indicators in the database track only short term outcomes, such as trees planted, and do not capture longer term outcomes, such as tr ee survival or maturation rate. Given the awareness; impro these short term proximate outcomes do not capture the full scope of the environmental impact. 4. The database tracked quantitative data on a relatively limited set of indicators (de fined by initial donor, USAID). These variables do not capture the more qualitative and longitudinal components of attitude and behavior change. In addition to the project database, IOM has reported on program activities in weekly, monthly and quarterly publications. As these publications were used as a fundraising tool, they are promotional and do not provide an objective evaluation of activiti es. 34 Fournier 2013


25 However, d espite these limitations, these records still serve as valuable source s of program data for documentary review during the program evaluation. Practicum Evaluation Measures: The following measures were used to determine if the criteria for each stage of impact model were satisfied in order to advance from inputs outputs and finally to short, medium and then long term outcomes I developed the imp act model with input from the organization as one of the practicum deliverables see Attachment 2 The model is intended to clarify and focus the mission of t he watershed management program as one of the deliverables of my practicum. It also serves to guide the evaluation of the program. The evaluation measures for the program can be divided into two categories: Knowledge Attitude and Behavior (KAP) impact and environmental impact. The Practicum Design, Methodology and Results sections of this report are organized around these two categories. Knowledge Attitude and Behavior Impact s : of the program were assessed by comparing the following measurements in intervention and control groups: Change in perception/understanding of environmental processes Change in attitude toward the environment Change in sense of responsibility toward the environment Accrual of knowledge of best land management practices Change in sense of agency/capacity in environmental management Demonstration of new practices on private plots Demonstration of maintenance/protection on project sites These objectives wer e measured first on an individual level through oral surveys The results were then aggregated to provide an indicator of overall community norms. The se objectives assess the performance of the program in fostering pro environmental attitudes, providing knowledge of best practices and promoting responsible


26 environmental behavior. Change was measured by comparing responses between program participants and a matching control group. In this case, the control group served as a proxy baseline (since no befor e data exists). I also reviewed the demographic composition of survey participants to compare the practices of different sub sections of the population. Analyzing this data provide d insights into the determinants of environmental knowledge, attitudes and behavior. E nvironmental I mpact : The IOM watershed management program has the increasing environmental awareness; improving tree cover; mitigating flo S tructured observation provided direct evidence of I conducted this structured observation at intervention sites with the support of a field assistant. I developed a field visit checklist comprised of measures of environmental health to guide an environmental assessment of intervention sites I organized the indicators around the environmental goals of the IOM watershed management program See Table 2 The overall program assessment also included the sustainability of activities as measured by erosion barrier durability and erosion barrier maintenance. Table 2 Environmental impact indicators by IOM program goal Program Goal Relevant Environmental Impact Indicators Improve tree cover tree spacing, tree cover/ survival tree diameter at breast height, basel area, tree predation Decrease erosion (increase soil fertility) soil accumulation vegetation cover quality of erosion barrier construction Mitigate flooding stream discharge, stream bank width, distance of water from stream bank Not measured. In order to assess flood mitigation measurements needed to be taken during the wet and dry seasons. Given the limitations of the practicum, anecdotal evidence was collected from downstream residents to provide a qualitative measure of flood mitigation.


27 Methodology Part 1: IOM Evaluation Process Evaluation Instrumentation : The evaluation design employ ed several data collection methods in order to assess the imp act of the program These methods contribute to one of four objectives : the design of evaluation instrumentation selection of intervention sites and survey participants, measurement of knowledge, attitude and behavior impact and measurement of environmental impact see Table 3 Table 3 Data collection methods Collection Method Type of Data Data Objective Participants/ Sources 1. Documentary Review Quantitative and Qualitative Selection of intervention sites and selection of survey participants Program reports, program manual, IOM project database 2. Open ended interview Qualitative Design of evaluation instrumentation K ey informants 3. Oral survey Quantitative and Qualitative Measurement of KAP impact Heads of household control and intervention 4. Structured observation Quantitative Measurement of environmental impact Principal investigator 1. Documentary Review: I review ed program documents and monitoring data to gather information on project sites, intervention activities and participation figures. This information inform ed my selection of comm unities and field site visits for the evaluation Intervention sites were selected using a random number generator Matching control sites were selected based on relevant social, demographic and geographic characteristics. A review of the project database provided an estimate of the cost efficiency of activities. I calculated cost efficiency by dividing the program outputs by the cost of achieving them (i.e. cost per meter of stone wall constructed). 2. Open Ended Interviews: I interviewed key informants program staff, donor representatives and com munity leaders activities and accomplishment of outcomes. These criteria were in turn applied to the


28 design of survey and observation methods in order to determine if IOM was successful by their own definition. Key informants included IOM program staff, donor representatives and community leaders. I also drew on interviews to identify confounding factors, threats to succ ess, as well as lessons learned. Methodology Part 2: IOM KAP Impact 3. Oral Surveys : I conducted oral surveys with members of the control and intervention communities. Since no baseline survey was conducted prior to implementation of the program, the control group served as proxy baseline. By comparing the survey responses of control and intervention groups, I was able to approximate the impact of the program on partici pants. The surveys gather ed qualitative data from participants about their attitudes, priorities and practices. Participants were assessed on their knowledge of environmental services and best management practices. In addition, the survey asked interven tion participants about their role in the program and how participation has impacted their current livelihood strategies. See int erview questions in Attachment 2 I organized q uestions into four blocks to address the behavioral objectives of the programs outlined earlier in the Practicum Design section of the report. A. Introduction, participation, planning and goal setting: This section was inte nded to introduce the survey assess the goals of the respondent and determine their participation in implementation (or why they did not participate). Questions 2 3 assess ed or her community and its most pressing challenges. Questions 4 6 assess ed ved personal and community satisfaction with the project. B Environmental knowledge and attitudes: The opening question aimed to gather general perspectives on trees. Responses were coded into categorie s during analysis. Questions 9 13 sought to underst barriers to tree planting. Questions 12 14 inquire d into the driver s of deforestation.


29 C Sense of responsibility and agency: These questions asses s ed outlook on responsibility for the environmental stewardship. At the conclusion of the program, the community signed a maintenance agreement, pledging to water and mulch trees and protect tr ees from predation. Questions 15 and 16 aim ed t o assess whether the participant is aware of and committed to th is responsibility. Question 17 evaluate d the sense of agency of the respondent. Questions 17 and 18 assess ed the tree planting behavior of the r espondent. Question 19 inquired into the perc eived threats to tree survival. D Environmental behavior and technical skills: Questions 20 and 21 were open response questions and maintenance techniq ues The follow up questions probe d into whether they learned these techniques from the program and how effective they are in practice. I also collected d emograp hic data from survey participants to further analyze their interview responses. These independent variables are utilized to assess how program participation and socio cultural, economic and ecological conditions influence environmental behavior (the de pendent variable) see Table 4 These variables provide information on the demographic composition of the intervention commun ities studied They also provided another point of comparison between intervention and control groups. I selected these v ariables based on a literature review of Haitian agrarian culture, livelihood strategies and theories of environmental behavior. Table 4 Independent v ariables collected as part of survey participation in project N of people per household age land tenure/ownership sex other land holdings religion size of total land owned literacy slope of land primary livelihood soil types These variables are operationalized in Appendix A.


30 S urvey S ampling: I selected a matching control community based on relevant criteria in order to best approximate the intervention community : population size and density of community primary livelihood s level of education, land tenure and geographic features of community (i.e. elevation, state of erosion, tree coverage and proximity to urban areas). I collected this information from interviews with key informants and a review of program documents. I systematically selected i nterview and survey participants from worker payroll records in the case of intervention group members, and from school enrollment lists in the case of control group members. Public schools are one of the few sources of reliable d emographic and community data in rural sections of Haiti. A parent of the selected student was selected for the survey. These lists contain ed contact information in addition to some demographic data. I employed a random number generator to complete the selection for each group Where the selected participant could not be located, the next selected name on the list w as contacted I originally designed the sample size to reflect population size and the coverage of the program in each intervention community and planned to survey equal members of the intervention and control groups However in practice, I was not able to achieve the same sample sizes for both groups due the limite d t ime and human resources available in each region. As such, the sample size in the two regions, Petit Gove and Jacmel, are different. In addition, the intervention group is significantly larger than the control group. These differences are in keeping with the wishes of the host organization that requested that I study intervention participants and the longer established Petit Go ve sites more in depth A total of 106 randomly selected adults took the survey. 35 Of these respondents, 72 participated in the program (68%), and 34 did not participate (32%). The distribution of 35 Adult as defined as 18 or more years of age


31 survey participants is detailed in Ta ble 5 Table 5 IOM field site selection Commune Community No. of Survey Participants Jacmel Ternier 8 Jacmel Laviale 9 Jacmel Durivier 9 Jacmel Lakoupe 15 Petit Gove Courtis 8 Petit Gove Volant 7 Petit Gove Nan Bazile 8 Petit Gove Vallue 8 Petit Gove Verger 9 Petit Gove La Regale 6 Petit Gove Lakoupe 10 Petit Gove Sous Pierreau 9 Total 12 communities (9 intervention, 3 control) 106 Matching control sites selected for comparison with intervention communities Respondent composition: Table 5 describes the demographics and land ownership patterns of respondents. These independent variables were used to compare the two groups as well as to compare respondents within each group.


32 Table 5 Comparison of survey respondents: intervention and control Independent Variable Value Intervention Group (n=72) Control Group (n=34) participation in project 100% 0% age Average of 41, median of 37 Average of 36, median of 30 sex 56% Female, 44% Male 52% Female, 48% Male religion 55% Catholic, 26% Protestant, 8 % Evangelist, 6% Baptist, 2 % Vodou, %2 other 48 % Catholic, 30 % Protestant, 5 % Evangelist, 10% Baptist, 6 % other literacy 48% literate 66% literate primary livelihood 40% farmer, 25% merchant, 10% construction, 8% student, 5% tailor, 12% other 37% farmer, 35% merchant, 3% construction, 18% student, 0% tailor, 7% other N of people per household Average of 5 Average of 5 land tenure/ownership 90% own (61% inherited, 29% purchased), 3% rent, 7% no land 92% own (72% inherited, 22% purchased), 5% rent, 3% no land estimated total land size Average 1.3 ha Average 1.1 ha Differences between groups: The most pronounced differences between respondents in the intervention and control groups were found in age literacy primary livelihood and to a lesser degree, estimated total land size The rationale for each of these differences, and their potential influence on the survey results are considered in the Practicum Results section. Methodology Part 3: IOM Environmental Impact 4. Structured observation : As part of the environmental impact assessment, I undertook direct observation at a sample of project sites. I randomly selected sites from a list generated by the IOM project database of projects implement in Petit Goave and Jacmel. These two regions provided a longitudinal component to the evaluation, as watersh ed management activities were started in Petit Goave in 2006 and in Jacmel in 2012. Comparing intervention sites at different stages of implementation offered insights into the impact and sustainability of activities over time. It should be noted that I did not assess any control sites due to time and methodological constraints. Not having this


33 point of comparison limits what can be inferred from assessing the environmental state of intervention sites. A checklist comprised of indicators of environmenta l health guided f ield visit observations see Table 7 Table 7 Field visit observation checklist Variable Measure Notes Tree spacing Distance (m) between trees Average measurement taken at randomly selected hedgerow Tree cover/survival % trees remaining in hedgerow Average count taken at randomly selected hedgerow Tree diameter at breast height* Diameter (m) Average measurement of 10 trees randomly selected from stratified sections Basel area* Prism wedge count Function of coverage and diameter Tree predation Scale (1 5 ) Signs of livestock predation. Scale developed for site comparison purposes only not generalizable Soil accumulation Volume (cm 2 ) Average v olume of top soil accumulated at randomly selected barrier s Vegetation cover Scale (%) Cover in a 5m 2 randomly selected sample Quality of construction of erosion barriers Scale (1 5) Scale developed for site comparison purposes only using best practice example Durability of erosion barriers Scale (1 5) Scale developed for site comparison purposes only using best practice example Stream discharge * Continuous (m 3 /second) Area = Depth x Width Discharge = Area x Velocity Stream bank width * Continuous (m) Measurement taken at widest point of stream Distance of water from stream bank ** Continuous (m) Measurement taken at widest point of stream Measurements taken where applicable most trees were too thin to register ** Measurements of stream fluctuation not used, potential for further study of impact Practicum Results To understand the best practices of these two tree planting programs, I evaluated activities and documented programs as part of a comparative study of their impact and application a cross different contexts. The MCC agroforestry approach was studied in order to produce a program


34 manual of their best practices. Both programs were evaluated on the basis of their environmental impact by measuring tree survival rate. IOM was evaluated in terms of other indicators of environmental impact and evaluated in terms of their KAP impact. The results of this comparative tree plannting study contribute to our understanding of what works where in Haiti. Part 1: IOM and MCC Evaluation Results O verview Evaluation Process: I relied on several data collection and analysis methods to derive the evaluation results I employed information collected through the documentary review and interviews of key informants to select evaluation sites and design the survey and observation guide. Together, these collection methods provided a measure of the sustainability, as well as its application in different contexts. This evaluation process is described in Table 7 and illustrated in Figure 2 Table 7 Data collection and analysis methods Collection Method Participants/ Sources Information/Results Provided (Organization) Documentary Review Program reports, program manual, IOM project database Intervention site selection (IOM) Control community selection (IOM) Cost efficiency (IOM) Open ended interview K ey informants Profile of success (IOM and MCC) Program manual content (MCC) Oral survey Heads of household control and intervention Magnitude of knowledge, attitude and behavior change (IOM) Structured o bservation Principal investigator Tree survival rate (IOM and MCC) Erosion barrier effectiveness (IOM) Erosion barrier durability (IOM) Community maintenance (IOM) Data analysis* Principal investigator Program impact, sustainability and application (IOM and MCC) A review of overall evaluation results used to draw conclusions and lessons learned.


35 Figure 2 : Evaluation process In this illustration, the different data collection and analysis methods employed in the evaluation are organized by the two objectives of t he watershed management program. The grey arrow represents the KAP impact evaluation and the white arrow represents the environmental impact evaluation. These methods progress from left to right, where each method contributes to the completion of the next method. I n order to complete the KAP impact component ( grey arrow) a documentary review was done to make a site and community selection and to inform the design of survey Once these steps were completed, data collection, data analysis, conclusions and lessons learned proceeded. The white arrow documents the process of design ing the observation guide for the environmental impact component of the eva luation Practicum Results Part 2: MCC Program Manual I compiled the MCC program manual by undertaking a documentary review of program records, interviews with key informants and by shadowing MCC staff at 20 community processes by which trees are produced, distributed and managed through the different components of the agroforestry program. The manual is intended to be instructional: it describes program activities and discusses how they might be applied in other settings in order to make the approach accessible to partners. The MCC program manual can be accessed in Attachment 3.


36 Practicum Results Part 3: IOM and MCC Program Comparison Tree survival rate: I measured survival by sampling field sites and comparing the number of trees remaini ng against the number of trees planted. In the case of IOM, I randomly selected planting sites from a list of projects in the region. Trees were planted at regular intervals and missing trees were easily identified. For MCC, I randomly selected planting sites from attendance sheets compiled during tree distributions. The measurement relied on participants to recall what was planted in previous years which invited the potential for recall error However, t tree survival. 36 Results from these studies are compared with the minimum tree survival rate of 30% after three years established by the Ministry of Environment for government endorsed activities 37 See Figure 3 Three years is considered the time necessary for trees to become established, after which they are less vulnerable to predation, drought and disease. While 30% appears to be a modest goal at face value, it accurately reflects the harsh prospects for trees in Haiti observed during the evalu ation. Figure 3 Tree survival by program and by species type after three years 36 MCC 2011 37 Ministry of Environment 2013 44% 66% 18% 65% 72% 25% 30% 0% 10% 20% 30% 40% 50% 60% 70% 80% Watershed Mgmt IOM Agroforestry MCC Ministry of Environment Forestry Overall Fruit Overall


37 Tree Survival Rates: Discussion Survival varied greatly across species, with forestry species surviving at much higher rates than fruit species. Overall survival rates far exceeded the minimum rate set by the Ministry of Environment for Government of Haiti supported tree planting programs. The overall rate was buoyed by a handful of forestry species, namely eucalyptus ( Eucalyptus cinerea ), cassia ( Senna siamea ), casuari nas ( Casuarina cristata ) and fwnn ( Simarouba glauca ). Species less adapted to marginal soil and drought conditions fared much worse; with those that did survive showing signs of stunting. These highlighted species are also the most resistant to predatio n. In contrast, fruit species appear to be the favored forage for livestock. Most of the fruit trees that remained showed signs of grazing, contributing to their much lower survival rate : 25% of MCC trees, 18% of IOM trees Wind is another environment al threat which must factor in species selection. At one exposed mountain site, only casuarinas showed sustained growth due its needle shaped leaves. The leaves of most other species at this site had been shredded by the wind with only eucalyptus showin g signs of resistance This consideration extends to the grass species used in hedgerows as well. Vetiver ( Chrysopogon zizanioides ) fared dramatically better in drought conditions and showed fewer signs of predation than did gautemala ( Tripsacum laxum ) o r elephant grass ( Pennisetum purpureum ). The timing of planting is also critical to survival, as trees planted during the dry season fared much worse, even with the benefit of hand irrigation. Cost efficiency : The following cost per unit measurements were calculated using inputs and outputs tracked in the IOM project database. 38 These measurements do not include overhead, staff and office costs. They do however include the labor costs to implement 38 Further analysis might compare cost efficiency with the effectiveness of interventions in order to provide an estimate of cost effectiveness One of the challenges of such analysis is separating these different interventions, as they are mutually reinfor cing (e.g. contour canals planted with trees form hedgerows).


38 each activity. A discussion of MCC activity costs foll ows. As their activities differ, they are not directly comparable. While both organizations plant trees, IOM pays participants to plant trees, MCC does not. This payment is factored into the following cost efficiency calculations for the I OM watershed m anagement program in the Petit Goave and Jacmel regional offices. Calculations from other IOM regional offices are drawn in the discussion that follows. Table 9 Cost efficiency by IOM activity Indicator Unit Cost Per Unit (US $) Trees planted Tree 2.24* Vetiver grass planted Plug 4.68** Stone walls constructed Meter 16.40*** Contour canals constructed Meter 6.20 Short term jobs created 2 week rotation 62.50 Employment days created One day 6.25**** Fruit and forestry tree species disaggregated species data not available ** Vetiver grass purchased from private nursery more expensive to plant than trees *** T aken from a limited sample size: s tone walls are constructed by IOM in regions where there is a greater supply of granite for construction, primarily Gonaives and Saint Marc. **** The government mandated minimum wage Watershed management projects are extremely labor intensive. Since there are no construction materials required, wage labor is the principal cost In this way, IOM pursues the goals of both watersh ed management and job creation by paying participants to construct erosion barriers and plant trees. The IOM offices in Petit Gove and Jacmel are able to plant trees at a much lower cost than the other IOM offices engaged in watershed management. The average cost per tree for the other four offices is $4.14. This can be attributed to the practice of subsidizing the production of trees at community run nurseries as opposed to purchasing trees employed b y the other offices. Trees are outplanted at a minimum of 6 9 months after germination. MCC also subsidizes tree production, but does not pay participants to plant trees,


39 nor do they pay to transport the trees to the planting site As a result they only take responsibility for the price of producing trees at unit price of $0.15 and $0.35 In addition, MCC does not engage in the other activities of planting grass or constructing erosion barriers. Consequently, comparisons of cost efficiency between the t wo organizations are problematic. Practicum Results Part 4 : IOM KAP Impact S urvey Results : I administered a n oral watershed management program. I asked program participants about their satisfaction with the program. Participant Satisfaction : The survey polled members from the intervention group (n=72) on their perceptions of the program: how well the project met the expectations of the communit y, as well as their own personal satisfaction with the program. The wording of the two questions was intentionally similar in order to gauge to what extent program participation contributed to community cohesion and collective action. Roughly half of par ticipants responded positively to both questions. Several participants commented and the same. A minority of respondents answered that the program failed to meet the expectations of the community (12%) or felt personally unsatisfied with the program (15%). Follow up questions revealed that personal dissatisfaction was commonly associated with not working more shifts on the programs, and not necessarily a reflection on the quality of the project. These res ults are illustrated in Figure 4


40 Figure 4 Summary of selected survey results intervention group only How well did the program meet the expectations of the community? How satisfied are you personally with the program? Knowledge, Attitude and Behavior Impact : I compared the responses of the intervention group to the responses of the control group to estimate the magnitude of change that can be attributed to project participation in Tables 10 I organized the r esults by the KAP impact measurements in my evaluation plan. I calculated the statistical significance of each of these results using chi square tests. A discussion of the composition of respondent group s and the potential confounding factors foll ows these results Table 10 Survey results by KAP impact indicator Key : Q# = Question number Int = Interve ntion Ctrl = Control Sig = S tatistical s ignificance P Value categories : Not significant = 0.05alpha>0.05; significant = 0.01

41 2. Change in attitude toward the environment Q# Indicator Int Ctrl P Value Sig 1 Identify environmental problems (in top 3) 36% 38% 0.841 Not 2 Prioritize environmental interventions (in top 3) 72% 26% < 0.00 High 7 Describe trees as "important" or "very important" 92% 84% 0.211 Not 3. Change in sense of responsibility toward the environment Q# Indicator Int Ctrl P Value Sig 15 Claim responsibility to plant trees 60% 51% 0.382 Not 16 Claim responsibility to manage trees 68% 42% 0.010 High 11 Agree to plant trees voluntarily (w/out payment) 24% 44% 0.036 Sig 4. Accrual of knowledge of best land management practices Q# Indicator Int Ctrl P Value Sig 9 Identification of 4 or more services provided by trees 68% 66% 0.837 Not 21 Identification of 3 or more soil conservation methods 40% 33% 0.487 Not 5. Change in sense of agency/capacity in environmental management Q# Indicator Int Ctrl P Value Sig 17 Feel capable of planting trees 76% 52% 0.013 Sig 19 Plan on planting more trees 70% 46% 0.017 Sig 6. Demonstration of new practices on private plots Q# Indicator Int Ctrl P Value Sig 22 Practice 3 or more soil conservation methods 40% 33% 0.487 Not 18 Planted trees on property in last year 27% 34% 0.459 Not


42 Figure 5 Comparison of selected survey results : intervention and control Survey Results: Discussion These results suggest that the IOM watershed management program was successful in changing environmental knowledge and attitudes, but ultimately unsuccessful in changing tree planting behavio r. Compared with the control group, program participants responded favorably across all measurements of behavior change except for Demonstration of new practices on pri vate plots Change in sense of responsibility toward the environment Participants were less likely to agree to voluntarily plant trees (24%) than were members of the control group (4 4%). This resulted in participants planting fewer trees on their property in the last year (27% v. 34%). Raising awareness and changing attitudes towards the environment is clearly not enough if it does not translate into voluntarily planting trees. It is impractical and unsustainable to expect external dono rs to pay to produce and plant every last tree; so voluntary planting is essential to reforestation in Haiti. While the difference in actual tree planting behavior between these groups is relatively small (7%), the rate itself is still low for both groups. Essentially only one in three, or less, of respondents planted trees in the last year despite resid ing in heavily deforested areas Participants were more likely to respond that they are plan o n planting 0% 10% 20% 30% 40% 50% 60% 70% 80% Identify impacts of deforestation Demonstrate understanding of role of trees Prioritize environmental interventions Claim responsibility for planting trees Independently plant trees Intervention Control


43 more trees in the future (70% v. 46%) but the fact that they have not done so already raises questions about their willingness to follow through. Since the program actively promoted tree planting, it is possible that participants are more like ly to state their intent to plant trees in order to appear to conform to the program norms in the hope of securing future employment or benefits More research is required to understand the barriers to this c r itical step of behavior change. The greatest differences in responses between groups were the intervention ability to identify the consequences of deforestation (66% v. 42%), to p rioritize environmental interventions (72% v. 26%), to claim responsibility to manage trees (68% v. 42%), feel capab le of plant ing trees (76% v. 52%) and as noted previously, to plan on planting trees (70% v. 46%). In contrast, the intervention group were much less likely to a gre e to plant trees with out payment (24% v. 44% ) This raises concerns about 1) the impa ct of compensation on tree planting attitudes and behavior; and 2) the validity of the pro environmental responses expressed by the intervention group if they did not translate to willingness to plant trees as discussed under the following confounding fa ctors. Survey Results: Confounding factors Composition d ifferences between groups: The most pronounced differences between respondents in the intervention and control groups were found in age literacy primary livelihood and to a lesser degree, estimated total land size The rationale for each of these differences, and their potential influence on the survey results are considered in the following discussion: Age: M embers of the control group were younger on average (36 years) then the intervention group (41 years) The IOM program staff attributed the difference in age to two factors: o C ommunities tend to prioritize older residents for employment in Cash for Work activities leading to a higher average age in the intervention group o Many of the younger m embers of these rural communities are absent


44 during the academic year as they attend secondary school in urban centers. As a result, they are not employed by the program. These students return home for the summer holidays when the survey was conducted. This may explain why the control group is skewed towards a younger demographic. The greater proportion of students in the control group substantiates this explanation. Literacy : Respondents in the control group had a higher rate of literacy (66%) than those in the intervention group (48%). Both the younger average age and greater proportion of students of this group likely contribute to this difference. Since the environment is increasingly a part of the national education curriculum in Haiti, 39 one w ould associate literacy with greater awareness of environmental concerns. By extension, the control group would be more likely to begin with more knowledge of deforestation than their intervention counterparts. This would only bolster the findings that the program contributed to positive changes in environmental attitudes and knowledge, since the intervention group scored better on these sections of the survey despite any disadvantage incurred by lower literacy rates. Primary livelihood : In addition to being comprised of more students (18% v. 8%), the control group also had a higher proportion of merchants (35% v. 25%). As both of these occupations are associated with travel, it is possible these respondents were not present in the community to take advantage of program employment. This could bias the results of the survey by measuring knowledge, attitudes and behavior accrued outside of the matching control community (and therefore, not representative of the general population). Combined, these two professions account for 53% of control respondents, as compared to 33% of intervention respondents. Further analysis is needed to determine whether this difference in composition between the two groups impacts their survey responses. 39 Fournier 2013


45 Estimated total la nd size : The intervention group (1.3 ha) had slightly more land per capita than the control group (1.1 ha). This may be a reflection of the younger average age of the control respondents as well, since younger respondents have had less time to amass land holdings. Land ownership was primarily gained through inheritance in both groups. Younger respondents are more likely to be waiting for their inheritance to be passed down, and thus have fewer holdings. Data analysis of respondents from both groups ind icates that respondents with greater land holdings (2 ha or more) are 12% more likely to plant trees on their property than those with less land. This finding is corroborated by studies of the determinants of tree planting in Haiti done by Balzano 40 and Ba nnister and Nair. 41 However, the control group was slightly more likely (34% v. 27%) to have planted trees in the last year than the intervention group. Cognitive bias : Survey respondents from both the intervention and control groups were subject to a degree of response bias. Having participated in IOM program activities, objectives. In ad dition, the team administering the survey with me was comprised of past employees of the organization. It is likely that these associations influenced some participants to formulate their responses according to what they believed the organization wanted t hem to answer, rather than according to their own beliefs. This bias may have been true of the control group as well. The majority of control respondents were at least familiar with the program. The control group had the added motivation of lobbying for the organization to implement a project in their own community. Finally, this bias most likely worked both ways for intervention group as a way of voicing their dissatisfac tion with the program. To limit the extent of this bias, I piloted questions with members of both groups before 40 Balzano 1986 41 Bannister and Nair 2003


46 finalizing the survey. In addition the survey team presented the survey as a stand alone activity, not connected with future program prospects It is difficult to say whether this bias was stronger among intervention or control group participants. Externalities of membership: In order to approximate the conditions found in the intervention commun ities, I selected control communities from the same commune This proximity raises the possibility that the control group was subject to spillover effects through interactions with the intervention group. Further confounding the comparison of groups was the presence of other environmental int erventions in the region. In order to limit the influence of these spillover effects, I selected control groups from communities with no active interventions. Practicum Results Part 5 : IOM Environmental Impact Structured observation Results : The watershed management program produced results on tree survival, community maintenance and erosion barrier quality, durability and effectiveness. Of these results, only tree survival was comparable between the IOM a nd MCC programs. The remaining results are presented here. Tree diameter at breast height : Due to the limited diameter of trees, this measurement was not considered applicable to the evaluation. Project trees ranged from six months to seven years old and many were stunted in their growth. This measurement will be more appropriate in the future. Community maintenance : At the conclusion of each watershed management project, the host community signs a maintenance cont ract with IOM In this contract the community pledges to maintain the project structures for a minimum of 10 years. The extent to which the community is in compliance with this contract is assessed by a series of yes/no ovided an assessment of maintenance to compar e di fferent sites see Table 11 Continued


47 maintenance of sites is part of the criteria IOM uses when deciding to implement follow up activities in communities. Table 11 Community maintenance record at field sites Field Site Commune Maintenance committee active? (Y/N) Livestock tied or corralled? (Y/N) Livestock predation? (1 5 scale) Ban on root crops respected? (Y/N) Erosion barriers maintained? (Y/N) Ternier Jacmel Y Y 2 Y Y Laviale Jacmel Y Y 2 Y Y Durivier Jacmel Y N/A 1 Y Y Verger Petit Gove Y N/A 3 Y N Vallue Petit Gove Y Y 1 Y N Nan Bazil Petit Gove N N/A 3 Y N Haute Courtis Petit Gove N N 3 Y N Courtis Petit Gove N N 5 Y N Regale Petit Gove Y Y 1 Y N 2eme Plaine Petit Gove N N 3 Y N Total 6/10 4/10 Avg = 2.4 10/10 3/10 Communi ty M aintenance : Discussion Maintenance committee active : I observed various levels of activity across committees. Some committees met regularly, performed routine site visits and intervened when committee for the purpose of the evaluation. Livestock predation: I measured predation was on a 1 5 scale based on observed signs of livestock damage to trees. Sites with active maintenance committees and livestock management had lower rates of predation. However, it is difficult to deduct if this relationship is caus al since some areas simply had more livestock


48 than others (and the role of the maintenance committee in managing them differed from community to community). Erosion barriers maintained: I observed that barriers were maintained at all of the field sites in Jacmel and none of the sites in Petit Goave. Several factors may explain the difference in maintenance between these two regions: o IOM activities are still ongoing in Jacmel and may serve as motivation for the community to comply with maintenance requireme nts. o The sites at Petit Goave are considerably older than those in Jacmel. Thes community may have initially maintained these sites, but has since ceased maintenance activities. Erosion Barrier Assessment: As part of the structured observation, I assessed barriers at three points of each field site: the top, middle and bottom of the slope. For each point I randomly selected a barrier and evaluated it on three criteria: quality of construction, durability and effectiveness. The measurements take n from the three points were averaged to provide an overall site measurement for these criteria. Quality of construction of erosion barriers : measures the extent to which specifications were followed in the construction of barriers on a 1 5 scale see Ta ble 12 These specifications include whether the barrier is perpendicular to the slope, maintains a level plane respects standard dimensions and is correctly spaced in relation to other barriers.


49 Table 12 Quality of erosion barrier construction by site Field Site Commune Type of Erosion Barrier Quality Rating (1 5) Ternier North Jacmel contour canal 4.2 Ternier South Jacmel contour canal 4.5 Laviale Jacmel contour canal 4.0 Durivier East Jacmel contour canal 4.1 Durivier West Jacmel contour canal 3.8 Verger Petit Gove contour canal and stone wall 3.2 Vallue Petit Gove contour canal 3.4 Vallue Northeast Petit Gove contour canal 3.6 Nan Bazil East Petit Gove contour canal 2.6 Nan Bazil West Petit Gove contour canal 3.1 Haute Courtis Petit Gove contour canal and stone wall 3.7 Courtis Petit Gove contour canal and stone wall 3.7 Regale Petit Gove contour canal 2.9 Deuxieme Plaine Petit Gove contour canal 3.5 Total and Average 14 communities, 2 communes 3.6 Durability of erosion b arrier s : is a measurement of the state of contour canals over time. Durability depends not only the quality of construction, but also the type of soil. However canals are expected to maintain their form for at least two years. As they deteriorate canals become less effective at capturing runoff and sediment, and can even contribute to erosion if not repaired. A scale of 1 5 based on the uniformity of the barrier was employed see Table 13 In practice, quality and durability wer e difficult to distinguish from one another; as such, these measurements were used as a composite assessment to


50 identify problematic sites for follow up by the organization. Table 13 Durability of erosion barriers by site Field Site Commune Type of Erosion Barrier Durability Rating (1 5) Ternier North Jacmel contour canal 4.8 Ternier South Jacmel contour canal 4.6 Laviale Jacmel contour canal 4.6 Durivier East Jacmel contour canal 4.8 Durivier West Jacmel contour canal 4.5 Verger Petit Gove contour canal and stone wall 3 .8 Vallue Petit Gove contour canal 3.2 Vallue Northeast Petit Gove contour canal 3.0 Nan Bazil East Petit Gove contour canal 3.1 Nan Bazil West Petit Gove contour canal 2.7 Haute Courtis Petit Gove contour canal and stone wall 2.4 Courtis Petit Gove contour canal and stone wall 2.3 Regale Petit Gove contour canal 2.5 Deuxieme Plaine Petit Gove contour canal 4.1 Total and Average 14 communities, 2 communes 3.5 Erosion b arrier effectiveness: is a measurement of the depth of soil accumulated in each canals are intended to capture rain and soil runoff. The depth of soil in the canal is a proximate measure of the e rosion prevented at the field site. There is expected to be less soil accumulation further down the slope, since the soil runoff accumulates in the higher canals. Taking a stratified sample of canals at the top, middle and bottom of the slope is therefore considered more representative of global conditions at the field site.


51 I measured soil accumulation by excavating loose soil in the contour canal until the original soil horizon was visible and the depth of accumulation could be measured. I then average d the measurements for the three contour canals to provide a rate of accumulation that could be applied to all canals at the site 42 see Table 14 Table 14 Soil accumulation at sampled field sites Field Site Commune Soil Accumulation Rate (m 2 ) Projected length of contour canals (m)* Total Estimated Soil Accumulation (m 3 ) Ternier North Jacmel 0.30 1,200 360 Ternier South Jacmel 0. 26 600 156 Laviale Jacmel 0.22 2,100 462 Durivier East Jacmel 0. 24 400 96 Durivier West Jacmel 0. 16 900 144 Verger Petit Gove 0.42 6,500 2,730 Vallue Petit Gove 0. 36 3,800 1,368 Vallue Northeast Petit Gove 0. 34 2,200 748 Nan Bazil East Petit Gove 0. 38 1,600 608 Nan Bazil West Petit Gove 0.44 1,600 704 Haute Courtis Petit Gove 0. 46 2,500 1,150 Courtis Petit Gove 0. 28 1,400 392 Regale Petit Gove 0. 30 1,000 300 Deuxieme Plaine Petit Gove 0. 33 1,800 594 Total 14 communities, 2 communes 27,600 9,812 Projecti ons from IOM project database a ctual measurements of total canal length not 42 Soil accumulation depth is an average of three readings in each canal left side, middle and right side. Soil depth is multiplied by the width and length of the canal to provide an estimate of accumulation volume. This calculation assumes a standard can al width and depth of 0.1m x 0.2m.


52 possible due to time constraints Erosion Barrier Assessment: Confounding Factors Since conditions vary widely, erosion barrier effectiveness is difficult to compare between field sites. Factors such as slope, precipitation or soil composition all impact soil accumu lation in addition to the quality of construction and/or quantity of barriers T he age of sites also affects this measurement, as older sites have had more time to accumulate soil runoff. To some extent, all three of the erosion barrier evaluation criter ia quality durability and effectiveness are a function of time. Activities in Jacmel started in 2012, while Petit Gove has been active since 2006. Sites in Jacmel scored better on quality ( average of 4.1 vs. 3.3) and durability (average of 4.7 vs. 3.0) than those in Petit Goave. This trend also reflects the difficulty to separate measurements of quality from durability. Higher quality barrier construction contributes to longer barrier durability, but other factors can also inf luence durability and confound comparison across sites. Another confounding factor in barrier assessment is the effect of maintenance on barrier durability and soil accumulation levels. While all communities are instructed to clear out canals that have filled in with soil, few communities take this initiative after the conclusion of the project. The fact that maintenan ce is inconsistent across sites poses another c hallenge to making an accurate assessment The principal value of these measurement s is t o provide a means of capturing some of the environmental impacts of for all IOM sites and presented to donors as part of a funding proposal to continue activities.


53 Program Analysis The primary means of implementing tree planting in Haiti is currently Cash for Work activities (as employed by IOM) While these meet the immediate goal of injecting cash into households and markets, they can undermine community participa tion and ownership. Paying beneficiaries creates the expectation that planting trees is wage labor and not an activity inherently beneficial in and of itself. If designed incorrectly Cash for Work schemes can discourage independent initiative and erode volunteerism. This can be seen in other sectors as well, in the wake of Cash for Work programming after the earthquake. The pervasiveness of these activities around the country has created a difficult precedent for future programs to follow. This may ex plain the evaluation finding that IOM Cash for Work participants were less likely to agree to plant trees without payment than members of the control group. Despite these drawbacks, paid labor is all but essential in addressing deforestation on kin owne d la nds and remote upper catchments. In contrast to these conditions, MCC agroforestry participants primarily plant trees on land for which they are the sole owner, that is located closer to their home and that is flat to moderately sloped. The decision to plant trees in this fashion reflects the land tenure security and acc essibility (or ability to monito r) of these holdings. This decision making criteria is mirrored in previous studies of agroforestry programs in Haiti Murray and Bannister go on to identify the threats to distant plots or (1) free ranging livestock, (2) nocturnal 43 These threats are all present to varying degrees on aid labor is also necessary to motivate people to plant trees that they will not be able to harvest for economic benefit. The trees planted by IOM Cash for Work participants are th e property of the landowner(s). Short term employment also meets an important need for cash in rural communities. 43 Murray and Bannister 2004: 389


54 survey participants. IOM has tried to strike a balance between paid and voluntary labor by requiring participants to contribute one employed the Haitian konbit practice of organizing labor by preparing food for participants. While cheaper and perhaps more participatory than Cash for Work this traditional method has limitations when scaling up interventions. Both organizations currently subsidize the production of trees at community run nurseries by paying participants a small amount per tree produced. IOM pays nursery workers between $0.40 and $0.60 per tree, while MCC pays workers $0.15 p er forestry tree and $0.35 per fruit tree produced. The long term outcome of current environmental programs depends on the extent to which they achieve stakeholder participation and ownership. As demonstrated by previous reforestation efforts in Haiti i t is easier to plant trees than it is to protect and maintain them. Both organizations invest considerable effort in mobilizing communities through planning meetings. This participation is sustained by continued visits, capacity trainings and recruitment of residents to superv isory positions. However, where private funding allows them to make an open ended commitment to communities, IOM is limited by the deadlines of public donors specifically USAID and AECID This is reflected in the higher rat es of survival of trees planted through agroforestry program ( overall rate of 65% by MCC vs. 44% by IOM) The success of the MCC program of expanding to 22 communities and planting 7.6 million trees is due in large part to its continuity and longevity. The program has remarkably continued its services through periods of immense social u pheaval and economic hardship over the last 20+ years. For many Haitian farmers, environmental considerations are out of necessity secondary to immediate econ omic needs. However, the two are not mutually exclusive. Under the right conditions, trees can help meet both objectives. Resource ecologist Joel Timyan allows li ttle consideration for trees without immediate value to the household economy.


55 its elimination, many times in favor of other agricultural activities, mainly grazing a nd 44 IOM promotes trees foremost for their environmental benefits, and not necessarily their trees. Fruit bearing species comprise less than 15% of trees planted. IOM also plants trees in the upper catchments to maximize their environmental services based on the logic that flooding and erosion starts upstream The marginal soil on these steep, eroded slopes rules out planting more frui t or commercial timber species. At the same time, the remote location of these sites poses challenges for protecting plantations as well as harvesting and transporting tree products. Finally, t he rigid deadlines under which IOM operates restrict the up activities necessary to ensure higher rates of survival of the trees planted. In contrast, MCC distributes trees to farmers during the rainy season. Farmers tend to plant trees on more fertile valley lands that are both closer to their home and under greater supervision. The s e different conditions complicate comparisons of tree survival rate between the two programs. for age, fruit, timber, firewood and charcoal. While charcoal has been widely vilified as a driver of deforestation, it has also played an instrumental role in reforesting Haiti. 45 As the size of agricultural plots shrinks in Haiti, trees are increasingly in competition with crops for available land. Charcoal provides farmers with the economic motivation to maintain trees. It is a low tech, value added product that can be easily transported to urban markets. While it has a lower market value than timber, ch arcoal can be produced much sooner. MCC has championed charcoal and timber production through its agroforestry program, mobilizing farmers to voluntarily plant trees. These farmers times of need. Trees also serve as insurance against crop failure, sickness, injury or other 44 Timyan 1996: xi 45 Lea 1995: 42


56 emerg encies. 46 These benefits have resulted in higher school enrollment rates, better housing materials and more livestock in beneficiary communities. Survey participants identified the time required for trees to mature for harvest as one of the main barriers to planting trees. IOM works to address this concern by promoting backyard gardens to provide income and nutrition in the interim. Participants are supplied with materials and trained in permaculture methods to intensively cultivate these small plots. Since the Haitian lakou is traditionally separate from la rger farmlands these backyard gardens do not compete for space with trees or other crops. Their proximity to the house also makes maintaining gardens less labor intensive. Lessons Learned Agrofore stry and reforestation both play an important role in addressing deforestation in Haiti, though neither is sufficient in itself: planting trees for harvest on private lands neglects some of the most vulnerable sections of watersheds; at the same time, wate rshed management does not provide the same long term economic incentives for farmers to integrate trees into their production systems. Therefore a comprehensive reforestation strategy might draw on the most promising elements of both of these approaches. The MCC and IOM programs represent successful examples of each approach. Studying these programs provides us with important lessons learned in designing future tree planting efforts in Haiti. These lessons detailed below apply to community nurseries, voluntary vs. paid labor, appropriate species selection, diversified funding, sustainability vs. impact tradeoffs and the value of separate approaches. 1. Community Nurseries: Decentralizing seedling production from professionally run to community run operat ions can reduce the cost of production and transport while still delivering high quality seedlings. In fact, several of the communities 46 Murrary 1987


57 the trees on site in community nurserie s eliminated the cost of transportatio n and its potential strain on seedlings. Engaging communities to produce their own seedlings builds the capacity of participants to do so and ensures that preferred species are produced (as seeds are available) The benefits of capacity building extend to the rest of the community through extension visits conducted by community nursery committees. In order to maintain high quality production standards, nursery committees must be trained, monitored and paid by an exte rnal party In that regard, IOM and MCC demonstrate the most cost efficient means of paying participants is by providing the production based incentive of a per tree subsidy. There are challenges inherent in this arrangement. The amount IOM and MCC pay nursery committees for production is not enough for their members to work exclusively at nurseries. In some cases, this has led to the neglect of nursery responsibilities in favor of other income generating activities. Additionally, requiring nurseries t o collect or purchase seeds can create shortfalls in desired species. While these measures may make nurseries more financially independent, they create a trade off in terms of quality and quantity of seedlings. This disparity between community and profes sional nurseries was observed by IOM staff working in different regions. The professional nurseries were reported by IOM to produce more trees with slightly higher rates of survival 47 which confirms findings from previous studies. 48 Overall however, the MCC and IOM nurseries I observed during the study met production standards: the majority of nurseries produced or exceeded the agreed upon number of seedlings. These seedlings were healthy and mature at the time of outplanting. In su m, professionally run nurseries may provide more consistent results, but at higher operating costs. This finding is born out in the cost 47 Fournier 2013 48 Murray and Bannister 2004: 394


58 comparison of tree planting across IOM regional offices detailed in the Practicum Results section. 2. Voluntary Labor: MCC has mobilized farmers in 22 communities to voluntarily plant trees on their property through its agroforestry program. This success can be attributed to a number of factors: providing large quantities of free seedlings through annual nursery distribut ions; r ecognizing communities should control the type of tree species produced ; guaranteeing tree tenure and harvest rights to participants (producers have full permission to manage trees as they see fit); demonstrating the economic value of planting trees through planning meetings; providing technical training to plant and manage tree stands; and most recently, introducing silviculture practices that support tree production alongside traditional crops. Each MCC nursery hosts an a nnual tree distribution in the associated community. The distribution coincides with the rainy season to give trees the best chance of survival. The date is widely publicized in the community to attract as many participants as possible. MCC staff demonstrate correct planting pro cedures at the time of distribution, and conduct follow up household visits to ensure they are followed. In this way, the program is able to mobilize farmers to plant some 450,000 trees annually on a minimal budget. The MCC agroforestry program budget co nsists of the initial start up materials for nurseries, the subsidies paid for producing seedlings and the salaries of program staff. IOM has developed a mixed system of paid and voluntary labor that requires Cash for Work participants to contribute one day of unpaid labor per week essentially being paid five days for six days work. While this system accomplishes more work for the same amount of money, the primary purpose of volunteer days is to create a sense of persona l investment in the project among participants. This sense of encourage beneficiaries to protect the trees.


59 3. Cash for Work : As indicated earlier, it is necessary to compensate participants for planting trees that they themselves will not be able to harvest. This has been accomplished by supplying food stocks, paying cash wages and providing services to participants in exchange for their labor Putting cash in the hands of par ticipants empowers them to make their own decisions on how to meet their household needs. Cash for Work tree planting meets the immediate demand for wage labor in rural communities with few opportunities for employment while accomplishing long term ecolog ical objectives (that would otherwise be impossible without such collective action). IOM has developed a rotation system to provide the opportunity for employment to as many community members as possible. Each participant works for two weeks (including t he two volunteer days) during their rotation on the project. They are paid the national minimum wage of 250 HTG, or approximately, $6.25 per day. While this may not amount to very much money, the households interviewed were still able to use their income to make long term investments such as paying school fees, acquiring livestock and purchasing merchandise for resale, in addition to meeting their daily needs for nutrition and shelter. 4. Appropriate Species Selection: sites in Petit Gove revealed the importance of selecting tree and grass species adapted to site conditions. Compared to other tree planting programs, the overall survival rate of t rees was relatively high at 42%. However, this figure was buoyed by a handful of forestry species The common characteristics of these key species are drought resistance, tolerance to poor soil, bitter leaves (to discourage predation) and in the case of sites exposed to wind, a narrow leaf shape The fruit tree species employed in the program have none of these characteristics, and as a result survived at a much lower rate of 18%. This consideration extends to the grass spe cies used in hedgerows as well.


60 Given this experience, the selection of tree and grass species may be the single most important indicator of survival at reforestation sites. The timing of planting is also critical to survival, as trees and grasses planted during the dry season fared much worse, even with the benefit of han d irrigation. long term resiliency. Another reforestation program, implemented by the Presbyterian Church over the last 20 years in Leogne has attempted to address this ch allenge by planting tree species in succession. The mission plants eucalyptus as a frontline to stabilize soil and reduce wind, followed by nitrogen fixing species such as leucaena ( Leucaena leucocephala ), before finally planting fruit trees. These fruit trees are often accompanied by agriculture as soil fertility is restored The trees planted in stages one and two are thinned to clear space once they have fulfilled their respective roles. The resulting landscape encompasses much more diverse species. 49 5. Diversified Long Term Funding: The priorities and means of donors change frequently. Competent programs close due to funding shortfalls before they have achieved their objectives and/or handed over implementation to the community USAID funded Agroforestry Outreach Program (AOP) referenced in this paper as one of the few success stories in tree planting is one such example. Funding for the AOP was not renewed after 20 years of implementation unprecedented success. The longevity of the AOP raises questions about the long term role of external interventions. At the same time, the abrupt end of the program also underscores the n eed for an exit strategy by both donors and implementing partners. From the accounts of members of the AOP, funding for the 49 CODEP 2013


61 AOP stopped before the program could fully capitalize on the economy of scale it had built to produce and distribute trees. 50 MCC was one of the original NGOs contracted under the AOP by one of the implementing partner s PADF. While the other local NGOs contracted by PADF and CARE closed or changed course after the AOP concluded, MCC continued to execute its agroforestry activit ies. MCC was able to secure private funding for operations through international fundraising. This example demonstrates the continuity Such institutional sustainability is critical in promoting the beha vioral changes implicit in the 51 In recent years, MCC has initiated a handover process by transitioning from international staff to a smaller team of national staff, decentralizing tree production to community run nurseries, and in the process, building the capacity of communities to produce and plant trees using locally available materials. 6. Sustainability vs. Impact: The pursuit of financial sustainability has resulted in p ositive and negative outcomes. MCC developed a demand for tree seedlings through many years of free distribution. In recent years, MCC has started to capitalize on this d emand by reducing the number of trees distributed and encouraging participants to purchase trees. The proceeds from tree sales go directly to sustaining tree production as part of a long term handover strategy to make nurseries financially autonomous. Th is shift has had some success in raising funds, but it has come at the cost of participants taking home fewer trees to plant. At the same time, MCC still subsidizes 95% of the trees produced. Anthropologist Gerald Murray, who roject in which peasants pay for 10% of the seedlings 50 See Murray and Bannister 2004 for a discussion of the circumstances of 51 Hildebrand 2013


62 is no more sustainable than one in which the project simply donates them What you do by imposing a per seedling cost is reduce by 90% the number of farmers 52 Since the trees ser ve secondary ecological and public functions, no single farmer should be required to pay the cost of producing them. Instead, they should be widely subsidized by governments and other donors as they are in many other countries, including the United Stat es. 7. Separate Approaches: A final lesson learned in studying the IOM and MCC programs, is that agroforestry and reforestation are viable but distinct approaches to tree planting. While their objectives and impacts overlap at certain levels, implementing both activities as part of the same program has the potential to be counterproductive. Agroforestry provides trees for cultivation. Reforestation provides cash to plant trees. Merging these two benefit flows may have negative consequences for the sustai nability of activities. People traditionally paid to plant trees are less likely to continue to do so without compensation At the same time, farmers motivated to plant trees on their agricultural plots for direct benefits cannot be expected to take owne rship of the survival of trees planted on communal lands without receiving tree tenure rights. The trees planted through these two approach es serve different purposes and should be managed as such. F uture programs should be cognizant of these differences 52 Murray personal communication, 2013


63 Conclusion The IOM and MCC programs in their current forms represent the evolution of countless lessons learned and applied to their methodology. Their approach to tree planting has been refined through years of field implementation and continues to be refined as highlighted in this study fall on this continuum of refinement. I present these lessons here to contribute to the ongoing development of best practices for planting trees in Haiti. These lessons also call attention to the need for multiple approaches to address deforestatio n. Many such practices have been already been tested and proven in this context. Environmental partners would be wise to build on what has already been accomplished and learn from the mistakes made over the course of the 50 year venture in tree planting in Haiti. The success of the MCC agroforestry program and AOP point to the restructuring of the relationship between smallholder and tree what Murray describes as the generating cro 53 This evolutionary step offers hope for the widespread adoption of trees into livelihood systems through the extension of the correct incentives of seedling supply, tree tenure and harvest rights. At the same time communities must be mobilized by wag e labor to reforest remote mountaintops that are not suitable for agroforestry. The outlook for reforestation in Haiti is tempered by the reality that such incentives will have to be financed by external donors for the foreseeable future. 53 Murray 1987


64 Appendix A Definitions of Variables Participation in project: This dichotomous categorical variable measures whether or not a household participated in Pwoj Pyebwa, as reported by the head of household. Age: This discrete quantitative variable represents the numerical value of a head of Sex: This dichotomous categorical variable measures the sex of the head of household, as reported by the head of household. Religion: This categorical variable measures th e religion practiced by the head of household. Literacy: This dichotomous categorical variable represent whether or not heads of household are literate, as reported by heads of household. Primary livelihood: This discrete categorical variable identifie s the primary source of income of the household, as reported by the head of household. Number of children: This discrete quantitative variable measures the numerical value of total children produced by the head of household. Number of people per househol d: This discrete quantitative variable measures the numerical value of people living at a particular household, as reported by the head of household. Housing materials: This dichotomous categorical variable measures whether the house walls and roof are co nstructed of (a combination of) concrete, corrugated sheet iron,


65 brick or mud as observed by the principal investigator. Land ownership/tenure: This dichotomous categorical variable represents whether or not the household owns the land as reported by the head of household. Other land holdings: This dichotomous categorical variable represents whether or not a household has ownership or access to more than one parcel of land, as reported by the head of household. Size of total land owned: This continuous quantitative variable measures the total size of land owned by a household, in hectares, as reported by the head of household. Slope of land: This dichotomous categorical variable measures whether the slope of the land is steep, medium or soft. Soil type s: This categorical variable measures the type of soil on land owned or access by a household, as reported by the head of household.


66 References Aquastat Country Profile: Haiti." Food and Agriculture Organization (FAO). United Nations, 2012. Accessed 14 Oct 2012. . Bannister, M.E. and Josiah S.J 1993. Agroforestry training and extension: the experience from Haiti. Agroforestry Systems 23:239 251. Bannister M.E. and Nair 2003. Agroforestry adoption in Haiti: the importance of household and farm characteristics. Agroforest Syst 57: 149 157. Bannister M. and Nair P.K.R 1990. Alley Cropping as a sustainable agricultural technology for the hillsides in Haiti: Exper ience of an agroforestry outreach project, American Journal of Alternative Agriculture vol 5: 51 59. on, Port au Prince, Haiti, March 2012. Chazdon R. 2008. Beyond deforestation: Restoring forests and ecosystem services on degraded lands. Science 320: 1458 1460. "Climate Change Vulnerability Index." Climate Change and Environment Risk Atlas Maplecroft, 2011. Accessed 28 Oct 2012. <>. Country Profile: Haiti Food and Agriculture Organization (FAO). United Nations, 2012. Accessed 7 Oct 2012. .


67 Cohen, M. "Planting No w: Agricultural Challenges and Opportunities for Haiti's Reconstruction." OXFAM Haiti OXFAM, Oct 2010. Farmer, P. Aid and Accusation: The History of Blame in Haiti. Berkeley California: University of California Press, 1992. Farmer, P. Uses of Haiti 3r d ed. New York: Common Courage Press, 2005. Fournier, F Head of Office. International Organization for Migration (IOM). Interview, Port au Prince, Haiti, 23 Jan 2013. Foxx, T. "Te a Fatigue 'The Earth is Tired': Reversing Deforestation in Haiti." Behavorial Interventions 27.2 (2012): 105 108. Web. 30 Sep. 2012. < >. Gopaul, H. "Best Practices for Public/Private Sector and Community Participation in Integrated Wate rshed Management in Caribbean Small Islands Developing States (SIDS)." Caribbean Environmental Health Institute (CEHI), 17 Nov 2005. Web. 27 Sep 2012. . n.d. Accessed 5 Nov 2012. . Havemann, T. 2011. Financing mitigation in smallholder agricultural systems: Issues and opportunities. CCAFS Working Paper 6. Copenhagen, Denmark: CCAFS. Human Development Report 2007 2008 United Nations Development Program (UNDP). United Nations, 2008. Accessed Oct 7 2012. < 2008>.


68 International Organization for Migration (IOM) Port au Prince, Haiti. January 2012. Interlandi, J. "A Tree Grows in Haiti." Newsweek Magazine. 16 Jul 2010: n. page. Web. 3 Oct. 2012. < tree grows in haiti.html >. Lea, J. D. "A Review of Literature on Charcoal in Haiti." South East Consortium for International Development Web. 27 Sep. 2012. < >. Lilin, C., and Koohafkan A. P 1987. Techniqu es Biologiques de Conservation des Sole en Haiti. Projet: Centre de Formation en Amenagement Integre des Mornes, Organization des Nations Unis Pour L'Alimentation de L'Agriculture and Ministere de L'Agriculture de Ressources Naturelles et du Developpement Rural. Lundahl, M., "The Haitian Dilemma Reexamined: Lessons From the Past in the Light of Some New Economic Theory," in Rotberg, ed., Haiti Renewed, 2002: p. 85. Ministry of the Environment. Plan of Action for the Environment. Port au Prince, Haiti 2013 Mennonite Central Committee Port au Prince, Haiti. March 2011. Murray G. 1987. The domestication of wood in Haiti: A case study in applied evolution, pp. 223 242. In: Wulff R.M. and Fiske S.J. (eds). Anthropological Praxis: Translating Knowledge into Action. Westview Press, Boulder CO.


69 Murray, G. F. "Terraces, Trees, and the Haitian Peasant: Twenty five Years of Erosion Control in Rural Haiti." Report for U.S. Agency for International Development, Port au Prince, Hai ti. Print. 1979. Murray, G.F., and Bannister, M.E. "Peasants, agroforesters, and anthropologists: A 20 year venture in income generating trees and hedgerows in Haiti." Agroforestry Systems 61 62.1 (2004): 383 397. Web. 27 Sep. 2012. < > Perfecto, I., Rice R. A., Greenberg, R. and van der Voort, M. E. Shade Coffee: A Disappearing Refuge for Biodiversity." BioScience Vol. 46 No. 8 (Sep 1996): 598 608. Port au Prince, Haiti, 15 Jul 2012. au Prince, Haiti, March 2011. Rudel T. K. (1998), Is There a Forest Transition? Deforestation, Reforestation, and Development. Rural Sociology, 63: 533 552. doi: <10.1111/j.1549 0831.1998.tb00691>. The Jou rnal of Developing Areas. 24.1 (1989): 59 76. Web. 27 Sep. 2012. < > Schwilch, G., Bachmann F., Liniger, H. "Appraising and selecting conservation measures to mitigat e desertification and land degradation based on stakeholder participation and global best practices." Land Degradation and Development 20.3 (2009): 308 326. Web. 27 Sep. 2012. < >


70 Smucker, G. R., et al. United States Government. United States Agency for Development. Environmental Vulnerability in Haiti Apr 2007. Web. . Smuck er G. and Timyan J. 1995. Impact of tree planting in Haiti: 1982 1995. Haiti P roductive Land Use Systems Pro ject, Southeast Consortium for International Development and Auburn University, Petionville, Haiti. Smucker, G., R., White, T.A., and Bannister, M. "Land Tenure and the Adoption of Agricultural Technology in Haiti." Victoria Falls, Zambia, 2002. "Timeline of Events: Haiti." British Broadcasting Corporation (BBC) 2012. Accessed 17 Sep 2012. . Schulz D., and Marcella, G., Reconciling the Irreconcilable: The Troubled Outlook for U.S. Policy Toward Haiti (Carlisle, Pa.: Strategic Studies Institute, 1994), p. 5. Schulz, D. Haiti: Will Things Fall Apart? Parameters Winter 1997 98, pp. 73 91. United Accessed 3 Mar 2014. < > United Nations Development Program (UNDP). "Rapport sur l'assistance externe en Haiti: 19 76 1990." Port au Prince, Haiti. 2010. United States Government. Central Intelligence Agency (CIA). World Fact Book: Haiti 2012. Accessed 18 Sep 2012. . United States Government. Library of Congress. S.1183 Haiti Reforestation Act of 2010. 21 Sept 2011.


71 United States Government. Federal Research Division, Library of Congress. Haiti: Country Studies Mar 2011. Accessed 16 Sept 2012. . U.S. Department of State. Bureau of Public Affairs. Office of Public Communication. White, T.A., and J.L. Jickling, Peasants, experts, and land use in Haiti: Lessons from indigenous and project technology, 1995. Whi te, T.A. "Policy Lessons from History and Natural Resource Projects in Rural Haiti." Environmental and Natural Resources Policy and Training Project (EPAT), USAID, Nov 1994. Web. 30 Sep 2012. < http :// >. Common Property and Collective Action: Lessons from Economic Development and Cultural Change. 43.1 (1994): 1 41. Web. 30 Sep 2012. < >. Wilentz, Amy. The Rainy Season: Haiti Since Duvalier New York: Simon & Schuster, 1989.